Home

_module.args

Additional arguments passed to each module in addition to ones like lib, config, and pkgs, modulesPath.

This option is also available to all submodules. Submodules do not inherit args from their parent module, nor do they provide args to their parent module or sibling submodules. The sole exception to this is the argument name which is provided by parent modules to a submodule and contains the attribute name the submodule is bound to, or a unique generated name if it is not bound to an attribute.

Some arguments are already passed by default, of which the following cannot be changed with this option:

• lib: The nixpkgs library.

• config: The results of all options after merging the values from all modules together.

• options: The options declared in all modules.

• specialArgs: The specialArgs argument passed to evalModules.

• All attributes of specialArgs

Whereas option values can generally depend on other option values thanks to laziness, this does not apply to imports, which must be computed statically before anything else.

For this reason, callers of the module system can provide specialArgs which are available during import resolution.

For NixOS, specialArgs includes modulesPath, which allows you to import extra modules from the nixpkgs package tree without having to somehow make the module aware of the location of the nixpkgs or NixOS directories.

{ modulesPath, ... }: { imports = [ (modulesPath + "/profiles/minimal.nix") ]; }

For NixOS, the default value for this option includes at least this argument:

• pkgs: The nixpkgs package set according to the nixpkgs.pkgs option.

Type: lazy attribute set of raw value

Declared by: - \

cudaPackagesPackageSets

CUDA Package Sets

Type: submodule

Default:

{
  cudaPackages_10_0 = {
    cudaVersion = "10.0";
  };
  cudaPackages_10_1 = {
    cudaVersion = "10.1";
  };
  cudaPackages_10_2 = {
    cudaVersion = "10.2";
  };
  cudaPackages_11_0 = {
    cudaVersion = "11.0";
  };
  cudaPackages_11_1 = {
    cudaVersion = "11.1";
  };
  cudaPackages_11_2 = {
    cudaVersion = "11.2";
  };
  cudaPackages_11_3 = {
    cudaVersion = "11.3";
  };
  cudaPackages_11_4 = {
    cudaVersion = "11.4";
  };
  cudaPackages_11_5 = {
    cudaVersion = "11.5";
  };
  cudaPackages_11_6 = {
    cudaVersion = "11.6";
  };
  cudaPackages_11_7 = {
    cudaVersion = "11.7";
  };
  cudaPackages_11_8 = {
    cudaVersion = "11.8";
  };
  cudaPackages_12_0 = {
    cudaVersion = "12.0";
  };
  cudaPackages_12_1 = {
    cudaVersion = "12.1";
  };
  cudaPackages_12_2 = {
    cudaVersion = "12.2";
  };
  cudaPackages_12_3 = {
    cudaVersion = "12.3";
  };
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_10_0

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_10_0.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.0"
  "3.2"
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "6.0"
  "6.1"
  "7.0"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.0";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.2";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.0"
  "3.2"
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.0";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.2";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "10"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_0.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_1

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_10_1.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.0"
  "3.2"
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "6.0"
  "6.1"
  "7.0"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.0";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.2";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.0"
  "3.2"
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.0";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.2";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "10.1"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "10"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_1.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_2

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_10_2.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.0"
  "3.2"
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "6.0"
  "6.1"
  "7.0"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.0";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.2";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.0"
  "3.2"
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.0";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.2";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "10"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_10_2.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_0

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_11_0.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "6.0"
  "6.1"
  "7.0"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "11.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "11"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_0.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_1

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_11_1.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "11.1"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "11"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_1.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_2

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_11_2.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "11.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "11"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_2.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_3

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_11_3.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "11.3"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "11"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_3.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_4

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_11_4.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "11.4"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "11"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_4.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_5

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_11_5.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
  "8.7"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "11.5"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "11"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_5.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_6

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_11_6.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
  "8.7"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "11.6"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "11"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_6.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_7

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_11_7.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default: [ ]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
  "8.7"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "11.7"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "11"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_7.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_8

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_11_8.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default:

[
  "Ada"
]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
  "8.9"
  "9.0"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "3.5"
  "3.7"
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
  "8.7"
  "8.9"
  "9.0"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "11.8"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "11"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_11_8.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_0

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_12_0.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default:

[
  "Ada"
]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
  "8.9"
  "9.0"
  "9.0a"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
  "8.7"
  "8.9"
  "9.0"
  "9.0a"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "12.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "12"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_0.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_1

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_12_1.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default:

[
  "Ada"
]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
  "8.9"
  "9.0"
  "9.0a"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
  "8.7"
  "8.9"
  "9.0"
  "9.0a"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "12.1"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "12"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_1.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_2

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_12_2.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default:

[
  "Ada"
]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
  "8.9"
  "9.0"
  "9.0a"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
  "8.7"
  "8.9"
  "9.0"
  "9.0a"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "12.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "12"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_2.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_3

A CUDA package set for a particular version of CUDA

Type: submodule

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

cudaPackagesPackageSets.cudaPackages_12_3.cudaAtLeast

A utility to check if the CUDA version is at least a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaAtMost

A utility to check if the CUDA version is at most a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.archNames

The names of the architectures to build for

Type: list of string

Default:

[
  "Ada"
]

Example:

[
  "Turing"
  "Ampere"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.arches

The architectures to build for. Includes a virtual architecture if forward compatibility is enabled.

Type: list of (A real/physical CUDA architecture name. or A virtual CUDA architecture name.)

