TensorFlow 2.11 gpu, tensorRT, no AVX build tf_ops_n_z.cc fail

kjl

12/16/23, 3:29 AM

I'm trying to build TensorFlow 2.11 with gpu but no avx in Ubuntu 20.04

cuda version is 11.8
cuDNN version is 8.7
tensorRT version is 8.5
gcc version is 9.4.0

I'm using basilisk to run bazel version 5.3.0

The build stops with the following error:

bazelisk --output_base=~/install/tensorflow/tftmp build --config=cuda --local_ram_resources=4096 //tensorflow/tools/pip_package:build_pip_package
<br>...omitted lines...<br>
'build' options: --action_env PYTHON_BIN_PATH=/home/io/anaconda3/bin/python3 --action_env PYTHON_LIB_PATH=/home/io/anaconda3/lib/python3.8/site-packages --python_path=/home/io/anaconda3/bin/python3 --config=tensorrt --action_env CUDA_TOOLKIT_PATH=/usr/local/cuda-11.8 --action_env TF_CUDA_COMPUTE_CAPABILITIES=6.1 --action_env GCC_HOST_COMPILER_PATH=/usr/bin/x86_64-linux-gnu-gcc-9 --config=cuda
<br>...omitted lines...<br>
ERROR: /home/io/install/tensorflow/tensorflow/compiler/mlir/lite/BUILD:291:11: 
Compiling tensorflow/compiler/mlir/lite/ir/**tfl_ops.cc failed: (Exit 1):** **crosstool_wrapper_driver_is_not_gcc failed: error executing command** external/local_config_cuda/crosstool/clang/bin/**crosstool_wrapper_driver_is_not_gcc** -MD -MF 
bazel-out/k8-opt/bin/tensorflow/compiler/mlir/lite/_objs/tensorflow_lite/**tfl_ops.pic.d** ... (remaining 307 arguments skipped)<br>
x86_64-linux-gnu-gcc-9: fatal error: Killed signal terminated program cc1plus
compilation terminated.<br>
Target //tensorflow/tools/pip_package:build_pip_package failed to build<br>
Use --verbose_failures to see the command lines of failed build steps.<br>
INFO: Elapsed time: 717.825s, Critical Path: 660.88s<br>
INFO: 86 processes: 9 internal, 77 local.<br>
FAILED: Build did NOT complete successfully<br>

Any help would be greatly appreciated.

343

0 + 0

compiling

gcc

tensorflow

20.04

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: TensorFlow 2.11 gpu, tensorRT, no AVX build tf_ops_n_z.cc fail

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.