tensorflow: Bazel build failed in compiling TensorFlow from source.

I am compiling TensorFlow from source on an Ubuntu machine. This is a verbose output of the error log generated in the bazel build //tensorflow/tools/pip_package:build_pip_package execution.

Error Log:

ERROR: /home/annanay/.cache/bazel/_bazel_root/9b62c7240de3d9136528cdded945b550/external/llvm/BUILD:418:5: Generating code from table: lib/Target/AArch64/AArch64.td @llvm//:aarch64_target_gen__gen_fast_isel_genrule failed: bash failed: 
error executing command 
(cd /home/annanay/.cache/bazel/_bazel_root/9b62c7240de3d9136528cdded945b550/execroot/tensorflow && \
exec env - \
PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/snap/bin \
/bin/bash -c 'source external/bazel_tools/tools/genrule/genrule-setup.sh; bazel-out/host/bin/external/llvm/llvm-tblgen -I external/llvm/include -I external/llvm/tools/clang/include -I $(dirname external/llvm/lib/Target/AArch64/AArch64.td) -gen-fast-isel external/llvm/lib/Target/AArch64/AArch64.td -o bazel-out/local-opt/genfiles/external/llvm/lib/Target/AArch64/AArch64GenFastISel.inc'): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 127.
bazel-out/host/bin/external/llvm/llvm-tblgen: relocation error: bazel-out/host/bin/external/llvm/llvm-tblgen: symbol _ZNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEE9_M_createERmm, version GLIBCXX_3.4.21 not defined in file libstdc++.so.6 with link time reference
Target //tensorflow/tools/pip_package:build_pip_package failed to build

I built TensorFlow from source, and here is the relevant information:

$ git rev-parse HEAD
4c3bb1aeb7bb46bea35036433742a720f39ce348

$ bazel version
Build label: 0.4.4
Build target: bazel-out/local-fastbuild/bin/src/main/java/com/google/devtools/build/lib/bazel/BazelServer_deploy.jar
Build time: Wed Feb 1 18:54:21 2017 (1485975261)
Build timestamp: 1485975261
Build timestamp as int: 1485975261

System:

x86_64 GNU/Linux

About this issue

  • Original URL
  • State: closed
  • Created 7 years ago
  • Comments: 17 (7 by maintainers)

Commits related to this issue

Most upvoted comments

@kirk86 u just need to change se to stream_executor, and the mpiops.cc just complies successfully. Looks like someone shortened the name of stream_executor as se, and unfortunately the se was deleted in newer code. check the file //tensorflow/stream_executor/lib/statusor.h

@ghostindex good catch mate. I confirm that indeed changing se to stream_executor did solve the problem, as you said someone apparently tried to use a shortened version but without linking that somehow to the original namespace.

Also pretty sure they are without MPI support.