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaCapabilities

CUDA capabilities (hardware generations) to build for

Type: list of A CUDA capability version.

Default:

[
  "8.9"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaForwardCompat

Build with forward compatibility gencode (+PTX) to support future GPU generations

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaGpus

CUDA GPUs to build for

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaJetsonGpus

CUDA GPUs to build for, including only Jetson devices

Type: list of (submodule)

Default: [ ]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaNonJetsonGpus

CUDA GPUs to build for, excluding Jetson devices

Type: list of (submodule)

Default:

[
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaNonJetsonGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaNonJetsonGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaNonJetsonGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaNonJetsonGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaNonJetsonGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaNonJetsonGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.cudaSupport

Build packages with CUDA support

Type: boolean

Default: true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.defaultCapabilities

The default compute capabilities for a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "6.0"
  "6.1"
  "7.0"
  "7.5"
  "8.0"
  "8.6"
  "8.9"
  "9.0"
  "9.0a"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.defaultGpus

The default GPUs for a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.defaultGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.defaultGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.defaultGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.defaultGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.defaultGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.defaultGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.gencodeArgs

The gencode arguments to pass to NVCC

Type: list of A `gencode` argument for the `nvcc` compiler.

Default:

[
  "-gencode=arch=compute_89,code=sm_89"
]

Example:

[
  "-gencode=arch=compute_75,code=sm_75"
  "-gencode=arch=compute_86,code=sm_86"
  "-gencode=arch=compute_86,code=compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.gencodeString

A space-separated string of CUDA gencode arguments to pass to NVCC

Type: string

Default: "-gencode=arch=compute_89,code=sm_89"

Example: "-gencode=arch=compute_75,code=sm_75 ... -gencode=arch=compute_86,code=compute_86"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.isJetsonBuild

Whether the build is for a Jetson device.

Jetson devices cannot be targeted by the same binaries which target non-Jetson devices. While NVIDIA provides both linux-aarch64 and linux-sbsa packages, which both target aarch64, they are built with different settings and cannot be mixed.

Type: boolean

Default: false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.realArches

The physical architectures to build for

Type: list of A real/physical CUDA architecture name.

Default:

[
  "sm_89"
]

Example:

[
  "sm_75"
  "sm_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.supportedCapabilities

The compute capabilities supported by a given CUDA version

Type: list of A CUDA capability version.

Default:

[
  "5.0"
  "5.2"
  "5.3"
  "6.0"
  "6.1"
  "6.2"
  "7.0"
  "7.2"
  "7.5"
  "8.0"
  "8.6"
  "8.7"
  "8.9"
  "9.0"
  "9.0a"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.supportedGpus

The GPUs supported by a given CUDA version

Type: list of (submodule)

Default:

[
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.supportedGpus.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.supportedGpus.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.supportedGpus.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.supportedGpus.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.supportedGpus.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.supportedGpus.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.data.virtualArches

The virtual architectures to build for. These are typically used for forward compatibility, when trying to support an architecture newer than the CUDA version allows.

Type: list of A virtual CUDA architecture name.

Default:

[
  "compute_89"
]

Example:

[
  "compute_75"
  "compute_86"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.utils.archMapper

Maps a feature across a list of compute capabilities to produce a list of architectures.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “sm_80” “sm_86” “sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.utils.dropDot

Drops dots from a string.

Type: function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.utils.gencodeMapper

Maps a feature across a list of compute capabilities to produce a list of gencode arguments.

For example, “sm” and [ “8.0” “8.6” “8.7” ] yield [ “-gencode=arch=compute_80,code=sm_80” “-gencode=arch=compute_86,code=sm_86” “-gencode=arch=compute_87,code=sm_87” ].

Type: function that evaluates to a(n) function that evaluates to a(n) list of string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.utils.getNixSystem

Maps NVIDIA redist arch to Nix system.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “aaarch64-linux”, “x86_64-linux”, “ppc64le-linux”, “x86_64-windows”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.utils.getRedistArch

Maps Nix system to NVIDIA redist arch.

NOTE: We swap out the default linux-sbsa redist (for server-grade ARM chips) with the linux-aarch64 redist (which is for Jetson devices) if we’re building any Jetson devices. Since both are based on aarch64, we can only have one or the other, otherwise there’s an ambiguity as to which should be used.

NOTE: This function will be called by unsupported systems because cudaPackages is part of all-packages.nix, which is evaluated on all systems. As such, we need to handle unsupported systems gracefully.

Type: function that evaluates to a(n) (one of “unsupported”, “linux-aarch64”, “linux-sbsa”, “linux-x86_64”, “linux-ppc64le”, “windows-x86_64”)

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.utils.isDefault

Whether a GPU is included in the default capabilities

Type: function that evaluates to a(n) boolean

Default: <function, args: {dontDefaultAfter, isJetson}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaFlags.utils.isSupported

Whether a CUDA version supports a given GPU

Type: function that evaluates to a(n) boolean

Default: <function, args: {maxCudaVersion, minCudaVersion}>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-flags.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaMajorMinorVersion

The major and minor version of the CUDA package set

Type: string

Default: "12.3"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaMajorVersion

The major version of the CUDA package set

Type: string

Default: "12"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaNewer

A utility to check if the CUDA version is newer than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaOlder

A utility to check if the CUDA version is older than a given version

Type: function that evaluates to a(n) boolean

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaPackagesPackageSets.cudaPackages_12_3.cudaVersion

The CUDA version of the package set

Type: A version number with a major and minor component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/cuda-packages-package-set.nix

cudaVersions

A list of CUDA versions to create package sets for

Type: list of A version number with a major and minor component.

