Convert Dropout to new style Function. #1914
Closed
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Edward Z. Yang <[email protected]>
Member
|
This looks good, but does dropout actually need to be an autograd Function? Can it just be implemented as normal Python function? Roughly: def dropout(input, p=0.5, train=False, inplace=False):
noise = Variable(make_noise(...))
return input * noise |
Contributor
Author
|
Yeah, it does seem like a normal Python function would be good enough. |
jjsjann123
added a commit
to jjsjann123/pytorch
that referenced
this pull request
Aug 30, 2022
1. Mostly mechanical changes to refactor some of KernelArgumentHolder in our stack instead of direct use of at::Tensor/IValue: Note: we are still holding a ref counted at::Tensor within kernel arg holder for tensor entries, simply because we want to forward it in case of aliased output. This is quite unsatisfying. But to properly strip framework Tensor from codegen stack, we need quite some refactor to abstract away the ownership of memory and allocator. That's for some future PRs. 2. Separate compilation from execution of kernels, currently using FusionExecutorCache::compileFusion and FusionExecutorCache::runFusionWithInputs. Note that the compilation API is still experimental. We currently kick off compilation into a separate thread. This part would need to be exposed & integrated into our python API. TODO for follow up PRs: - trivial forwarding input to outputs - infer outputs should switch from meta tensor to fake tensor in order to preserve device - segmented fusion should/could be compiled in parallel, since we can infer outputs without a compiled kernel. - inputs_id_lookup should be refactored into KernelArgumentHolder, since we currently use args for passing inputs around. - index mode currently is per fusion. which is not neccesary and could be refactored into per segmented fuion instead. - bind kernel inputs should also try to bind cpu scalar with int type, since the runtime value can also be used in shape inference. Generally speaking, cpu scalar dtype should also be checked during validation. - high water mark could be refactored into using occupancy API after compilation, so we are not unnecessarily recompile when we don't have to.
pytorchmergebot
pushed a commit
that referenced
this pull request
Sep 8, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
pytorchmergebot
pushed a commit
that referenced
this pull request
Sep 8, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 8, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 8, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 9, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 9, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 9, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 9, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 9, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser ghstack-source-id: d0d88cf Pull Request resolved: #84626
jjsjann123
added a commit
that referenced
this pull request
Sep 19, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 19, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 19, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 19, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 19, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser ghstack-source-id: 0443a2a Pull Request resolved: #84626
jjsjann123
added a commit
that referenced
this pull request
Sep 21, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 21, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) [ghstack-poisoned]
jjsjann123
added a commit
that referenced
this pull request
Sep 21, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser ghstack-source-id: 34c0b92 Pull Request resolved: #84626
pytorchmergebot
pushed a commit
that referenced
this pull request
Sep 23, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) Pull Request resolved: #84626 Approved by: https://github.com/malfet
mehtanirav
pushed a commit
that referenced
this pull request
Oct 4, 2022
Syncing nvfuser devel branch to upstream master. https://github.com/csarofeen/pytorch/ Codegen changes include: - codegen improvement: i. improved view support on pointwise and transpose scheduler ii. grouped grid welford added for better outer-norm grid persistence in normalization - misc: i. new composite ops added: variance_mean , arange, ii. fixes misaligned address for transpose scheduler iii. refactor on separation of compilation API from execution API to prepare us for async compilation iv. double type support on expression evaluator v. PYTORCH_NVFUSER_DUMP refactor to save PTX and CUBIN Commits that's in this PR from the devel branch: ``` 89330aa Tensor factories must set the output shape as its input (#1939) b2fd01e arange support (#1933) 56c00fd Double support on all expression evaluators (#1937) 371f282 Improve trivial reduction merge support (#1931) 1d0c267 Test `rand` in a fusion with zero tensor input (#1932) 0dab160 Fix softmax bwd sizes. (#1890) ef98f36 Fix a bug (#1936) 63132a0 Propagate permissive mapping information into indexing pass (#1929) b4ac2c8 Map IterationDomains through view operations. (#1919) c0a187a do not use deprecated functions (#1935) 88de85e Upstream cherry pick fixes 0811 (#1934) b247dcf Separate kernel compilation API from kernel execution API (#1914) b34e3b9 Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924) 14a53e6 Nullary RNGOp (#1892) 3c3c89e Misc fixes/tuning for transpose scheduler (#1912) 20cf109 Grouped grid welford (#1921) 6cf7eb0 Transpose scheduler small dim sizes better support (#1910) 9341ea9 Disabled ViewPersistentShmoo sizes that results in NAN (#1922) 057237f Fix CUDA driver error: misaligned address for transpose scheduler (#1918) 3fb3d80 Add variance_mean function using Welford (#1907) 98febf6 Remove DisableOption::UnrollWithRng (#1913) ee8ef33 Minor fix for the debug interface of using PTX directly (#1917) 6e8f953 Add PYTORCH_NVFUSER_DUMP options to save PTX and CUBIN (#1916) 5eefa9a dopt is only available since nvrtc 11.7 (#1915) 2ec8fc7 Kill computeAtBetween (#1911) d0d106a Improve view support on pointwise and transpose scheduler (#1906) e71e1ec Fix name clash of RNG with shared memory (#1904) 3381793 Fix mutator and sameAs for expanded IterDomain (#1902) ``` RUN_TORCHBENCH: nvfuser Differential Revision: [D39324552](https://our.internmc.facebook.com/intern/diff/D39324552) Pull Request resolved: #84626 Approved by: https://github.com/malfet
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Signed-off-by: Edward Z. Yang [email protected]