tensorflow: Tensorflow 2.12 bazel failed!Cuda11.2+cuDNN8.1+Tensorrt7.2 under Ubuntu2004

Click to expand!

Issue Type

Build/Install

Have you reproduced the bug with TF nightly?

No

Source

source

Tensorflow Version

tf2.12

Custom Code

Yes

OS Platform and Distribution

Ubuntu2004

Mobile device

No response

Python version

3.8

Bazel version

5.3

GCC/Compiler version

9.4

CUDA/cuDNN version

CUDA11.2/cuDNN8.1

GPU model and memory

No response

##Current Behaviour?

TensorRT Version: 7.2.3.4 CUDA Version: 11.2 CUDNN Version: 8.1.1.33-1 Operating System: Ubuntu-20.04

https://github.com/tensorflow/tensorflow/blob/master/.bazelrc

Bazel build latest 2.12 source code failed:

##Standalone code to reproduce the issue

# Configuration: 605fee2c8aa1b68d8dcb7abeac1c0e77048ffd56710c9ebb1fee2729853263d3
# Execution platform: @local_execution_config_platform//:platform
tensorflow/compiler/xla/stream_executor/cuda/cuda_graph.cc: In function ‘tsl::StatusOr<stream_executor::gpu::OwnedCudaGraph> stream_executor::gpu::CaptureCudaGraph(stream_executor::Stream*, absl::lts_20220623::AnyInvocable<tsl::Status()>, cudaStreamCaptureMode)’:
tensorflow/compiler/xla/stream_executor/cuda/cuda_graph.cc:135:21: **error**: ‘cudaGraphDebugDotFlagsVerbose’ was not declared in this scope
  135 |         int flags = cudaGraphDebugDotFlagsVerbose;
      |                     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
tensorflow/compiler/xla/stream_executor/cuda/cuda_graph.cc:136:24: **error**: ‘cudaGraphDebugDotPrint’ was not declared in this scope
  136 |         if (auto err = cudaGraphDebugDotPrint(graph, file.c_str(), flags);
      |                        ^~~~~~~~~~~~~~~~~~~~~~
Target //tensorflow/tools/pip_package:build_pip_package failed to build
INFO: Elapsed time: 11358.490s, Critical Path: 270.65s
INFO: 26270 processes: 9762 internal, 16508 local.
FAILED: Build did NOT complete successfully

Relevant log output

No response

About this issue

  • Original URL
  • State: closed
  • Created a year ago
  • Reactions: 1
  • Comments: 17 (9 by maintainers)

Most upvoted comments

I succeeded building TensorFlow 2.12 in tensorflow/build:2.12-python3.8 docker.