Default:

[
  "10.0"
  "10.1"
  "10.2"
  "11.0"
  "11.1"
  "11.2"
  "11.3"
  "11.4"
  "11.5"
  "11.6"
  "11.7"
  "11.8"
  "12.0"
  "12.1"
  "12.2"
  "12.3"
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules

generic.data.gpus.all

Known GPUs

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.0";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.2";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.all.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.all.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.all.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.all.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.all.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.all.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.jetsons

Known GPUs that are part of NVIDIA’s Jetson line of embedded computers

Type: list of (submodule)

Default:

[
  {
    archName = "Maxwell";
    computeCapability = "5.3";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.2";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.7";
    dontDefaultAfter = null;
    isJetson = true;
    maxCudaVersion = null;
    minCudaVersion = "11.5";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.jetsons.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.jetsons.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.jetsons.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.jetsons.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of true, false

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.jetsons.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.jetsons.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.nonJetsons

Known GPUs that are not part of NVIDIA’s Jetson line of embedded computers

Type: list of (submodule)

Default:

[
  {
    archName = "Kepler";
    computeCapability = "3.0";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.2";
    dontDefaultAfter = "10.2";
    isJetson = false;
    maxCudaVersion = "10.2";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.5";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Kepler";
    computeCapability = "3.7";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = "11.8";
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.0";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Maxwell";
    computeCapability = "5.2";
    dontDefaultAfter = "11.0";
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Pascal";
    computeCapability = "6.1";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Volta";
    computeCapability = "7.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Turing";
    computeCapability = "7.5";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "10.0";
  }
  {
    archName = "Ampere";
    computeCapability = "8.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ampere";
    computeCapability = "8.6";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.2";
  }
  {
    archName = "Ada";
    computeCapability = "8.9";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "11.8";
  }
  {
    archName = "Hopper";
    computeCapability = "9.0a";
    dontDefaultAfter = null;
    isJetson = false;
    maxCudaVersion = null;
    minCudaVersion = "12.0";
  }
]

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.nonJetsons.*.archName

Name of the microarchitecture

Type: A CUDA architecture name, like Ampere or Ada.

Example: "Kepler"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.nonJetsons.*.computeCapability

Compute capability of the GPU

Type: A CUDA capability version.

Example: "3.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.nonJetsons.*.dontDefaultAfter

The CUDA version after which to exclude this GPU from the list of default capabilities we build. null means we always include this GPU in the default capabilities if it is supported.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.nonJetsons.*.isJetson

Whether a GPU is part of NVIDIA’s line of Jetson embedded computers.

This field is notable because it tells us what architecture to build for (as Jetson devices are aarch64). More on Jetson devices here: https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

NOTE: These architectures are only built upon request.

Type: one of false, true

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.nonJetsons.*.maxCudaVersion

The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.

Type: null or A version number with a major and minor component.

Example: "10.2"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.gpus.nonJetsons.*.minCudaVersion

The minimum (inclusive) CUDA version that supports this GPU.

Type: A version number with a major and minor component.

Example: "10.0"

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.data.nvccCompatibilities

Mapping of CUDA version to known NVCC compatibilities with GCC and Clang

Type: attribute set of (submodule)

Default:

{
  "10.0" = {
    clangMaxMajorVersion = "6";
    clangMinMajorVersion = "6";
    gccMaxMajorVersion = "7";
    gccMinMajorVersion = "5";
  };
  "10.1" = {
    clangMaxMajorVersion = "7";
    clangMinMajorVersion = "6";
    gccMaxMajorVersion = "8";
    gccMinMajorVersion = "5";
  };
  "10.2" = {
    clangMaxMajorVersion = "8";
    clangMinMajorVersion = "6";
    gccMaxMajorVersion = "8";
    gccMinMajorVersion = "5";
  };
  "11.0" = {
    clangMaxMajorVersion = "9";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "9";
    gccMinMajorVersion = "6";
  };
  "11.1" = {
    clangMaxMajorVersion = "10";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "10";
    gccMinMajorVersion = "6";
  };
  "11.2" = {
    clangMaxMajorVersion = "11";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "10";
    gccMinMajorVersion = "6";
  };
  "11.3" = {
    clangMaxMajorVersion = "11";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "10";
    gccMinMajorVersion = "6";
  };
  "11.4" = {
    clangMaxMajorVersion = "12";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "10";
    gccMinMajorVersion = "6";
  };
  "11.5" = {
    clangMaxMajorVersion = "12";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "10";
    gccMinMajorVersion = "6";
  };
  "11.6" = {
    clangMaxMajorVersion = "12";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "11";
    gccMinMajorVersion = "6";
  };
  "11.7" = {
    clangMaxMajorVersion = "13";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "11";
    gccMinMajorVersion = "6";
  };
  "11.8" = {
    clangMaxMajorVersion = "14";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "11";
    gccMinMajorVersion = "6";
  };
  "12.0" = {
    clangMaxMajorVersion = "14";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "12";
    gccMinMajorVersion = "6";
  };
  "12.1" = {
    clangMaxMajorVersion = "15";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "12";
    gccMinMajorVersion = "6";
  };
  "12.2" = {
    clangMaxMajorVersion = "16";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "12";
    gccMinMajorVersion = "6";
  };
  "12.3" = {
    clangMaxMajorVersion = "16";
    clangMinMajorVersion = "7";
    gccMaxMajorVersion = "12";
    gccMinMajorVersion = "6";
  };
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/nvcc-compatibilities

