tensorflow2.4.0编译安装（适配CPU型号 flag）
1.GCC编译器版本7.3.1
https://www.jianshu.com/p/b3f96dde3f61
gcc7.3.1
sudo yum install centos-release-scl
sudo yum install devtoolset-7-gcc*
scl enable devtoolset-7 bash
which gcc
gcc --version
# scl enable devtoolset-7 bash # bash生效
# which gcc
/opt/rh/devtoolset-7/root/usr/bin/gcc
# gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/opt/rh/devtoolset-7/root/usr/libexec/gcc/x86_64-redhat-linux/7/lto-wrapper
Target: x86_64-redhat-linux
Configured with: ../configure --enable-bootstrap --enable-languages=c,c++,fortran,lto --prefix=/opt/rh/devtoolset-7/root/usr --mandir=/opt/rh/devtoolset-7/root/usr/share/man --infodir=/opt/rh/devtoolset-7/root/usr/share/info --with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-shared --enable-threads=posix --enable-checking=release --enable-multilib --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-unique-object --enable-linker-build-id --with-gcc-major-version-only --enable-plugin --with-linker-hash-style=gnu --enable-initfini-array --with-default-libstdcxx-abi=gcc4-compatible --with-isl=/builddir/build/BUILD/gcc-7.3.1-20180303/obj-x86_64-redhat-linux/isl-install --enable-libmpx --enable-gnu-indirect-function --with-tune=generic --with-arch_32=i686 --build=x86_64-redhat-linux
Thread model: posix
gcc version 7.3.1 20180303 (Red Hat 7.3.1-5) (GCC)
2.安装python3环境
安装tensorflow编译需要的依赖包
pip3 install numpy==1.19.4 -i http://pypi.doubanio.com/simple/ --trusted-host pypi.doubanio.com
pip3 install wheel -i http://pypi.doubanio.com/simple/ --trusted-host pypi.doubanio.com
pip3 install keras_preprocessing --no-deps -i http://pypi.doubanio.com/simple/ --trusted-host pypi.doubanio.com
pip3 install setuptools==54.2.0 -i http://pypi.doubanio.com/simple/ --trusted-host pypi.doubanio.com
3.tensorflow 源代码 2.4.0
从github上检出2.4.0分支 下载
git clone https://github.com/tensorflow/tensorflow.git
/opt/tflow/tensorflow-2.4.0/configure.py 包含tensorflow需要的bazel的版本信息
tensorflow-2.4.0
4. bazel 环境
yum install unzip
bazel 3.1.0
https://github.com/bazelbuild/bazel/releases?after=3.3.1
chmod +x bazel-3.1.0-installer-linux-x86_64.sh
./bazel-3.1.0-installer-linux-x86_64.sh
5.编译tensorflow
5.1 为安装TensorFlow配置环境
/opt/tflow/tensorflow-2.4.0
./configure
5.2 编译tensorflow
gcc5以下版本
bazel build --local_ram_resources=2048 -c opt --copt=-mssse3 --copt=-mcx16 --copt=-mavx --copt=-msse4.1 --copt=-msse4.2 --copt=-mpopcnt -k //tensorflow/tools/pip_package:build_pip_package
gcc5以上版本
bazel build --local_ram_resources=2048 --cxxopt="-D_GLIBCXX_USE_CXX11_ABI=0" -c opt --copt=-mssse3 --copt=-mcx16 --copt=-mavx --copt=-msse4.1 --copt=-msse4.2 --copt=-mpopcnt -k //tensorflow/tools/pip_package:build_pip_package > log 2>&1 &
6.构建软件包
------------------------------------------------------------------------------
[root@node01 tensorflow-2.4.0]# ls
ACKNOWLEDGMENTS bazel-tensorflow-2.4.0 configure ISSUE_TEMPLATE.md SECURITY.md
arm_compiler.BUILD bazel-testlogs configure.cmd LICENSE tensorflow
AUTHORS BUILD configure.py models.BUILD third_party
bazel-bin CODE_OF_CONDUCT.md CONTRIBUTING.md README.md tools
bazel-out CODEOWNERS ISSUES.md RELEASE.md WORKSPACE
[root@node01 tensorflow-2.4.0]# ./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg
Thu Apr 1 12:46:24 CST 2021 : === Preparing sources in dir: /tmp/tmp.wHCYLV3NHl
/usr/local/src/tensorflow-2.4.0 /usr/local/src/tensorflow-2.4.0
/usr/local/src/tensorflow-2.4.0
/usr/local/src/tensorflow-2.4.0/bazel-bin/tensorflow/tools/pip_package/build_pip_package.runfiles/org_tensorflow /usr/local/src/tensorflow-2.4.0
/usr/local/src/tensorflow-2.4.0
/tmp/tmp.wHCYLV3NHl/tensorflow/include /usr/local/src/tensorflow-2.4.0
/usr/local/src/tensorflow-2.4.0
Thu Apr 1 12:46:48 CST 2021 : === Building wheel
warning: no files found matching 'README'
warning: no files found matching '*.pyd' under directory '*'
warning: no files found matching '*.pyi' under directory '*'
warning: no files found matching '*.pd' under directory '*'
warning: no files found matching '*.dylib' under directory '*'
warning: no files found matching '*.dll' under directory '*'
warning: no files found matching '*.lib' under directory '*'
warning: no files found matching '*.csv' under directory '*'
warning: no files found matching '*.h' under directory 'tensorflow/include/tensorflow'
warning: no files found matching '*.proto' under directory 'tensorflow/include/tensorflow'
warning: no files found matching '*' under directory 'tensorflow/include/third_party'
Thu Apr 1 12:47:26 CST 2021 : === Output wheel file is in: /tmp/tensorflow_pkg
bazel build 命令会创建一个名为 build_pip_package 的可执行文件，此文件是用于构建 pip 软件包的程序。如下所示地运行该可执行文件，以在 /tmp/tensorflow_pkg 目录中构建 .whl 软件包。
./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg
安装软件包
生成的 .whl 文件的文件名取决于 TensorFlow 版本和您的平台。例如，使用 pip install 安装软件包：
pip3 install /tmp/tensorflow_pkg/tensorflow-version-tags.whl
pip3 install /tmp/tensorflow_pkg/NAME_OF_WHEEL.whl
参考链接
https://tech.amikelive.com/node-887/how-to-resolve-error-illegal-instruction-core-dumped-when-running-import-tensorflow-in-a-python-program/
https://tech.amikelive.com/node-882/how-to-build-and-install-the-latest-tensorflow-without-cuda-gpu-and-with-optimized-cpu-performance-on-ubuntu/
https://www.tensorflow.org/install/source
libstdc++.so.6.0.24 放到 /usr/lib64/目录下
[root@node04 ~]# cd /usr/lib64/
[root@node04 lib64]# ll /usr/lib64/libstdc++.so.6
lrwxrwxrwx. 1 root root 19 Mar 23 14:31 /usr/lib64/libstdc++.so.6 -> libstdc++.so.6.0.19
[root@node04 lib64]# mv libstdc++.so.6 libstdc++.so.6_bak
[root@node04 lib64]# ll libstdc++.so.6.0.24
-rwxr-xr-x 1 root root 11482848 Apr 1 14:13 libstdc++.so.6.0.24
[root@node04 lib64]# ln -s libstdc++.so.6.0.24 /usr/lib64/libstdc++.so.6
[root@node04 lib64]# ll /usr/lib64/libstdc++.so.6
lrwxrwxrwx 1 root root 19 Apr 1 14:15 /usr/lib64/libstdc++.so.6 -> libstdc++.so.6.0.24