| Summary: | sci-libs/tensorflow-2.4.0_rc4 fails to compile with cuda flag | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Oscar <oscar.mueller> |
| Component: | Current packages | Assignee: | Jason Zaman <perfinion> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | jstein |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | AMD64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
cudnn 8.0.5.39 nvidia-cuda-toolkit 11.1.1 |
message: ERROR: sci-libs/tensorflow-2.4.0_rc4::gentoo failed (compile phase): ebazel failed Call stack: ebuild.sh, line 125: Called src_compile environment, line 4245: Called ebazel 'build' '//tensorflow:libtensorflow_framework.so' '//tensorflow:libtensorflow.so' environment, line 2456: Called die The specific snippet of code: "${@}" || die "ebazel failed" TF_CUDA_COMPUTE_CAPABILITIES=7.5 in /etc/portage/make.conf mv_ed cudnn-10.2-linux-x64-v7.6.5.32.tgz from NVIDIA to /var/cache/distfiles/ due to "fetch restriction" versions: dev-libs/cudnn Latest version available: 8.0.4.30 Latest version installed: 7.6.5.32-r2 dev-util/bazel Latest version installed: 3.2.0 dev-util/nvidia-cuda-sdk Latest version available: 10.2.89 Latest version installed: 10.2.89 dev-util/nvidia-cuda-toolkit Latest version available: 11.1.1 Latest version installed: 10.2.89-r2 Python 3.8 emerge --info '=sci-libs/tensorflow-2.4.0_rc4::gentoo' (this build works) sci-libs/tensorflow-2.4.0_rc4::gentoo was built with the following: USE="python xla -cuda -mpi" ABI_X86="(64)" CPU_FLAGS_X86="avx sse sse2 sse3 sse4_1 sse4_2 -avx2 -fma3 -fma4" PYTHON_TARGETS="python3_8 -python3_6 -python3_7" CFLAGS="-march=ivybridge -O3 -pipe -msse -msse2 -msse3 -msse4.1 -msse4.2 -mavx" CXXFLAGS="-march=ivybridge -O3 -pipe -msse -msse2 -msse3 -msse4.1 -msse4.2 -mavx" emerge -pqv '=sci-libs/tensorflow-2.4.0_rc4::gentoo' [ebuild R ] sci-libs/tensorflow-2.4.0_rc4 USE="cuda* python xla -mpi" CPU_FLAGS_X86="avx sse sse2 sse3 sse4_1 sse4_2 -avx2 -fma3 -fma4" PYTHON_TARGETS="python3_8 -python3_6 -python3_7"