generic.data.nvccCompatibilities.\.clangMaxMajorVersion

Maximum Clang major version to support this CUDA version

Type: A version number with a major component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/nvcc-compatibilities

generic.data.nvccCompatibilities.\.clangMinMajorVersion

Minimum Clang major version to support this CUDA version

Type: A version number with a major component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/nvcc-compatibilities

generic.data.nvccCompatibilities.\.gccMaxMajorVersion

Maximum GCC major version to support this CUDA version

Type: A version number with a major component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/nvcc-compatibilities

generic.data.nvccCompatibilities.\.gccMinMajorVersion

Minimum GCC major version to support this CUDA version

Type: A version number with a major component.

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/nvcc-compatibilities

generic.types

A set of types used in building cudaPackages.

Type: attribute set of optionType

Default: { }

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/nvcc-compatibilities - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic

generic.types.archName

A CUDA architecture name, like Ampere or Ada.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "A CUDA architecture name, like Ampere or Ada.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \"^[[:alpha:]]+\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "archName";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.types.cudaArch

A real/physical CUDA architecture name.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "A real/physical CUDA architecture name.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \"^sm_[[:digit:]]+[a-z]?\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "cudaArch";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.types.cudaCapability

A CUDA capability version.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "A CUDA capability version.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \"^[[:digit:]]+\\\\.[[:digit:]]+[a-z]?\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "cudaCapability";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.types.gencodeArg

A gencode argument for the nvcc compiler.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "A `gencode` argument for the `nvcc` compiler.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \"^-gencode=arch=compute_[[:digit:]]+[a-z]?,code=(sm|compute)_[[:digit:]]+[a-z]?\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "gencodeArg";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.types.gpu

A GPU

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "submodule";
  descriptionClass = null;
  emptyValue = {
    value = { };
  };
  functor = {
    binOp = <function>;
    name = "submodule";
    payload = {
      class = null;
      description = null;
      modules = [
        {
          options = {
            archName = {
              _type = "option";
              description = "Name of the microarchitecture";
              example = "Kepler";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A CUDA architecture name, like Ampere or Ada.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^[[:alpha:]]+\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "archName";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            computeCapability = {
              _type = "option";
              description = "Compute capability of the GPU";
              example = "3.0";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A CUDA capability version.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^[[:digit:]]+\\\\.[[:digit:]]+[a-z]?\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "cudaCapability";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            dontDefaultAfter = {
              _type = "option";
              description = ''
                The CUDA version after which to exclude this GPU from the list of default capabilities
                we build. null means we always include this GPU in the default capabilities if it is
                supported.
              '';
              example = "10.2";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "null or A version number with a major and minor component.";
                descriptionClass = "conjunction";
                emptyValue = {
                  value = null;
                };
                functor = {
                  binOp = <function>;
                  name = "nullOr";
                  payload = null;
                  type = <function>;
                  wrapped = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "nullOr";
                nestedTypes = {
                  elemType = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            isJetson = {
              _type = "option";
              description = ''
                Whether a GPU is part of NVIDIA's line of Jetson embedded computers.

                This field is notable because it tells us what architecture to build for (as Jetson devices are
                aarch64). More on Jetson devices here:
                https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

                NOTE: These architectures are only built upon request.
              '';
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "one of true, false";
                descriptionClass = "conjunction";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "enum";
                  payload = [
                    true
                    false
                  ];
                  type = <function>;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "enum";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            maxCudaVersion = {
              _type = "option";
              description = "The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.";
              example = "10.2";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "null or A version number with a major and minor component.";
                descriptionClass = "conjunction";
                emptyValue = {
                  value = null;
                };
                functor = {
                  binOp = <function>;
                  name = "nullOr";
                  payload = null;
                  type = <function>;
                  wrapped = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "nullOr";
                nestedTypes = {
                  elemType = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            minCudaVersion = {
              _type = "option";
              description = "The minimum (inclusive) CUDA version that supports this GPU.";
              example = "10.0";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
          };
        }
      ];
      shorthandOnlyDefinesConfig = false;
      specialArgs = { };
    };
    type = <function, args: {class?, description?, modules, shorthandOnlyDefinesConfig?, specialArgs?}>;
    wrapped = null;
  };
  getSubModules = [
    {
      options = {
        archName = {
          _type = "option";
          description = "Name of the microarchitecture";
          example = "Kepler";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A CUDA architecture name, like Ampere or Ada.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^[[:alpha:]]+\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "archName";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        computeCapability = {
          _type = "option";
          description = "Compute capability of the GPU";
          example = "3.0";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A CUDA capability version.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^[[:digit:]]+\\\\.[[:digit:]]+[a-z]?\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "cudaCapability";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        dontDefaultAfter = {
          _type = "option";
          description = ''
            The CUDA version after which to exclude this GPU from the list of default capabilities
            we build. null means we always include this GPU in the default capabilities if it is
            supported.
          '';
          example = "10.2";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "null or A version number with a major and minor component.";
            descriptionClass = "conjunction";
            emptyValue = {
              value = null;
            };
            functor = {
              binOp = <function>;
              name = "nullOr";
              payload = null;
              type = <function>;
              wrapped = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "nullOr";
            nestedTypes = {
              elemType = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        isJetson = {
          _type = "option";
          description = ''
            Whether a GPU is part of NVIDIA's line of Jetson embedded computers.

