Skip to content

Unable to build Tensorflow 2.11.0 from source code #58953

@kevnzhao

Description

@kevnzhao
Click to expand!

Issue Type

Build/Install

Source

source

Tensorflow Version

tf 2.11.0

Custom Code

No

OS Platform and Distribution

Amazon Linux 2

Mobile device

No response

Python version

3.7

Bazel version

5.3.0

GCC/Compiler version

7.3.1

CUDA/cuDNN version

11.8.0/8.6.0

GPU model and memory

NVIDIA

Current Behaviour?

Met below compile error when running command `bazel --max_idle_secs=60 build --config=opt --config=cuda --action_env="LD_LIBRARY_PATH=${LD_LIBRARY_PATH}" --verbose_failures //tensorflow/tools/pip_package:build_pip_package`.


[29,045 / 33,182] Compiling tensorflow/core/kernels/pad_op_gpu.cu.cc; 235s local ... (72 actions, 71 running)
[30,249 / 33,182] Compiling tensorflow/core/kernels/pad_op_gpu.cu.cc; 435s local ... (72 actions, 71 running)
ERROR: /codebuild/output/src557356419/src/github.com/tensorflow/tensorflow/tensorflow/compiler/mlir/quantization/tensorflow/python/BUILD:13:11: Compiling tensorflow/compiler/mlir/quantization/tensorflow/python/quantize_model.cc failed: (Exit 1): crosstool_wrapper_driver_is_not_gcc failed: error executing command 
  (cd /root/.cache/bazel/_bazel_release/d8567c683bdb8c91b8e55044ae41426f/execroot/org_tensorflow && \
  exec env - \
    CUDA_TOOLKIT_PATH=/usr/local/cuda-11.8 \
    CUDNN_INSTALL_PATH=/usr/local/cuda \
    GCC_HOST_COMPILER_PATH=/usr/bin/gcc \
    LD_LIBRARY_PATH=/usr/local/cuda/lib64:/lib:/usr/local/cuda/lib64/stubs: \
    PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/codebuild/user/bin:/usr/local/cuda/bin \
    PWD=/proc/self/cwd \
    PYTHON_BIN_PATH=/usr/bin/python3 \
    PYTHON_LIB_PATH=/usr/local/lib64/python3.7/site-packages \
    TF2_BEHAVIOR=1 \
    TF_CUDA_COMPUTE_CAPABILITIES=3.7,5.2,7.0 \
    TF_CUDA_VERSION=11.8 \
    TF_CUDNN_VERSION=8.6 \
  external/local_config_cuda/crosstool/clang/bin/crosstool_wrapper_driver_is_not_gcc -MD -MF ...
  ......
# Configuration: ba96b98541fcc31cd39f04dd19ffaddc2a35225a697570c451b030de703660d8
# Execution platform: @local_execution_config_platform//:platform
tensorflow/compiler/mlir/quantization/tensorflow/python/quantize_model.cc: In function 'absl::lts_20220623::StatusOr<std::unique_ptr<llvm::raw_fd_ostream> > tensorflow::quantization::internal::{anonymous}::CreateMlirDumpFile(absl::lts_20220623::string_view)':
tensorflow/compiler/mlir/quantization/tensorflow/python/quantize_model.cc:140:10: error: could not convert 'dump_file' from 'std::unique_ptr<llvm::raw_fd_ostream>' to 'absl::lts_20220623::StatusOr<std::unique_ptr<llvm::raw_fd_ostream> >'
   return dump_file;
          ^~~~~~~~~
At global scope:
cc1plus: warning: unrecognized command line option '-Wno-array-parameter'
cc1plus: warning: unrecognized command line option '-Wno-unknown-warning'
Target //tensorflow/tools/pip_package:build_pip_package failed to build
INFO: Elapsed time: 1659.532s, Critical Path: 608.28s
INFO: 30987 processes: 10197 internal, 20790 local.
FAILED: Build did NOT complete successfully
FAILED: Build did NOT complete successfully


### Standalone code to reproduce the issue

```shell
bazel --max_idle_secs=60 build --config=opt --config=cuda --action_env="LD_LIBRARY_PATH=${LD_LIBRARY_PATH}" --verbose_failures //tensorflow/tools/pip_package:build_pip_package

Relevant log output

No response

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions