Thanks for your sharing. When I compile pytorch project using JIT or Setuptools (e.g., python3 pytorch/setup.py install), I have a error as follows:

/home/chenxingyu/Documents/NN-CUDA-Example/pytorch/add2_ops.cpp:20:14: error: expected constructor, destructor, or type conversion before ‘(’ token
 TORCH_LIBRARY(add2, m) {

Could you help me solve it?

Traceback (most recent call last): File "tensorflow/time.py", line 57, in cuda_module = tf.load_op_library('build/libadd2.so') File "/home/guowei/anaconda3/envs/gy_py3.7_tf2.3/lib/python3.7/site-packages/tensorflow/python/framework/load_library.py", line 58, in load_op_library lib_handle = py_tf.TF_LoadLibrary(library_filename) tensorflow.python.framework.errors_impl.NotFoundError: build/libadd2.so: undefined symbol: _ZTIN10tensorflow8OpKernelE

I have changed the train.py here into this. And I meet this error : RuntimeError: CUDA error: an illegal memory access was encountered

(My environment is OK...using normal train.py can work)

Would you provide any suggestion? Thank you!!!

This problem occurs when I compile

running install running bdist_egg running egg_info writing add2.egg-info\PKG-INFO writing dependency_links to add2.egg-info\dependency_links.txt writing top-level names to add2.egg-info\top_level.txt reading manifest file 'add2.egg-info\SOURCES.txt' writing manifest file 'add2.egg-info\SOURCES.txt' installing library code to build\bdist.win-amd64\egg running install_lib running build_ext D:\miniconda3\envs\torch-gpu\lib\site-packages\torch\utils\cpp_extension.py:304: UserWarning: Error checking compiler version for cl: [WinError 2] 系统找不到指定的文件。 warnings.warn(f'Error checking compiler version for {compiler}: {error}') building 'add2' extension Emitting ninja build file D:\git-bash\daima\NN-CUDA-Example\build\temp.win-amd64-3.8\Release\build.ninja... Compiling objects... Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N) 1.10.2.git.kitware.jobserver-1 C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x64\link.exe /nologo /INCREMENTAL:NO /LTCG /DLL /MANIFEST:EMBED,ID=2 /MANIFESTUAC:NO /LIBPATH:D:\miniconda3\envs\torch-gpu\lib\site- packages\torch\lib "/LIBPATH:C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\lib/x64" /LIBPATH:D:\miniconda3\envs\torch-gpu\libs /LIBPATH:D:\miniconda3\envs\torch-gpu\PCbuild\amd64 "/LIBPATH:C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\lib\x64" "/LIBPATH:C:\Program Files (x86)\Windows Kits\10\lib\10.0.10240.0\ucrt\x64" "/LIBPATH:C:\Program Files (x86)\Windows Kits\8.1\lib\winv6.3\um\x64" c10.lib torch.lib tor ch_cpu.lib torch_python.lib cudart.lib c10_cuda.lib torch_cuda_cu.lib torch_cuda_cpp.lib /EXPORT:PyInit_add2 D:\git-bash\daima\NN-CUDA-Example\build\temp.win-amd64-3.8\Release\pytorch/add2_ops.obj D:\git-bash\daima\NN-CUDA-Example\b uild\temp.win-amd64-3.8\Release\kernel/add2_kernel.obj /OUT:build\lib.win-amd64-3.8\add2.cp38-win_amd64.pyd /IMPLIB:D:\git-bash\daima\NN-CUDA-Example\build\temp.win-amd64-3.8\Release\pytorch\add2.cp38-win_amd64.lib LINK : fatal error LNK1181: 无法打开输入文件“D:\git-bash\daima\NN-CUDA-Example\build\temp.win-amd64-3.8\Release\pytorch\add2_ops.obj” error: command 'C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.16.27023\bin\HostX86\x64\link.exe' failed with exit status 1181

The environment configuration is as follows： pytorch1.8.1+cuda11.1 python3.8

How to solve this problem？

The environment is competely same....But when I am using pytorch with jit method, below error appears:

subprocess.CalledProcessError: Command '['where', 'cl']' returned non-zero exit status 1.

Any suggestion? Thank you!!!!

Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed. Compiler: /usr/local/cuda/bin/nvcc Build flags: Id flags: -v

The output was: No such file or directory

Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed. Compiler: /usr/local/cuda/bin/nvcc Build flags: Id flags: -v

The output was: No such file or directory

Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed. Compiler: /usr/local/cuda/bin/nvcc Build flags: Id flags: -v

The output was: No such file or directory

Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed. Compiler: /usr/local/cuda/bin/nvcc Build flags: Id flags: -v

The output was: No such file or directory

Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed. Compiler: /usr/local/cuda/bin/nvcc Build flags: Id flags: -v

The output was: No such file or directory

Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed. Compiler: /usr/local/cuda/bin/nvcc Build flags: Id flags: -v

The output was: No such file or directory

Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed. Compiler: /usr/local/cuda/bin/nvcc Build flags: Id flags: -v

The output was: No such file or directory

Compiling the CUDA compiler identification source file "CMakeCUDACompilerId.cu" failed. Compiler: /usr/local/cuda/bin/nvcc Build flags: Id flags: -v

The output was: No such file or directory

Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Checking whether the CUDA compiler is NVIDIA using "" did not match "nvcc: NVIDIA (R) Cuda compiler driver": Checking whether the CUDA compiler is Clang using "" did not match "(clang version)": Performing C++ SOURCE FILE Test CMAKE_HAVE_LIBC_PTHREAD failed with the following output: Change Dir: /data/jupyter/cuda_learn/NN-CUDA-Example/build/CMakeFiles/CMakeTmp

Run Build Command(s):/usr/bin/make -f Makefile cmTC_6d839/fast && /usr/bin/make -f CMakeFiles/cmTC_6d839.dir/build.make CMakeFiles/cmTC_6d839.dir/build make[1]: Entering directory '/data/jupyter/cuda_learn/NN-CUDA-Example/build/CMakeFiles/CMakeTmp' Building CXX object CMakeFiles/cmTC_6d839.dir/src.cxx.o /usr/bin/c++ -DCMAKE_HAVE_LIBC_PTHREAD -o CMakeFiles/cmTC_6d839.dir/src.cxx.o -c /data/jupyter/cuda_learn/NN-CUDA-Example/build/CMakeFiles/CMakeTmp/src.cxx Linking CXX executable cmTC_6d839 /home/huyang02/anaconda3/lib/python3.6/site-packages/cmake/data/bin/cmake -E cmake_link_script CMakeFiles/cmTC_6d839.dir/link.txt --verbose=1 /usr/bin/c++ -rdynamic CMakeFiles/cmTC_6d839.dir/src.cxx.o -o cmTC_6d839 CMakeFiles/cmTC_6d839.dir/src.cxx.o: In function main': src.cxx:(.text+0x3e): undefined reference topthread_create' src.cxx:(.text+0x4a): undefined reference to pthread_detach' src.cxx:(.text+0x56): undefined reference topthread_cancel' src.cxx:(.text+0x67): undefined reference to pthread_join' src.cxx:(.text+0x7b): undefined reference topthread_atfork' collect2: error: ld returned 1 exit status CMakeFiles/cmTC_6d839.dir/build.make:98: recipe for target 'cmTC_6d839' failed make[1]: *** [cmTC_6d839] Error 1 make[1]: Leaving directory '/data/jupyter/cuda_learn/NN-CUDA-Example/build/CMakeFiles/CMakeTmp' Makefile:127: recipe for target 'cmTC_6d839/fast' failed make: *** [cmTC_6d839/fast] Error 2

Source file was: #include <pthread.h>

static void* test_func(void* data) { return data; }

int main(void) { pthread_t thread; pthread_create(&thread, NULL, test_func, NULL); pthread_detach(thread); pthread_cancel(thread); pthread_join(thread, NULL); pthread_atfork(NULL, NULL, NULL); pthread_exit(NULL);

return 0; }

Determining if the function pthread_create exists in the pthreads failed with the following output: Change Dir: /data/jupyter/cuda_learn/NN-CUDA-Example/build/CMakeFiles/CMakeTmp

Run Build Command(s):/usr/bin/make -f Makefile cmTC_308c5/fast && /usr/bin/make -f CMakeFiles/cmTC_308c5.dir/build.make CMakeFiles/cmTC_308c5.dir/build make[1]: Entering directory '/data/jupyter/cuda_learn/NN-CUDA-Example/build/CMakeFiles/CMakeTmp' Building CXX object CMakeFiles/cmTC_308c5.dir/CheckFunctionExists.cxx.o /usr/bin/c++ -DCHECK_FUNCTION_EXISTS=pthread_create -o CMakeFiles/cmTC_308c5.dir/CheckFunctionExists.cxx.o -c /data/jupyter/cuda_learn/NN-CUDA-Example/build/CMakeFiles/CheckLibraryExists/CheckFunctionExists.cxx Linking CXX executable cmTC_308c5 /home/huyang02/anaconda3/lib/python3.6/site-packages/cmake/data/bin/cmake -E cmake_link_script CMakeFiles/cmTC_308c5.dir/link.txt --verbose=1 /usr/bin/c++ -DCHECK_FUNCTION_EXISTS=pthread_create -rdynamic CMakeFiles/cmTC_308c5.dir/CheckFunctionExists.cxx.o -o cmTC_308c5 -lpthreads /usr/bin/ld: cannot find -lpthreads collect2: error: ld returned 1 exit status CMakeFiles/cmTC_308c5.dir/build.make:98: recipe for target 'cmTC_308c5' failed make[1]: *** [cmTC_308c5] Error 1 make[1]: Leaving directory '/data/jupyter/cuda_learn/NN-CUDA-Example/build/CMakeFiles/CMakeTmp' Makefile:127: recipe for target 'cmTC_308c5/fast' failed make: *** [cmTC_308c5/fast] Error 2

