Not sure what category this belongs in (build??) and I see other references to the subject library but not in this regard. During build, some links are failing with a number of the following type of message:
/usr/bin/ld: /usr/local/cuda/lib64/libcublas.so: undefined reference to `cublasLtHHHMatmulAlgoInit@libcublasLt.so.11’
It seems that this library, introduced after CUDA 10.0, is not included in the link, e.g.,
/usr/bin/c++ -Wno-unused-variable -Wno-unused-result -Wno-deprecated-declarations -Wno-deprecated-copy -Wno-uninitialized -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -DEDGE_PROFILER_USE_KINETO -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Werror=cast-function-type -Wno-stringop-overflow -DHAVE_AVX512_CPU_DEFINITION -DHAVE_AVX2_CPU_DEFINITION -O3 -DNDEBUG -DNDEBUG -rdynamic test_cpp_c10d/CMakeFiles/ProcessGroupGlooAsyncTest.dir/ProcessGroupGlooAsyncTest.cpp.o -o bin/ProcessGroupGlooAsyncTest -L/opt/conda/lib -Wl,-rpath,/opt/pytorch/build/lib:/opt/conda/lib:/usr/local/cuda/lib64 lib/libtorch_cpu.so lib/libc10d_cuda_test.so lib/libgtest_main.a -lpthread lib/libtorch_cuda.so lib/libc10_cuda.so /usr/local/cuda/lib64/libcudart.so /opt/conda/lib/libnvToolsExt.so lib/libprotobuf.a lib/libc10.so -lmkl_intel_lp64 -lmkl_gnu_thread -lmkl_core -fopenmp /usr/lib/x86_64-linux-gnu/libpthread.so -lm /usr/lib/x86_64-linux-gnu/libdl.so lib/libdnnl.a -ldl -Wl,–no-as-needed,"/opt/pytorch/build/lib/libtorch_cpu.so" -Wl,–as-needed /usr/local/cuda/lib64/libcufft.so /usr/local/cuda/lib64/libcurand.so /usr/local/cuda/lib64/libcublas.so /usr/lib/x86_64-linux-gnu/libcudnn.so lib/libgtest.a -pthread
adding the library results in success
If there’s a proper way to do this, I’m all ears, otherwise, FYI.