TensorFlow doesnt build with debug mode

Question

We are trying to build a TensorFlow test case with debug flag:

bazel build -c dbg //tensorflow/python/kernel_tests:sparse_matmul_op_test

However the build is failing with below error:

/usr/include/features.h:330:4: error: #warning _FORTIFY_SOURCE requires compiling with optimization (-O) [-Werror=cpp]
warning _FORTIFY_SOURCE requires compiling with optimization (-O)

cc1: all warnings being treated as errors

Target //tensorflow/python/kernel_tests:sparse_matmul_op_test failed to build

We have tried below options to resolve this:

built by exporting export CFLAGS and CXXFLAGS to "-Wno-error"
bazel build -c dbg --cxxopt="-Wno-all" --cxxopt="-Wno-error" //tensorflow/python/kernel_tests:sparse_matmul_op_test
Tried commenting compiler_flag from third_party/gpus/crosstool/CROSSTOOL.tpl

What is the correct way to suppress these warnings for the build to proceed?

We are using gcc v5.4.0.

BernardoGO · Accepted Answer · 2016-11-10 05:40:45Z

2

I've had the same issue recently. It got solved by adding --copt=-O and -c opt to the build command.

Example:

bazel build --copt=-O -c dbg -c opt //tensorflow/python/kernel_tests:sparse_matmul_op_test

answered Nov 10, 2016 at 5:40

BernardoGO

1,8571 gold badge20 silver badges36 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Dmitry Mikushin · Accepted Answer · 2018-02-09 09:38:16Z

Solution by @BernardoGO does not work when CUDA build is enabled:

$ bazel build --copt=-O -c opt --config cuda -c dbg --strip=never //tensorflow/tools/pip_package:build_pip_package -s

/usr/include/c++/6/bits/stl_pair.h(327): error: calling a __host__ function("std::_Rb_tree_const_iterator< ::tensorflow::NcclManager::NcclStream *> ::_Rb_tree_const_iterator") from a __device__ function("std::pair< ::std::_Rb_tree_const_iterator< ::tensorflow::NcclManager::NcclStream *> , bool> ::pair< ::std::_Rb_tree_iterator< ::tensorflow::NcclManager::NcclStream *>  &, bool &, (bool)1> ") is not allowed

/usr/include/c++/6/bits/stl_pair.h(327): error: identifier "std::_Rb_tree_const_iterator< ::tensorflow::NcclManager::NcclStream *> ::_Rb_tree_const_iterator" is undefined in device code

/usr/include/c++/6/bits/stl_algobase.h(1009): error: calling a __host__ function("__builtin_clzl") from a __device__ function("std::__lg") is not allowed

3 errors detected in the compilation of "/tmp/tmpxft_00007abb_00000000-6_nccl_manager.cpp1.ii".

Works only if --copt=-O is replaced with --copt=-O1, but -O1 is too much for comfortable debugging.

Collectives™ on Stack Overflow

TensorFlow doesnt build with debug mode

2 Answers 2

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related