36 lines
2.2 KiB
Python
36 lines
2.2 KiB
Python
import os
|
|
from setuptools import setup, find_packages
|
|
from torch_musa.utils.simple_porting import SimplePorting
|
|
from torch_musa.utils.musa_extension import MUSAExtension
|
|
|
|
SimplePorting(cuda_dir_path="./csrc", mapping_rule={
|
|
"x.device().is_cuda()": "true",
|
|
"#include <ATen/cuda/CUDAContext.h>": "#include \"torch_musa/csrc/aten/musa/MUSAContext.h\"",
|
|
"#include <c10/cuda/CUDAGuard.h>": "#include \"torch_musa/csrc/core/MUSAGuard.h\"",
|
|
"#include <ATen/cuda/Exceptions.h>": "#include \"torch_musa/csrc/core/MUSAException.h\"",
|
|
"#include <c10/cuda/CUDAStream.h>": "#include \"torch_musa/csrc/core/MUSAStream.h\"",
|
|
"at::kCUDA": "at::musa::kMUSA",
|
|
"at::cuda::getCurrentCUDAStream()": "at::musa::getCurrentMUSAStream()",
|
|
"__nv_bfloat16": "__mt_bfloat16",
|
|
"at::cuda::OptionalCUDAGuard": "at::musa::OptionalMUSAGuard",
|
|
"at::cuda::getCurrentCUDABlasHandle()": "at::musa::getCurrentMUSABlasHandle()",
|
|
"ATen/cuda/CUDATensorMethods.cuh": "ATen/musa/MUSA_PORT_TensorMethods.muh",
|
|
"#include \"attention_generic.cuh\"": "#include \"attention_generic.muh\"",
|
|
"#include \"reduction_utils.cuh\"": "#include \"reduction_utils.muh\"",
|
|
"#include <THC/THCAtomics.cuh>": "#include <THC/THCAtomics.muh>",
|
|
"#include \"dtype_float16.cuh\"": "#include \"dtype_float16.muh\"",
|
|
"#include \"dtype_float32.cuh\"": "#include \"dtype_float32.muh\"",
|
|
"#include \"custom_all_reduce.cuh\"": "#include \"custom_all_reduce.muh\"",
|
|
"#include \"dtype_bfloat16.cuh\"": "#include \"dtype_bfloat16.muh\"",
|
|
"#include \"dtype_fp8.cuh\"": "#include \"dtype_fp8.muh\"",
|
|
"#include \"attention_utils.cuh\"": "#include \"attention_utils.muh\"",
|
|
"cuPointerGetAttribute": "muPointerGetAttribute",
|
|
"CUdeviceptr": "MUdeviceptr",
|
|
"CUDA_SUCCESS": "MUSA_SUCCESS",
|
|
"CU_POINTER_ATTRIBUTE_RANGE_START_ADDR": "MU_POINTER_ATTRIBUTE_RANGE_START_ADDR",
|
|
"c10::cuda": "c10::musa",
|
|
"cudaStreamIsCapturing": "at::musa::musaStreamIsCapturing",
|
|
"AT_CUDA_CHECK": "C10_MUSA_CHECK",
|
|
"nv_bfloat16": "mt_bfloat16",
|
|
"struct __align__(16) RankData { const void *__restrict__ ptrs[8]; };":"struct __align__(16) RankData { const void *__restrict__ ptrs[8]; RankData& operator=(const RankData& ){return *this;} };"
|
|
}).run() |