rdht_init.sh

cd
CONDA_PATH="miniconda2"

wget https://gist.githubusercontent.com/yuchsiao/e2f9c1481976b4d5c524/raw/870969f072c461ad47da76e5822bdd3b807a9829/.screenrc
sudo yum -y install screen
screen

# Download and install miniconda
wget https://repo.continuum.io/miniconda/Miniconda-latest-Linux-x86_64.sh
chmod u+x Miniconda-latest-Linux-x86_64.sh
./Miniconda-latest-Linux-x86_64.sh -b -p $CONDA_PATH  # -b for batch no prompt

# Add miniconda to $PATH
echo '
# added by Miniconda2 3.19.0 installer
export PATH="/home/ec2-user/'$CONDA_PATH'/bin:$PATH"
' >> .bashrc
source .bashrc

conda install numpy -y
conda install scipy -y
conda install scikit-learn -y

# Install GPU driver
sudo /etc/grub.conf  # Append rdblacklist=nouveau at the end of kernel lines
sudo /etc/modprobe.d/blacklist.conf  # Append blacklist=nouveau
sudo reboot
sudo yum install gcc -y
sudo yum install gcc-c++ -y
sudo yum groupinstall -y "Development tools"
sudo yum install kernel-devel-`uname -r`
wget http://us.download.nvidia.com/XFree86/Linux-x86_64/340.46/NVIDIA-Linux-x86_64-340.46.run
chmod u+x NVIDIA-Linux-x86_64-340.46.run
sudo ./NVIDIA-Linux-x86_64-340.46.run

# Install CUDA
wget http://developer.download.nvidia.com/compute/cuda/7.5/Prod/local_installers/cuda_7.5.18_linux.run
chmod u+x cuda_7.5.18_linux.run
sudo ./cuda_7.5.18_linux.run
# move include and lib64 under /usr/local/cuda
sudo chmod 755 /usr/local/cuda-7.5


# Install JDK
wget --no-cookies --no-check-certificate --header "Cookie: gpw_e24=http%3A%2F%2Fwww.oracle.com%2F; oraclelicense=accept-securebackup-cookie" "http://download.oracle.com/otn-pub/java/jdk/8u72-b15/jdk-8u72-linux-x64.tar.gz"
tar xzf jdk-8u72-linux-x64.tar.gz
sudo mv jdk1.8.0_72 /opt
sudo alternatives --install /usr/bin/java java /opt/jdk1.8.0_72/bin/java 2
sudo alternatives --config java
echo 'export JAVA_HOME=/opt/jdk1.8.0_72
export JRE_HOME=/opt/jdk1.8.0_72/jre
export PATH="/opt/jdk1.8.0_72/bin:/opt/jdk1.8.0_72/jre/bin:$PATH"
' | tee -a $HOME/.bashrc 
source $HOME/.bashrc
echo "export JAVA_HOME=/opt/jdk1.8.0_72
export JRE_HOME=/opt/jdk1.8.0_72/jre
" | sudo tee -a /etc/environment 

# Install glibcxx, glibc
## Check versions: strings /usr/lib64/libstdc++.so.6 | grep GLIBC
wget ftp://ftp.gwdg.de/pub/misc/gcc/releases/gcc-5.3.0/gcc-5.3.0.tar.gz
tar -xzf  gcc-5.3.0.tar.gz 
cd gcc-5.3.0
./contrib/download_prerequisites
mkdir objdir
cd objdir
../configure --prefix=$HOME/gcc-5.3.0 --disable-multilib
make
make install
sudo mv /usr/lib64/libstdc++.so.6 /usr/lib64/libstdc++.so.6.bak
sudo cp $HOME/gcc-5.3.0/lib64/libstdc++.so.6 /usr/lib64/libstdc++.so.6
sudo cp $HOME/gcc-5.3.0/lib64/libstdc++.so.6.0.21 /usr/lib64/libstdc++.so.6.0.21
sudo chmod 755 /usr/lib64/libstdc++.so.6
sudo chmod 755 /usr/lib64/libstdc++.so.6.0.21

##
wget http://ftp.gnu.org/gnu/glibc/glibc-2.21.tar.gz
tar -xzf glibc-2.21.tar.gz
cd glibc-2.21
mkdir glibc-build
cd glibc-build
../configure --prefix='/usr'
make
sudo make install

##
GLIBC_VERSION=2.20
wget http://ftp.gnu.org/gnu/glibc/glibc-$GLIBC_VERSION.tar.gz
tar -xzf glibc-$GLIBC_VERSION.tar.gz
cd glibc-$GLIBC_VERSION
mkdir glibc-build
cd glibc-build
../configure --prefix='/usr'
make
sudo make install


# Install Bezel
## CAVEAT: cannot use non-default path for g++
## Need to dig in the code to change the path
#wget https://github.com/bazelbuild/bazel/releases/download/0.2.0/bazel-0.2.0-installer-linux-x86_64.sh
#chmod u+x bazel-0.2.0-installer-linux-x86_64.sh
#./bazel-0.2.0-installer-linux-x86_64.sh --user


# Install tensorflow
## GPU support
sudo pip install --upgrade https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow-0.7.1-cp27-none-linux_x86_64.whl
## CPU only
sudo pip install --upgrade https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.7.1-cp27-none-linux_x86_64.whl