我在SUSE Linux Enterprise Server 11 SP3上使用anaconda安装TensorFlow的过程记录。
准备安装包：
gcc48 glibc-2.18.tar.gz SLES-11-SP4-DVD-x86_64-GM-DVD1.iso tensorflow_install
ls gcc48/
gcc-4.8.2.tar.gz gmp-5.1.3.tar.gz mpc-1.0.1.tar.gz mpfr-3.1.2.tar.bz2
ls tensorflow_install/
Anaconda2-5.0.1-Linux-x86_64.sh mock-2.0.0-py27_0.tar.bz2 pbr-3.1.1-py27_0.tar.bz2 protobuf-3.5.1-py27_1.tar.bz2 tensorflow-1.4.0-py27_0.tar.bz2 tflearn-0.3.2-py27_0.tar.bz2 tldextract-1.6-py27_0.tar.bz2
开始安装过程，命令如下：
mount -t iso9660 -o loop SLES-11-SP4-DVD-x86_64-GM-DVD1.iso /media/
zypper ar /media/ suse11sp4iso
zypper in gcc
zypper in gcc-c++
cd gcc48/
tar xzf gcc-4.8.2.tar.gz
tar xzf mpc-1.0.1.tar.gz
tar xzf gmp-5.1.3.tar.gz
tar xjf mpfr-3.1.2.tar.bz2
cd gmp-5.1.3/
./configure
make -j20
make install
cd ..
cd mpfr-3.1.2/
./configure
make -j20
make install
cd ..
cd mpc-1.0.1/
./configure
make -j20
make install
cd ..
cd gcc-4.8.2/
mkdir build
cd build/
../configure –enable-checking=release –enable-languages=c,c++ –disable-multilib
export LD_LIBRARY_PATH=/usr/local/lib:/usr/local/lib64
make -j20
make install
cd ..
cd ..
tar xzf glibc-2.18.tar.gz
cd glibc-2.18/
mkdir build
cd build
../configure –prefix=/usr –disable-profile –enable-add-ons –with-headers=/usr/include –with-binutils=/usr/bin
make -j20
make install
echo “export LD_LIBRARY_PATH=/root/anaconda2/lib/:/usr/local/lib:/usr/local/lib64:$LD_LIBRARY_PATH” >> ~/.bashrc
cd ..
cd tensorflow_install/
bash Anaconda2-5.0.1-Linux-x86_64.sh
conda install –use-local tldextract-1.6-py27_0.tar.bz2
conda install –use-local mock-2.0.0-py27_0.tar.bz2
conda install –use-local pbr-3.1.1-py27_0.tar.bz2
conda install –use-local protobuf-3.5.1-py27_1.tar.bz2
conda install –use-local tensorflow-1.4.0-py27_0.tar.bz2
conda install –use-local tflearn-0.3.2-py27_0.tar.bz2
更多SUSE相关信息见SUSE 专题页面 https://www.linuxidc.com/topicnews.aspx?tid=3
