Tensorflow compilation fails with error "redeclared here as 'Eigen::internal::ComparisonName cmp'"

[kkarnatak]

Hi,

I am trying to compile the tensorflow source with opencl 'syscl' flag. I got the following error:

In file included from <command-line>:0:0:
./bazel-out/local_linux-py3-opt/bin/tensorflow/core/kernels/_objs/relu_op/tensorflow/core/kernels/relu_op.pic.sycl:8149:22: note: 'const bool cl::sycl::detail::{anonymous}::arg_used [16]' previously defined here
 const bool           kernel_info< ::Eigen::TensorSycl::ExecExprFunctorKernel<const ::Eigen::TensorAssignOp< ::Eigen::TensorMap< ::Eigen::Tensor<float, 1, 1, long>, 16, ::Eigen::MakePointer>, const ::Eigen::TensorCwiseBinaryOp< ::Eigen::internal::scalar_product_op<const float, const float>, const ::Eigen::TensorMap< ::Eigen::Tensor<const float, 1, 1, long>, 16, ::Eigen::MakePointer>, const ::Eigen::TensorConversionOp<float, const ::Eigen::TensorCwiseBinaryOp< ::Eigen::internal::scalar_cmp_op<const float, const float, ::Eigen::internal::ComparisonName::cmp_GT>, const ::Eigen::TensorMap< ::Eigen::Tensor<const float, 1, 1, long>, 16, ::Eigen::MakePointer>, const ::Eigen::TensorCwiseNullaryOp< ::Eigen::internal::scalar_constant_op<const float>, const ::Eigen::TensorMap< ::Eigen::Tensor<const float, 1, 1, long>, 16, ::Eigen::MakePointer> > > > > >, ::Eigen::TensorSycl::internal::FunctorExtractor< ::Eigen::TensorEvaluator<const ::Eigen::TensorAssignOp< ::Eigen::TensorMap< ::Eigen::Tensor<float, 1, 1, long>, 16, ::Eigen::MakePointer>, const ::Eigen::TensorCwiseBinaryOp< ::Eigen::internal::scalar_product_op<const float, const float>, const ::Eigen::TensorMap< ::Eigen::Tensor<const float, 1, 1, long>, 16, ::Eigen::MakePointer>, const ::Eigen::TensorConversionOp<float, const ::Eigen::TensorCwiseBinaryOp< ::Eigen::internal::scalar_cmp_op<const float, const float, ::Eigen::internal::ComparisonName::cmp_GT>, const ::Eigen::TensorMap< ::Eigen::Tensor<const float, 1, 1, long>, 16, ::Eigen::MakePointer>, const ::Eigen::TensorCwiseNullaryOp< ::Eigen::internal::scalar_constant_op<const float>, const ::Eigen::TensorMap< ::Eigen::Tensor<const float, 1, 1, long>, 16, ::Eigen::MakePointer> > > > > >, const ::Eigen::SyclDevice> >, ::utility::tuple::Tuple< ::cl::sycl::accessor<unsigned char, 1, ::cl::sycl::access::mode::read_write, ::cl::sycl::access::target::global_buffer>, ::cl::sycl::accessor<unsigned char, 1, ::cl::sycl::access::mode::read, ::cl::sycl::access::target::global_buffer>, ::cl::sycl::accessor<unsigned char, 1, ::cl::sycl::access::mode::read, ::cl::sycl::access::target::global_buffer>, ::cl::sycl::accessor<unsigned char, 1, ::cl::sycl::access::mode::read, ::cl::sycl::access::target::global_buffer> > > >::arg_used[] = {
                      ^
In file included from <command-line>:0:0:
./bazel-out/local_linux-py3-opt/bin/tensorflow/core/kernels/_objs/relu_op/tensorflow/core/kernels/relu_op.pic.sycl:8012:83: error: template parameter 'int cmp'
 template <typename LhsScalar, typename RhsScalar, Eigen::internal::ComparisonName cmp> struct scalar_cmp_op;
                                                                                   ^
In file included from external/eigen_archive/unsupported/Eigen/CXX11/../../../Eigen/Core:423:0,
                 from external/eigen_archive/unsupported/Eigen/CXX11/Tensor:14,
                 from ./third_party/eigen3/unsupported/Eigen/CXX11/Tensor:1,
                 from ./tensorflow/core/kernels/relu_op.h:23,
                 from tensorflow/core/kernels/relu_op.cc:20:
external/eigen_archive/unsupported/Eigen/CXX11/../../../Eigen/src/Core/functors/BinaryFunctors.h:190:77: error: redeclared here as 'Eigen::internal::ComparisonName cmp'
 template<typename LhsScalar, typename RhsScalar, ComparisonName cmp> struct scalar_cmp_op;
                                                                             ^
Target //tensorflow/tools/pip_package:build_pip_package failed to build
Use --verbose_failures to see the command lines of failed build steps.
INFO: Elapsed time: 4754.029s, Critical Path: 191.56s

I have used the following command to initiate compilation:

bazel build -c opt --config=sycl //tensorflow/tools/pip_package:build_pip_package

I have AMD radeon 7670M GPU. I am using ubuntu 14.04.5 LTS and have latest compute-cpp module. My AMD drivers are working correctly ( The sole reason why i am using 14.04.5 :) )

~/TensorFlow/Packages/tensorflow$ glxinfo | grep OpenGL
OpenGL vendor string: Advanced Micro Devices, Inc.
OpenGL renderer string: AMD Radeon HD 7600M Series
OpenGL core profile version string: 4.3.13416 Core Profile Context 15.302
OpenGL core profile shading language version string: 4.40
OpenGL core profile context flags: (none)
OpenGL core profile profile mask: core profile
OpenGL core profile extensions:
OpenGL version string: 4.5.13416 Compatibility Profile Context 15.302
OpenGL shading language version string: 4.40
OpenGL context flags: (none)
OpenGL profile mask: compatibility profile
OpenGL extensions:

$ bazel test //tensorflow/python/kernel_tests:basic_gpu_test    INFO: Found 1 test target...
Target //tensorflow/python/kernel_tests:basic_gpu_test up-to-date:
  bazel-bin/tensorflow/python/kernel_tests/basic_gpu_test
INFO: Elapsed time: 2.213s, Critical Path: 1.83s
//tensorflow/python/kernel_tests:basic_gpu_test                          PASSED in 1.8s

Executed 1 out of 1 test: 1 test passes.

Any help will be appreciated.

Thanks.

[lukeiwanski]

Can you try this branch: https://github.com/lukeiwanski/tensorflow/tree/dev/eigen_mehdi ?

make sure you have latest computecpp :)