            This field is notable because it tells us what architecture to build for (as Jetson devices are
            aarch64). More on Jetson devices here:
            https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

            NOTE: These architectures are only built upon request.
          '';
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "one of true, false";
            descriptionClass = "conjunction";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "enum";
              payload = [
                true
                false
              ];
              type = <function>;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "enum";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        maxCudaVersion = {
          _type = "option";
          description = "The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.";
          example = "10.2";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "null or A version number with a major and minor component.";
            descriptionClass = "conjunction";
            emptyValue = {
              value = null;
            };
            functor = {
              binOp = <function>;
              name = "nullOr";
              payload = null;
              type = <function>;
              wrapped = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "nullOr";
            nestedTypes = {
              elemType = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        minCudaVersion = {
          _type = "option";
          description = "The minimum (inclusive) CUDA version that supports this GPU.";
          example = "10.0";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major and minor component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorMinorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
      };
    }
  ];
  getSubOptions = <function>;
  merge = <function>;
  name = "submodule";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.types.jetsonGpu

A GPU that is part of NVIDIA’s Jetson line of embedded computers

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "submodule";
  descriptionClass = null;
  emptyValue = {
    value = { };
  };
  functor = {
    binOp = <function>;
    name = "submodule";
    payload = {
      class = null;
      description = null;
      modules = [
        {
          options = {
            archName = {
              _type = "option";
              description = "Name of the microarchitecture";
              example = "Kepler";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A CUDA architecture name, like Ampere or Ada.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^[[:alpha:]]+\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "archName";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            computeCapability = {
              _type = "option";
              description = "Compute capability of the GPU";
              example = "3.0";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A CUDA capability version.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^[[:digit:]]+\\\\.[[:digit:]]+[a-z]?\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "cudaCapability";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            dontDefaultAfter = {
              _type = "option";
              description = ''
                The CUDA version after which to exclude this GPU from the list of default capabilities
                we build. null means we always include this GPU in the default capabilities if it is
                supported.
              '';
              example = "10.2";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "null or A version number with a major and minor component.";
                descriptionClass = "conjunction";
                emptyValue = {
                  value = null;
                };
                functor = {
                  binOp = <function>;
                  name = "nullOr";
                  payload = null;
                  type = <function>;
                  wrapped = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "nullOr";
                nestedTypes = {
                  elemType = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            isJetson = {
              _type = "option";
              description = ''
                Whether a GPU is part of NVIDIA's line of Jetson embedded computers.

                This field is notable because it tells us what architecture to build for (as Jetson devices are
                aarch64). More on Jetson devices here:
                https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

                NOTE: These architectures are only built upon request.
              '';
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "one of true, false";
                descriptionClass = "conjunction";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "enum";
                  payload = [
                    true
                    false
                  ];
                  type = <function>;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "enum";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            maxCudaVersion = {
              _type = "option";
              description = "The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.";
              example = "10.2";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "null or A version number with a major and minor component.";
                descriptionClass = "conjunction";
                emptyValue = {
                  value = null;
                };
                functor = {
                  binOp = <function>;
                  name = "nullOr";
                  payload = null;
                  type = <function>;
                  wrapped = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "nullOr";
                nestedTypes = {
                  elemType = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            minCudaVersion = {
              _type = "option";
              description = "The minimum (inclusive) CUDA version that supports this GPU.";
              example = "10.0";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
          };
        }
        {
          options = {
            isJetson = {
              _type = "option";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "value true (singular enum)";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "enum";
                  payload = [
                    true
                  ];
                  type = <function>;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "enum";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
          };
        }
      ];
      shorthandOnlyDefinesConfig = false;
      specialArgs = { };
    };
    type = <function, args: {class?, description?, modules, shorthandOnlyDefinesConfig?, specialArgs?}>;
    wrapped = null;
  };
  getSubModules = [
    {
      options = {
        archName = {
          _type = "option";
          description = "Name of the microarchitecture";
          example = "Kepler";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A CUDA architecture name, like Ampere or Ada.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^[[:alpha:]]+\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "archName";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        computeCapability = {
          _type = "option";
          description = "Compute capability of the GPU";
          example = "3.0";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A CUDA capability version.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^[[:digit:]]+\\\\.[[:digit:]]+[a-z]?\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "cudaCapability";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        dontDefaultAfter = {
          _type = "option";
          description = ''
            The CUDA version after which to exclude this GPU from the list of default capabilities
            we build. null means we always include this GPU in the default capabilities if it is
            supported.
          '';
          example = "10.2";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "null or A version number with a major and minor component.";
            descriptionClass = "conjunction";
            emptyValue = {
              value = null;
            };
            functor = {
              binOp = <function>;
              name = "nullOr";
              payload = null;
              type = <function>;
              wrapped = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "nullOr";
            nestedTypes = {
              elemType = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        isJetson = {
          _type = "option";
          description = ''
            Whether a GPU is part of NVIDIA's line of Jetson embedded computers.

