Description of problem: Pytorch unusable on aarch64 machines. Version-Release number of selected component (if applicable): pytorch-2.0.1-3.an23.aarch64 How reproducible: ``` yum install -y pytorch python3 -c 'import torch' ``` Actual results: ``` Traceback (most recent call last): File "/usr/lib64/python3.10/site-packages/torch/__init__.py", line 168, in _load_global_deps ctypes.CDLL(lib_path, mode=ctypes.RTLD_GLOBAL) File "/usr/lib64/python3.10/ctypes/__init__.py", line 374, in __init__ self._handle = _dlopen(self._name, mode) OSError: libcurand.so.10: cannot open shared object file: No such file or directory During handling of the above exception, another exception occurred: Traceback (most recent call last): File "<string>", line 1, in <module> File "/usr/lib64/python3.10/site-packages/torch/__init__.py", line 228, in <module> _load_global_deps() File "/usr/lib64/python3.10/site-packages/torch/__init__.py", line 189, in _load_global_deps _preload_cuda_deps(lib_folder, lib_name) File "/usr/lib64/python3.10/site-packages/torch/__init__.py", line 154, in _preload_cuda_deps raise ValueError(f"{lib_name} not found in the system path {sys.path}") ValueError: libcublas.so.*[0-9] not found in the system path ['', '/usr/lib64/python310.zip', '/usr/lib64/python3.10', '/usr/lib64/python3.10/lib-dynload', '/usr/lib64/python3.10/site-packages', '/usr/lib/python3.10/site-packages'] ``` Expected results: Additional info: cuda's library path set to x86_64 ``` cat /etc/ld.so.conf.d/000_cuda.conf /usr/local/cuda/targets/x86_64-linux/lib ```
aarch64下的: /etc/ld.so.conf.d/988_cuda-12.conf /etc/ld.so.conf.d/000_cuda.conf 两个配置文件中的目录不对导致的问题。 已在cuda-12.1.1-5.an23版本修复