mirror of
https://github.com/saymrwulf/pytorch.git
synced 2026-05-14 20:57:59 +00:00
Summary: In the [official documentation](https://pytorch.org/tutorials/advanced/torch_script_custom_ops.html), it is recommended to use `TORCH_LIBRARY` to register ops for TorchScript. However, that code is never tested with CUDA extension and is actually broken (https://github.com/pytorch/pytorch/issues/47493). This PR adds a test for it. It will not pass CI now, but it will pass when the issue https://github.com/pytorch/pytorch/issues/47493 is fixed. Pull Request resolved: https://github.com/pytorch/pytorch/pull/47524 Reviewed By: zou3519 Differential Revision: D24991839 Pulled By: ezyang fbshipit-source-id: 037196621c7ff9a6e7905efc1097ff97906a0b1c |
||
|---|---|---|
| .. | ||
| no_python_abi_suffix_test | ||
| self_compiler_include_dirs_test | ||
| torch_test_cpp_extension | ||
| cpp_c10d_extension.cpp | ||
| cpp_c10d_extension.hpp | ||
| cpp_frontend_extension.cpp | ||
| cuda_extension.cpp | ||
| cuda_extension.cu | ||
| cuda_extension_kernel.cu | ||
| cuda_extension_kernel2.cu | ||
| cudnn_extension.cpp | ||
| doubler.h | ||
| extension.cpp | ||
| jit_extension.cpp | ||
| jit_extension2.cpp | ||
| msnpu_extension.cpp | ||
| rng_extension.cpp | ||
| setup.py | ||
| torch_library.cu | ||