modify setup.py

setup.py

@@ -3,45 +3,19 @@ import subprocess
 
 import torch
 from setuptools import setup, find_packages
-from torch.utils.cpp_extension import BuildExtension, CUDAExtension, CUDA_HOME
+from torch.utils.cpp_extension import BuildExtension, CUDAExtension, ROCM_HOME
 
 # ninja build does not work unless include_dirs are abs path
 this_dir = os.path.dirname(os.path.abspath(__file__))
 
 
-def get_cuda_bare_metal_version(cuda_dir):
-    raw_output = subprocess.check_output([cuda_dir + "/bin/nvcc", "-V"], universal_newlines=True)
-    output = raw_output.split()
-    release_idx = output.index("release") + 1
-    release = output[release_idx].split(".")
-    bare_metal_major = release[0]
-    bare_metal_minor = release[1][0]
-
-    return raw_output, bare_metal_major, bare_metal_minor
-
-
 def check_cuda_torch_binary_vs_bare_metal(cuda_dir):
-    raw_output, bare_metal_major, bare_metal_minor = get_cuda_bare_metal_version(cuda_dir)
     torch_binary_major = torch.version.cuda.split(".")[0]
     torch_binary_minor = torch.version.cuda.split(".")[1]
 
     print("\nCompiling cuda extensions with")
-    print(raw_output + "from " + cuda_dir + "/bin\n")
-
-    if (bare_metal_major != torch_binary_major) or (bare_metal_minor != torch_binary_minor):
-        raise RuntimeError(
-            "Cuda extensions are being compiled with a version of Cuda that does " +
-            "not match the version used to compile Pytorch binaries.  " +
-            "Pytorch binaries were compiled with Cuda {}.\n".format(torch.version.cuda) +
-            "In some cases, a minor-version mismatch will not cause later errors:  " +
-            "https://github.com/NVIDIA/apex/pull/323#discussion_r287021798.  "
-            "You can try commenting out this check (at your own risk).")
-
 
 def append_nvcc_threads(nvcc_extra_args):
-    _, bare_metal_major, bare_metal_minor = get_cuda_bare_metal_version(CUDA_HOME)
-    if int(bare_metal_major) >= 11 and int(bare_metal_minor) >= 2:
-        return nvcc_extra_args + ["--threads", "4"]
     return nvcc_extra_args
 
 
@@ -57,12 +31,6 @@ if not torch.cuda.is_available():
         'and, if the CUDA version is >= 11.0, Ampere (compute capability 8.0).\n'
         'If you wish to cross-compile for a single specific architecture,\n'
         'export TORCH_CUDA_ARCH_LIST="compute capability" before running setup.py.\n')
-    if os.environ.get("TORCH_CUDA_ARCH_LIST", None) is None:
-        _, bare_metal_major, _ = get_cuda_bare_metal_version(CUDA_HOME)
-        if int(bare_metal_major) == 11:
-            os.environ["TORCH_CUDA_ARCH_LIST"] = "6.0;6.1;6.2;7.0;7.5;8.0"
-        else:
-            os.environ["TORCH_CUDA_ARCH_LIST"] = "6.0;6.1;6.2;7.0;7.5"
 
 print("\n\ntorch.__version__  = {}\n\n".format(torch.__version__))
 TORCH_MAJOR = int(torch.__version__.split('.')[0])
@@ -82,12 +50,12 @@ ext_modules = []
 # https://github.com/pytorch/pytorch/commit/eb7b39e02f7d75c26d8a795ea8c7fd911334da7e#diff-4632522f237f1e4e728cb824300403ac
 version_dependent_macros = ['-DVERSION_GE_1_1', '-DVERSION_GE_1_3', '-DVERSION_GE_1_5']
 
-if CUDA_HOME is None:
+if ROCM_HOME is None:
     raise RuntimeError(
         "Are you sure your environment has nvcc available?  If you're installing within a container from https://hub.docker.com/r/pytorch/pytorch, only images whose names contain 'devel' will provide nvcc."
     )
 else:
-    # check_cuda_torch_binary_vs_bare_metal(CUDA_HOME)
+    # check_cuda_torch_binary_vs_bare_metal(ROCM_HOME)
 
     def cuda_ext_helper(name, sources, extra_cuda_flags):
         return CUDAExtension(
@@ -96,11 +64,12 @@ else:
                 os.path.join('fastfold/model/fastnn/kernel/cuda_native/csrc', path) for path in sources
             ],
             include_dirs=[
-                os.path.join(this_dir, 'fastfold/model/fastnn/kernel/cuda_native/csrc/include')
+                os.path.join(this_dir, 'fastfold/model/fastnn/kernel/cuda_native/csrc/include'),
+                os.path.join(this_dir, 'fastfold/model/fastnn/kernel/cuda_native/csrc/'),
             ],
             extra_compile_args={
                 'cxx': ['-O3'] + version_dependent_macros,
-                'nvcc':
+                'hipcc':
                     append_nvcc_threads(['-O3', '--use_fast_math'] + version_dependent_macros +
                                         extra_cuda_flags)
             })
@@ -108,10 +77,6 @@ else:
 
 
     cc_flag = ['-gencode', 'arch=compute_70,code=sm_70']
-    _, bare_metal_major, _ = get_cuda_bare_metal_version(CUDA_HOME)
-    if int(bare_metal_major) >= 11:
-        cc_flag.append('-gencode')
-        cc_flag.append('arch=compute_80,code=sm_80')
 
     extra_cuda_flags = [
         '-std=c++14', '-maxrregcount=50', '-U__CUDA_NO_HALF_OPERATORS__',
@@ -140,5 +105,5 @@ setup(
     ext_modules=ext_modules,
     package_data={'fastfold': ['model/fastnn/kernel/cuda_native/csrc/*']},
     cmdclass={'build_ext': BuildExtension} if ext_modules else {},
-    install_requires=['einops', 'colossalai'],
+    #install_requires=['einops', 'colossalai'],
 )