Comments (5)
q10
commented on June 12, 2024
Could you show the full logs, and let me know the nvcc, gcc, and cmake versions you are using? We don't have official support for building FBGEMM_GPU on WSL2, but if WSL2 works exactly the same as a regular Linux environment, then it may be just an issue with the nvcc and/or gcc versions.
from fbgemm.
shadow150519
commented on June 12, 2024
Here is the full logs.
nvcc version: Cuda compilation tools, release 12.3, V12.3.52
gcc version: gcc (Ubuntu 9.4.0-1ubuntu1~20.04.2) 9.4.0
cmake version: cmake version 3.25.3
OS: wsl2 Ubuntu20.4 on Windows10
wsl kernel version: 5.15.133.1-1
wsl version: 2.0.9.0
C:\Windows\system32\wsl.exe --distribution Ubuntu-20.04 --exec /bin/bash -c "export CMAKE_COLOR_DIAGNOSTICS=ON && export CLION_IDE=TRUE && export JETBRAINS_IDE=TRUE && cd /home/wtx/workspace/cpp_project/FBGEMM/fbgemm_gpu/cmake-build-fbgemm-gpu-debug && /usr/bin/cmake -DCMAKE_BUILD_TYPE=Debug -DCMAKE_C_COMPILER=/usr/bin/gcc '-DCMAKE_CXX_COMPILER=/usr/bin/g++' '-DCMAKE_PREFIX_PATH=D:\workspace\cpp_project\libtorch' -G 'CodeBlocks - Unix Makefiles' -S /home/wtx/workspace/cpp_project/FBGEMM/fbgemm_gpu -B /home/wtx/workspace/cpp_project/FBGEMM/fbgemm_gpu/cmake-build-fbgemm-gpu-debug"
================================================================================
Building the CUDA variant of FBGEMM-GPU
================================================================================
================================================================================
Default C++ compiler flags
(values may be overridden by CMAKE_CXX_STANDARD and CXX_STANDARD):
================================================================================
-- The CXX compiler identification is GNU 9.4.0
-- The C compiler identification is GNU 9.4.0
-- The CUDA compiler identification is NVIDIA 12.3.52
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/g++ - skipped
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working C compiler: /usr/bin/gcc - skipped
-- Detecting C compile features
-- Detecting C compile features - done
-- Detecting CUDA compiler ABI info
-- Detecting CUDA compiler ABI info - done
-- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc - skipped
-- Detecting CUDA compile features
-- Detecting CUDA compile features - done
-- Performing Test C_HAS_AVX_1
-- Performing Test C_HAS_AVX_1 - Failed
-- Performing Test C_HAS_AVX_2
-- Performing Test C_HAS_AVX_2 - Success
-- Performing Test C_HAS_AVX2_1
-- Performing Test C_HAS_AVX2_1 - Failed
-- Performing Test C_HAS_AVX2_2
-- Performing Test C_HAS_AVX2_2 - Success
-- Performing Test C_HAS_AVX512_1
-- Performing Test C_HAS_AVX512_1 - Failed
-- Performing Test C_HAS_AVX512_2
-- Performing Test C_HAS_AVX512_2 - Failed
-- Performing Test C_HAS_AVX512_3
-- Performing Test C_HAS_AVX512_3 - Failed
-- Performing Test CXX_HAS_AVX_1
-- Performing Test CXX_HAS_AVX_1 - Failed
-- Performing Test CXX_HAS_AVX_2
-- Performing Test CXX_HAS_AVX_2 - Success
-- Performing Test CXX_HAS_AVX2_1
-- Performing Test CXX_HAS_AVX2_1 - Failed
-- Performing Test CXX_HAS_AVX2_2
-- Performing Test CXX_HAS_AVX2_2 - Success
-- Performing Test CXX_HAS_AVX512_1
-- Performing Test CXX_HAS_AVX512_1 - Failed
-- Performing Test CXX_HAS_AVX512_2
-- Performing Test CXX_HAS_AVX512_2 - Failed
-- Performing Test CXX_HAS_AVX512_3
-- Performing Test CXX_HAS_AVX512_3 - Failed
-- Skipping merge_pooled_embeddings sources
-- Configuring done
-- Generating done
-- Build files have been written to: /home/wtx/workspace/cpp_project/FBGEMM/fbgemm_gpu/cmake-build-fbgemm-gpu-debug
Cannot get compiler information:
Compiler exited with error code 1: \usr\local\cuda\bin\nvcc -Dfbgemm_gpu_py_EXPORTS -g --generate-code=arch=compute_75,code=[compute_75,sm_75] -Xcompiler=-fPIC -std=c++17 -mavx -mf16c -mfma -mavx2 -fopenmp --dryrun /mnt/c/Users/Lenovo/AppData/Local/Temp/compiler-file11626473163701549697.cu
nvcc fatal : Unknown option '-mavx'
from fbgemm.
q10
commented on June 12, 2024
Hmm, some of the tool versions you listed are not the recommended versions - in particular, CUDA should be at 12.1 and GCC should be at 10+.
I did quick Google search cases where users have run into a similar error signature before, and the solution was generally to upgrade the drivers and toolkit. That being said, we highly recommend using building and using FBGEMM in a full Linux environment instead, as WSL may contain intricacies that are beyond our scope for debugging issues like this.
from fbgemm.
q10
commented on June 12, 2024
@shadow150519 could you try latest main, and run the usual FBGEMM_GPU build command, but with -DWSL_MODE=1 added to the args?
python setup.py ...<usual args>... -DWSL_MODE=1
(See here for build instructions in detail)
Let us know how this goes.
from fbgemm.
shadow150519
commented on June 12, 2024
sorry for the late response, I've moved to a linux machine right now and it works correctly. I'll try this and give you the result as soon as I can. Thank for your help :)
@shadow150519 could you try latest
main, and run the usual FBGEMM_GPU build command, but with-DWSL_MODE=1added to the args?python setup.py ...<usual args>... -DWSL_MODE=1(See here for build instructions in detail)
Let us know how this goes.
from fbgemm.
Related Issues (20)
- Support MacOS? HOT 1
- `.to("meta")` is leaked to the public main branch in the tests. HOT 2
- Can't compile FBGEMM with GCC 12.3.0 HOT 4
- compiling FBGEMM for ARM HOT 4
- How `partition_avx512` is auto-tuned? HOT 3
- Compiling on windows with mingw
- Error importing fbgemm_gpu HOT 15
- Build failure on MacOS HOT 6
- momentum for SGD/Adagrad HOT 2
- RuntimeError: No such operator fbgemm::jagged_2d_to_dense HOT 3
- Having issue installing FBGEMM-gpu on MacOS HOT 8
- Latest FBGEMM doesn't build with latest PyTorch HOT 2
- fbgemm_gpu doesn't build for CPU because impl_abstract_pystub is not found HOT 1
- [FBGEMM_GPU Question] When should I use FusedEmbeddingBagCollection over EmbeddingBagCollection?
- [Question] What does device / embedding_specs.compute_device parameter in ctor of TBE mean? HOT 4
- AttributeError: '_OpNamespace' 'fbgemm' object has no attribute 'jagged_2d_to_dense' HOT 5
- Building error after C++20 HOT 3
- quantize_embeddings + KeyedJaggedTensor+ vbe cannot work
- AttributeError: '_OpNamespace' 'fbgemm' object has no attribute 'jagged_2d_to_dense' HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from fbgemm.