            This field is notable because it tells us what architecture to build for (as Jetson devices are
            aarch64). More on Jetson devices here:
            https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

            NOTE: These architectures are only built upon request.
          '';
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "one of true, false";
            descriptionClass = "conjunction";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "enum";
              payload = [
                true
                false
              ];
              type = <function>;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "enum";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        maxCudaVersion = {
          _type = "option";
          description = "The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.";
          example = "10.2";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "null or A version number with a major and minor component.";
            descriptionClass = "conjunction";
            emptyValue = {
              value = null;
            };
            functor = {
              binOp = <function>;
              name = "nullOr";
              payload = null;
              type = <function>;
              wrapped = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "nullOr";
            nestedTypes = {
              elemType = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        minCudaVersion = {
          _type = "option";
          description = "The minimum (inclusive) CUDA version that supports this GPU.";
          example = "10.0";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major and minor component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorMinorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
      };
    }
    {
      options = {
        isJetson = {
          _type = "option";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "value true (singular enum)";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "enum";
              payload = [
                true
              ];
              type = <function>;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "enum";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
      };
    }
  ];
  getSubOptions = <function>;
  merge = <function>;
  name = "submodule";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.types.libSoName

The name of a shared object file.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "The name of a shared object file.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \".*\\\\.so(\\\\.[[:digit:]]+)*\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "libSoName";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.types.majorMinorPatchBuildVersion

A version number with a major, minor, patch, and build component.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "A version number with a major, minor, patch, and build component.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\\\.([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "majorMinorPatchBuildVersion";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.types.majorMinorPatchVersion

A version number with a major, minor, and patch component.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "A version number with a major, minor, and patch component.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "majorMinorPatchVersion";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.types.majorMinorVersion

A version number with a major and minor component.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "A version number with a major and minor component.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "majorMinorVersion";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.types.majorVersion

A version number with a major component.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "A version number with a major component.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \"^([[:digit:]]+)\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "majorVersion";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.types.nonJetsonGpu

A GPU that is not part of NVIDIA’s Jetson line of embedded computers

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "submodule";
  descriptionClass = null;
  emptyValue = {
    value = { };
  };
  functor = {
    binOp = <function>;
    name = "submodule";
    payload = {
      class = null;
      description = null;
      modules = [
        {
          options = {
            archName = {
              _type = "option";
              description = "Name of the microarchitecture";
              example = "Kepler";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A CUDA architecture name, like Ampere or Ada.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^[[:alpha:]]+\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "archName";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            computeCapability = {
              _type = "option";
              description = "Compute capability of the GPU";
              example = "3.0";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A CUDA capability version.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^[[:digit:]]+\\\\.[[:digit:]]+[a-z]?\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "cudaCapability";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            dontDefaultAfter = {
              _type = "option";
              description = ''
                The CUDA version after which to exclude this GPU from the list of default capabilities
                we build. null means we always include this GPU in the default capabilities if it is
                supported.
              '';
              example = "10.2";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "null or A version number with a major and minor component.";
                descriptionClass = "conjunction";
                emptyValue = {
                  value = null;
                };
                functor = {
                  binOp = <function>;
                  name = "nullOr";
                  payload = null;
                  type = <function>;
                  wrapped = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "nullOr";
                nestedTypes = {
                  elemType = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            isJetson = {
              _type = "option";
              description = ''
                Whether a GPU is part of NVIDIA's line of Jetson embedded computers.

                This field is notable because it tells us what architecture to build for (as Jetson devices are
                aarch64). More on Jetson devices here:
                https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

                NOTE: These architectures are only built upon request.
              '';
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "one of true, false";
                descriptionClass = "conjunction";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "enum";
                  payload = [
                    true
                    false
                  ];
                  type = <function>;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "enum";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            maxCudaVersion = {
              _type = "option";
              description = "The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.";
              example = "10.2";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "null or A version number with a major and minor component.";
                descriptionClass = "conjunction";
                emptyValue = {
                  value = null;
                };
                functor = {
                  binOp = <function>;
                  name = "nullOr";
                  payload = null;
                  type = <function>;
                  wrapped = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "nullOr";
                nestedTypes = {
                  elemType = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major and minor component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorMinorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            minCudaVersion = {
              _type = "option";
              description = "The minimum (inclusive) CUDA version that supports this GPU.";
              example = "10.0";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
          };
        }
        {
          options = {
            isJetson = {
              _type = "option";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "value false (singular enum)";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "enum";
                  payload = [
                    false
                  ];
                  type = <function>;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "enum";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
          };
        }
      ];
      shorthandOnlyDefinesConfig = false;
      specialArgs = { };
    };
    type = <function, args: {class?, description?, modules, shorthandOnlyDefinesConfig?, specialArgs?}>;
    wrapped = null;
  };
  getSubModules = [
    {
      options = {
        archName = {
          _type = "option";
          description = "Name of the microarchitecture";
          example = "Kepler";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A CUDA architecture name, like Ampere or Ada.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^[[:alpha:]]+\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "archName";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        computeCapability = {
          _type = "option";
          description = "Compute capability of the GPU";
          example = "3.0";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A CUDA capability version.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^[[:digit:]]+\\\\.[[:digit:]]+[a-z]?\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "cudaCapability";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        dontDefaultAfter = {
          _type = "option";
          description = ''
            The CUDA version after which to exclude this GPU from the list of default capabilities
            we build. null means we always include this GPU in the default capabilities if it is
            supported.
          '';
          example = "10.2";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "null or A version number with a major and minor component.";
            descriptionClass = "conjunction";
            emptyValue = {
              value = null;
            };
            functor = {
              binOp = <function>;
              name = "nullOr";
              payload = null;
              type = <function>;
              wrapped = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "nullOr";
            nestedTypes = {
              elemType = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        isJetson = {
          _type = "option";
          description = ''
            Whether a GPU is part of NVIDIA's line of Jetson embedded computers.