我在Jupiter notebook里运行下面jit命令的编译报错如下 python3 time.py --compiler jit

Using /tmp/torch_extensions as PyTorch extensions root... Detected CUDA files, patching ldflags Emitting ninja build file /tmp/torch_extensions/add2/build.ninja... Building extension module add2... [1/2] c++ -MMD -MF add2_ops.o.d -DTORCH_EXTENSION_NAME=add2 -DTORCH_API_INCLUDE_EXTENSION_H -I/gby/NN-CUDA-Example-master/pytorch/include -isystem /usr/local/lib/python3.6/dist-packages/torch/include -isystem /usr/local/lib/python3.6/dist-packages/torch/include/torch/csrc/api/include -isystem /usr/local/lib/python3.6/dist-packages/torch/include/TH -isystem /usr/local/lib/python3.6/dist-packages/torch/include/THC -isystem /usr/local/cuda/include -isystem /usr/include/python3.6m -D_GLIBCXX_USE_CXX11_ABI=0 -fPIC -std=c++11 -c /gby/NN-CUDA-Example-master/pytorch/add2_ops.cpp -o add2_ops.o FAILED: add2_ops.o c++ -MMD -MF add2_ops.o.d -DTORCH_EXTENSION_NAME=add2 -DTORCH_API_INCLUDE_EXTENSION_H -I/gby/NN-CUDA-Example-master/pytorch/include -isystem /usr/local/lib/python3.6/dist-packages/torch/include -isystem /usr/local/lib/python3.6/dist-packages/torch/include/torch/csrc/api/include -isystem /usr/local/lib/python3.6/dist-packages/torch/include/TH -isystem /usr/local/lib/python3.6/dist-packages/torch/include/THC -isystem /usr/local/cuda/include -isystem /usr/include/python3.6m -D_GLIBCXX_USE_CXX11_ABI=0 -fPIC -std=c++11 -c /gby/NN-CUDA-Example-master/pytorch/add2_ops.cpp -o add2_ops.o /gby/NN-CUDA-Example-master/pytorch/add2_ops.cpp:2:10: fatal error: add2.h: No such file or directory #include "add2.h" ^~~~~~~~ compilation terminated. ninja: build stopped: subcommand failed. Traceback (most recent call last): File "/usr/local/lib/python3.6/dist-packages/torch/utils/cpp_extension.py", line 960, in _build_extension_module check=True) File "/usr/lib/python3.6/subprocess.py", line 438, in run output=stdout, stderr=stderr) subprocess.CalledProcessError: Command '['ninja', '-v']' returned non-zero exit status 1.

During handling of the above exception, another exception occurred:

Traceback (most recent call last): File "time.py", line 56, in verbose=True) File "/usr/local/lib/python3.6/dist-packages/torch/utils/cpp_extension.py", line 658, in load is_python_module) File "/usr/local/lib/python3.6/dist-packages/torch/utils/cpp_extension.py", line 827, in _jit_compile with_cuda=with_cuda) File "/usr/local/lib/python3.6/dist-packages/torch/utils/cpp_extension.py", line 880, in _write_ninja_file_and_build _build_extension_module(name, build_directory, verbose) File "/usr/local/lib/python3.6/dist-packages/torch/utils/cpp_extension.py", line 973, in _build_extension_module raise RuntimeError(message) RuntimeError: Error building extension 'add2'

运行下面cmake命令的编译报错如下 Traceback (most recent call last): File "time.py", line 60, in torch.ops.load_library("build/libadd2.so") File "/usr/local/lib/python3.6/dist-packages/torch/_ops.py", line 106, in load_library ctypes.CDLL(path) File "/usr/lib/python3.6/ctypes/init.py", line 348, in init self._handle = _dlopen(self._name, mode) OSError: /gby/NN-CUDA-Example-master/pytorch/build/libadd2.so: cannot open shared object file: No such file or directory

Traceback (most recent call last): File "time.py", line 60, in torch.ops.load_library("build/libadd2.so") File "/home/gzy/anaconda3/envs/pytorch1.7/lib/python3.6/site-packages/torch/_ops.py", line 105, in load_library ctypes.CDLL(path) File "/home/gzy/anaconda3/envs/pytorch1.7/lib/python3.6/ctypes/init.py", line 348, in init self._handle = _dlopen(self._name, mode) OSError: /home/gzy/NN-CUDA-Example/pytorch/build/libadd2.so: undefined symbol: THPVariableClass