search

gpgpu-sim / pytorch-gpgpu-sim

Modified version of PyTorch able to work with changes to GPGPU-Sim
Other
40 stars 23 forks source link

Installation Error libculibos.a(cuos_common_posix.o) #7

Closed sxzhang1993 closed 3 years ago

sxzhang1993 commented 3 years ago

Issue description

Hello I have problems when running setup.py. I have followed the installation guide on the pytorch-gpgpu-sim homepage.

Code example

[ 95%] Linking CXX executable ProcessGroupGlooTest [100%] Linking CXX executable ProcessGroupGlooAsyncTest /usr/bin/ld: /usr/local/cuda/lib64/libculibos.a(cuos_common_posix.o): undefined reference to symbol 'dlclose@@GLIBC_2.2.5' //lib/x86_64-linux-gnu/libdl.so.2: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status test/CMakeFiles/ProcessGroupGlooTest.dir/build.make:110: recipe for target 'test/ProcessGroupGlooTest' failed make[2]: [test/ProcessGroupGlooTest] Error 1 CMakeFiles/Makefile2:244: recipe for target 'test/CMakeFiles/ProcessGroupGlooTest.dir/all' failed make[1]: [test/CMakeFiles/ProcessGroupGlooTest.dir/all] Error 2 make[1]: Waiting for unfinished jobs.... /usr/bin/ld: /usr/local/cuda/lib64/libculibos.a(cuos_common_posix.o): undefined reference to symbol 'dlclose@@GLIBC_2.2.5' //lib/x86_64-linux-gnu/libdl.so.2: error adding symbols: DSO missing from command line collect2: error: ld returned 1 exit status test/CMakeFiles/ProcessGroupGlooAsyncTest.dir/build.make:110: recipe for target 'test/ProcessGroupGlooAsyncTest' failed make[2]: [test/ProcessGroupGlooAsyncTest] Error 1 CMakeFiles/Makefile2:282: recipe for target 'test/CMakeFiles/ProcessGroupGlooAsyncTest.dir/all' failed make[1]: [test/CMakeFiles/ProcessGroupGlooAsyncTest.dir/all] Error 2 Makefile:140: recipe for target 'all' failed make: [all] Error 2 Failed to run 'bash tools/build_pytorch_libs.sh --use-cuda --use-nnpack nccl caffe2 nanopb libshm gloo THD c10d'

System Info

Please copy and paste the output from our environment collection script (or fill out the checklist below manually).

CUDA 8.0/9.0/9.1 CUDNN 7.0/7.1 will all have this error.

Any help would be appreciated! Thanks so much!

baiyang100 commented 3 years ago

Brother, have you solved your problem?I'm running into the same issue, Ubuntu 18.04 install cuda8,cudnn7.1.3, if you fix it, can you share the solution?

IceCY commented 3 years ago

I comment out the following lines in torch/lib/c10d/test/CMakeLists.txt to bypass the compilation issue.

c10d_add_test(ProcessGroupGlooTest.cpp c10d c10d_cuda_test)
c10d_add_test(ProcessGroupGlooAsyncTest.cpp c10d c10d_cuda_test)

System: Ubuntu 18.04; CUDA 8; CUDNN 7.1.4

sxzhang1993 commented 3 years ago

Brother, have you solved your problem?I'm running into the same issue, Ubuntu 18.04 install cuda8,cudnn7.1.3, if you fix it, can you share the solution?

Hello, I have tried IceCY's method and it worked for me!

sxzhang1993 commented 3 years ago

I comment out the following lines in torch/lib/c10d/test/CMakeLists.txt to bypass the compilation issue.

c10d_add_test(ProcessGroupGlooTest.cpp c10d c10d_cuda_test)
c10d_add_test(ProcessGroupGlooAsyncTest.cpp c10d c10d_cuda_test)

System: Ubuntu 18.04; CUDA 8; CUDNN 7.1.4

Thanks a lot! Now I have it installed by following your solution.

baiyang100 commented 3 years ago

I comment out the following lines in torch/lib/c10d/test/CMakeLists.txt to bypass the compilation issue.

c10d_add_test(ProcessGroupGlooTest.cpp c10d c10d_cuda_test)
c10d_add_test(ProcessGroupGlooAsyncTest.cpp c10d c10d_cuda_test)

System: Ubuntu 18.04; CUDA 8; CUDNN 7.1.4

Thanks a lot! Now I have it installed by following your solution.But I ran into a new problem while running the mnists dataset.