            This field is notable because it tells us what architecture to build for (as Jetson devices are
            aarch64). More on Jetson devices here:
            https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/

            NOTE: These architectures are only built upon request.
          '';
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "one of true, false";
            descriptionClass = "conjunction";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "enum";
              payload = [
                true
                false
              ];
              type = <function>;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "enum";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        maxCudaVersion = {
          _type = "option";
          description = "The maximum (exclusive) CUDA version that supports this GPU. null means there is no maximum.";
          example = "10.2";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "null or A version number with a major and minor component.";
            descriptionClass = "conjunction";
            emptyValue = {
              value = null;
            };
            functor = {
              binOp = <function>;
              name = "nullOr";
              payload = null;
              type = <function>;
              wrapped = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "nullOr";
            nestedTypes = {
              elemType = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major and minor component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorMinorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        minCudaVersion = {
          _type = "option";
          description = "The minimum (inclusive) CUDA version that supports this GPU.";
          example = "10.0";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major and minor component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\\\.([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorMinorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
      };
    }
    {
      options = {
        isJetson = {
          _type = "option";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "value false (singular enum)";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "enum";
              payload = [
                false
              ];
              type = <function>;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "enum";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
      };
    }
  ];
  getSubOptions = <function>;
  merge = <function>;
  name = "submodule";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/gpus

generic.types.nvccCompatibilities

Mapping of CUDA version to known NVCC compatibilities with GCC and Clang

Taken from https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#host-compiler-support-policy

NVCC performs a version check on the host compiler’s major version and so newer minor versions of the compilers listed below will be supported, but major versions falling outside the range will not be supported.

NOTE: These constraints don’t apply to Jetson, which uses something else. NOTE: NVIDIA can and will add support for newer compilers even during patch releases. E.g.: CUDA 12.2.1 maxxed out with support for Clang 15.0; 12.2.2 added support for Clang 16.0. NOTE: Because all platforms NVIDIA supports use GCC and Clang, we omit the architectures here.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "attribute set of (submodule)";
  descriptionClass = "composite";
  emptyValue = {
    value = { };
  };
  functor = {
    binOp = <function>;
    name = "attrsOf";
    payload = null;
    type = <function>;
    wrapped = {
      _type = "option-type";
      check = <function>;
      deprecationMessage = null;
      description = "submodule";
      descriptionClass = null;
      emptyValue = {
        value = { };
      };
      functor = {
        binOp = <function>;
        name = "submodule";
        payload = {
          class = null;
          description = null;
          modules = [
            {
              options = {
                clangMaxMajorVersion = {
                  _type = "option";
                  description = "Maximum Clang major version to support this CUDA version";
                  type = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                clangMinMajorVersion = {
                  _type = "option";
                  description = "Minimum Clang major version to support this CUDA version";
                  type = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                gccMaxMajorVersion = {
                  _type = "option";
                  description = "Maximum GCC major version to support this CUDA version";
                  type = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                gccMinMajorVersion = {
                  _type = "option";
                  description = "Minimum GCC major version to support this CUDA version";
                  type = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
              };
            }
          ];
          shorthandOnlyDefinesConfig = false;
          specialArgs = { };
        };
        type = <function, args: {class?, description?, modules, shorthandOnlyDefinesConfig?, specialArgs?}>;
        wrapped = null;
      };
      getSubModules = [
        {
          options = {
            clangMaxMajorVersion = {
              _type = "option";
              description = "Maximum Clang major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            clangMinMajorVersion = {
              _type = "option";
              description = "Minimum Clang major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            gccMaxMajorVersion = {
              _type = "option";
              description = "Maximum GCC major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            gccMinMajorVersion = {
              _type = "option";
              description = "Minimum GCC major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
          };
        }
      ];
      getSubOptions = <function>;
      merge = <function>;
      name = "submodule";
      nestedTypes = { };
      substSubModules = <function>;
      typeMerge = <function>;
    };
  };
  getSubModules = [
    {
      options = {
        clangMaxMajorVersion = {
          _type = "option";
          description = "Maximum Clang major version to support this CUDA version";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        clangMinMajorVersion = {
          _type = "option";
          description = "Minimum Clang major version to support this CUDA version";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        gccMaxMajorVersion = {
          _type = "option";
          description = "Maximum GCC major version to support this CUDA version";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        gccMinMajorVersion = {
          _type = "option";
          description = "Minimum GCC major version to support this CUDA version";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
      };
    }
  ];
  getSubOptions = <function>;
  merge = <function>;
  name = "attrsOf";
  nestedTypes = {
    elemType = {
      _type = "option-type";
      check = <function>;
      deprecationMessage = null;
      description = "submodule";
      descriptionClass = null;
      emptyValue = {
        value = { };
      };
      functor = {
        binOp = <function>;
        name = "submodule";
        payload = {
          class = null;
          description = null;
          modules = [
            {
              options = {
                clangMaxMajorVersion = {
                  _type = "option";
                  description = "Maximum Clang major version to support this CUDA version";
                  type = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                clangMinMajorVersion = {
                  _type = "option";
                  description = "Minimum Clang major version to support this CUDA version";
                  type = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                gccMaxMajorVersion = {
                  _type = "option";
                  description = "Maximum GCC major version to support this CUDA version";
                  type = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
                gccMinMajorVersion = {
                  _type = "option";
                  description = "Minimum GCC major version to support this CUDA version";
                  type = {
                    _type = "option-type";
                    check = <function>;
                    deprecationMessage = null;
                    description = "A version number with a major component.";
                    descriptionClass = "noun";
                    emptyValue = { };
                    functor = {
                      binOp = <function>;
                      name = "strMatching \"^([[:digit:]]+)\\$\"";
                      payload = null;
                      type = null;
                      wrapped = null;
                    };
                    getSubModules = null;
                    getSubOptions = <function>;
                    merge = <function>;
                    name = "majorVersion";
                    nestedTypes = { };
                    substSubModules = <function>;
                    typeMerge = <function>;
                  };
                };
              };
            }
          ];
          shorthandOnlyDefinesConfig = false;
          specialArgs = { };
        };
        type = <function, args: {class?, description?, modules, shorthandOnlyDefinesConfig?, specialArgs?}>;
        wrapped = null;
      };
      getSubModules = [
        {
          options = {
            clangMaxMajorVersion = {
              _type = "option";
              description = "Maximum Clang major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            clangMinMajorVersion = {
              _type = "option";
              description = "Minimum Clang major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            gccMaxMajorVersion = {
              _type = "option";
              description = "Maximum GCC major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            gccMinMajorVersion = {
              _type = "option";
              description = "Minimum GCC major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
          };
        }
      ];
      getSubOptions = <function>;
      merge = <function>;
      name = "submodule";
      nestedTypes = { };
      substSubModules = <function>;
      typeMerge = <function>;
    };
  };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/nvcc-compatibilities

