🚀 The feature, motivation and pitch

Hello,

i would like to ask when PyTorch will support the sm_90 CUDA capability. My request is motivated by the necessity of this compute capability to use H100 GPUs.

When i try to send a tensor to a H100 GPU, the following warning is given:

/home/schoerob/cuda_stuff/testnev/lib/python3.8/site-packages/torch/cuda/init.py:155: UserWarning:
NVIDIA H100 PCIe with CUDA capability sm_90 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_37 sm_50 sm_60 sm_70 sm_75 sm_80 sm_86.
If you want to use the NVIDIA H100 PCIe GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/

warnings.warn(incompatible_device_warn.format(device_name, capability, " ".join(arch_list), device_name))

Thank you in advance.

Alternatives

No response

Additional context

No response

cc @ezyang @seemethere @malfet @ngimel