[Caffe2 CPU tests] Update CMakeLists.txt

[kunalb]

I was trying to build PyTorch with USE_GLOG=ON (so we could get better timestamps around the nccl logging) and ran into this error

[1/7] Linking CXX executable bin/verify_api_visibility
FAILED: bin/verify_api_visibility
: && /opt/rh/gcc-toolset-11/root/usr/bin/c++ -D_GLIBCXX_USE_CXX11_ABI=1 -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DUSE_KINETO -DLIBKINETO_NOROCTRACER -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=range-loop-construct -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-strict-overflow -Wno-strict-aliasing -Wno-stringop-overflow -Wsuggest-override -Wno-psabi -Wno-error=pedantic -Wno-error=old-style-cast -Wno-missing-braces -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow -DHAVE_AVX512_CPU_DEFINITION -DHAVE_AVX2_CPU_DEFINITION -O2 -g -DNDEBUG -rdynamic     -Wl,--no-as-needed caffe2/CMakeFiles/verify_api_visibility.dir/__/aten/src/ATen/test/verify_api_visibility.cpp.o -o bin/verify_api_visibility -L/lib/intel64   -L/lib/intel64_win   -L/lib/win-x64 -Wl,-rpath,/lib/intel64:/lib/intel64_win:/lib/win-x64:/usr/local/cuda/lib64:/root/conda/lib:/mnt/code/pytorch/build/lib:  lib/libgtest_main.a  -Wl,--no-as-needed,"/mnt/code/pytorch/build/lib/libtorch.so" -Wl,--as-needed  -Wl,--no-as-needed,"/mnt/code/pytorch/build/lib/libtorch_cpu.so" -Wl,--as-needed  lib/libprotobuf.a  /root/conda/lib/libmkl_intel_lp64.so  /root/conda/lib/libmkl_gnu_thread.so  /root/conda/lib/libmkl_core.so  -fopenmp  /usr/lib64/libpthread.so  -lm  /usr/lib64/libdl.so  -Wl,--no-as-needed,"/mnt/code/pytorch/build/lib/libtorch_cuda.so" -Wl,--as-needed  lib/libc10_cuda.so  lib/libc10.so  /root/conda/lib/libglog.so.0.4.0  /root/conda/lib/libgflags.so.2.2.2  -lpthread  /usr/local/cuda/lib64/libcudart.so  /usr/local/cuda/lib64/libnvToolsExt.so  lib/libgtest.a  -pthread && /root/conda/bin/cmake -E __run_co_compile --lwyu="ldd;-u;-r" --source=bin/verify_api_visibility && :
/opt/rh/gcc-toolset-11/root/usr/bin/ld: /mnt/code/pytorch/build/lib/libtorch.so: undefined reference to symbol '_ZTVN10__cxxabiv117__class_type_infoE@@CXXABI_1.3'
/opt/rh/gcc-toolset-11/root/usr/bin/ld: /usr/lib64/libstdc++.so.6: error adding symbols: DSO missing from command line
collect2: error: ld returned 1 exit status

Adding stdc++ explicitly to the list of libraries to link seems to fix the build, and I was able to get a working build of PyTorch.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/119643

• 📄 Preview Python docs built from this PR
• 📄 Preview C++ docs built from this PR
• ❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

Note: Links to docs will display an error until the docs builds have been completed.

✅ No Failures

As of commit ee14203 with merge base c4d24b5 ():
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[kunalb]

@pytorchbot merge

[pytorchmergebot]

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

[pytorchmergebot]

Merge failed

Reason: 1 mandatory check(s) failed. The first few are:

• pull / win-vs2019-cpu-py3 / build

Dig deeper by viewing the failures on hud

Details for Dev Infra team

Raised by workflow job

Failing merge rule: Core Maintainers

[kunalb]

@pytorchmergebot merge --rebase

[pytorchmergebot]

@pytorchbot started a rebase job onto refs/remotes/origin/viable/strict. Check the current status here

[pytorchmergebot]

Successfully rebased kunalb-patch-1 onto refs/remotes/origin/viable/strict, please pull locally before adding more changes (for example, via git checkout kunalb-patch-1 && git pull --rebase)

[pytorchmergebot]

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

[pytorchmergebot]

Merge failed

Reason: 2 jobs have failed, first few of them are: trunk / win-vs2019-cuda11.8-py3 / build, trunk / win-vs2019-cpu-py3 / build

Details for Dev Infra team

Raised by workflow job

[malfet]

Added a guard against not-windows (stdc++ is a very linux library), but this change feels very suspicious to me as dependency to C++ runtime should be added by automatically for C++ projects, and libtorch is one of them

[cyyever]

I suspect that gcc-11 was used in an old centOS system so it has the path '/opt/rh/gcc-toolset-11/', and the system's default libstdc++ may have older ABI. May be the proper libstdc++ with /opt/rh/gcc-toolset-11/ should be installed and CMake should be able to detect it. (By using some kind of environment variable?)

PS: A quick Google search gave pkg names like "gcc-toolset-11-libstdc++-devel-11.2.1-9.1.el8.i686.rpm", I think this is an issue of incomplete toolchain setting, use

sudo yum install gcc-toolset-11-libstdc++-devel -y 

and re-run CMake may solve the problem.

[kunalb]

Hm, I was using a full toolchain installed into centos 8 (gcc-toolset-11-11.1-1.el8.x86_64), then run with scl enable gcc-toolset-11 to have everything pointed to the right location.

I'll double check

[malfet]

@kunalb I don't think there is a big harm in adding libstdc++ as a dependency, so feel free to land it (windows builds are passing now and runtime failure is unrelated), but again this dependency feels a bit superfluous though not wrong

[kunalb]

Thanks, I'll go ahead and land it in that case and then keep looking on the side.

@pytorchmergebot merge

[pytorchmergebot]

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

[pytorchmergebot]

Merge failed

Reason: 1 mandatory check(s) failed. The first few are:

• pull / win-vs2019-cpu-py3 / build

Dig deeper by viewing the failures on hud

Details for Dev Infra team

Raised by workflow job

Failing merge rule: Core Maintainers

[kunalb]

@pytorchbot merge

[pytorchmergebot]

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here