generic.types.nvccCompatibility

Compatibility of an instance of NVCC with GCC and Clang

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "submodule";
  descriptionClass = null;
  emptyValue = {
    value = { };
  };
  functor = {
    binOp = <function>;
    name = "submodule";
    payload = {
      class = null;
      description = null;
      modules = [
        {
          options = {
            clangMaxMajorVersion = {
              _type = "option";
              description = "Maximum Clang major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            clangMinMajorVersion = {
              _type = "option";
              description = "Minimum Clang major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            gccMaxMajorVersion = {
              _type = "option";
              description = "Maximum GCC major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
            gccMinMajorVersion = {
              _type = "option";
              description = "Minimum GCC major version to support this CUDA version";
              type = {
                _type = "option-type";
                check = <function>;
                deprecationMessage = null;
                description = "A version number with a major component.";
                descriptionClass = "noun";
                emptyValue = { };
                functor = {
                  binOp = <function>;
                  name = "strMatching \"^([[:digit:]]+)\\$\"";
                  payload = null;
                  type = null;
                  wrapped = null;
                };
                getSubModules = null;
                getSubOptions = <function>;
                merge = <function>;
                name = "majorVersion";
                nestedTypes = { };
                substSubModules = <function>;
                typeMerge = <function>;
              };
            };
          };
        }
      ];
      shorthandOnlyDefinesConfig = false;
      specialArgs = { };
    };
    type = <function, args: {class?, description?, modules, shorthandOnlyDefinesConfig?, specialArgs?}>;
    wrapped = null;
  };
  getSubModules = [
    {
      options = {
        clangMaxMajorVersion = {
          _type = "option";
          description = "Maximum Clang major version to support this CUDA version";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        clangMinMajorVersion = {
          _type = "option";
          description = "Minimum Clang major version to support this CUDA version";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        gccMaxMajorVersion = {
          _type = "option";
          description = "Maximum GCC major version to support this CUDA version";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
        gccMinMajorVersion = {
          _type = "option";
          description = "Minimum GCC major version to support this CUDA version";
          type = {
            _type = "option-type";
            check = <function>;
            deprecationMessage = null;
            description = "A version number with a major component.";
            descriptionClass = "noun";
            emptyValue = { };
            functor = {
              binOp = <function>;
              name = "strMatching \"^([[:digit:]]+)\\$\"";
              payload = null;
              type = null;
              wrapped = null;
            };
            getSubModules = null;
            getSubOptions = <function>;
            merge = <function>;
            name = "majorVersion";
            nestedTypes = { };
            substSubModules = <function>;
            typeMerge = <function>;
          };
        };
      };
    }
  ];
  getSubOptions = <function>;
  merge = <function>;
  name = "submodule";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/data/nvcc-compatibilities

generic.types.virtualCudaArch

A virtual CUDA architecture name.

Type: optionType

Default:

{
  _type = "option-type";
  check = <function>;
  deprecationMessage = null;
  description = "A virtual CUDA architecture name.";
  descriptionClass = "noun";
  emptyValue = { };
  functor = {
    binOp = <function>;
    name = "strMatching \"^compute_[[:digit:]]+[a-z]?\\$\"";
    payload = null;
    type = null;
    wrapped = null;
  };
  getSubModules = null;
  getSubOptions = <function>;
  merge = <function>;
  name = "virtualCudaArch";
  nestedTypes = { };
  substSubModules = <function>;
  typeMerge = <function>;
}

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/types.nix

generic.utils

A set of utility functions used in building cudaPackages.

Type: anything

Default: { }

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic

generic.utils.mkVersionedPackageName

A function to generate a versioned package name

Type: function that evaluates to a(n) function that evaluates to a(n) string

Default: <function>

Declared by: - /nix/store/qbcb5y4gv6lw4mr9dap8f1whh7y0dw6h-source/cuda-modules/generic/utils.nix