Automatic GPUs: A reproducible R / Python approach to getting up and running quickly on GCloud with GPUs in Tensorflow

RandPythonCondaGPU.txt

### Automatic GPUs:
### A reproducible R / Python approach to getting up and running quickly on GCloud with GPUs in Tensorflow
### https://medium.com/@zecca/automatic-gpus-46aa08f01886

# Check for CUDA and try to install.
if ! dpkg-query -W cuda; then

    # Start Timer here
	START=$(date +%s) # Time script

	# Install Cuda from NVIDIA
	curl -O http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/cuda-repo-ubuntu1604_8.0.61-1_amd64.deb
	# Login as root
	sudo dpkg -i ./cuda-repo-ubuntu1604_8.0.61-1_amd64.deb
	rm -r cuda-repo-ubuntu1604_8.0.61-1_amd64.deb
	sudo apt-get update
	sudo apt-get -y install cuda-8.0

	# Install cuDNN v6.0
	CUDNN_TAR_FILE="cudnn-8.0-linux-x64-v6.0.tgz"
	wget http://developer.download.nvidia.com/compute/redist/cudnn/v6.0/${CUDNN_TAR_FILE}
	tar -xzvf ${CUDNN_TAR_FILE}
	rm -r cudnn-8.0-linux-x64-v6.0.tgz
	sudo cp -P cuda/include/cudnn.h /usr/local/cuda/include
	sudo cp -P cuda/lib64/libcudnn* /usr/local/cuda/lib64/
	sudo chmod a+r /usr/local/cuda/lib64/libcudnn*

	# Export Paths
	echo 'export CUDA_HOME=/usr/local/cuda' >> ~/.bashrc
	echo 'export PATH=$PATH:$CUDA_HOME/bin' >> ~/.bashrc
	echo 'export LD_LIBRARY_PATH=$CUDA_HOME/lib64' >> ~/.bashrc
	echo 'export PATH=$PATH:$HOME/anaconda3/bin' >> ~/.bashrc
	source ~/.bashrc

	# Install Anaconda
	mkdir Downloads
	cd Downloads
	wget "https://repo.continuum.io/archive/Anaconda3-5.0.1-Linux-x86_64.sh" -O "Anaconda3-5.0.1-Linux-x86_64.sh"
	chmod +x Anaconda3-5.0.1-Linux-x86_64.sh
	sudo sh "Anaconda3-5.0.1-Linux-x86_64.sh" -b
	cd $HOME
	rm -r Downloads

	# Create conda environment to work with Python/R
	# conda search python
	# conda search r
	mkdir prog_env
	# conda create --prefix=$HOME/prog_env python=3.6 -y
	source activate prog_env
	sudo apt-get update
	conda install -c anaconda --prefix=$HOME/prog_env python=3.6 -y
	conda install -c anaconda tensorflow-gpu --prefix=$HOME/prog_env -y
	conda install -c anaconda --prefix=$HOME/prog_env r=3.4 -y
	source deactivate prog_env
	
	# Shows Cuda Info
	nvidia-smi
	
	# End of timer
	END=$(date +%s)
	DIFF=$(( $END - $START ))
	echo "It took $DIFF seconds"

fi


### Manually Check Python and R
source activate prog_env

(prog_env)$ python

	# Test GPU is working
	from tensorflow.python.client import device_lib
	device_lib.list_local_devices()

		# name: "/device:CPU:0"
		# device_type: "CPU"
		# memory_limit: 268435456

		# name: "/device:GPU:0"
		# device_type: "GPU"
		# memory_limit: 11326131405
		
	quit("yes") # CTRL(or CMD) + Z 
		
(prog_env)$ R

	# install.packages("tensorflow")
	library(tensorflow)
	install_tensorflow(version = "gpu")
	use_condaenv("r-tensorflow")
	sess = tf$Session()
		# 2018-03-21 19:32:33.987396: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] 
		# Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: Tesla K80, pci bus id: 0000:00:04.0, compute capability: 3.7)
		
	GPUtest <- tf$constant('GPU is running!')
	sess$run(GPUtest)
		# "GPU is running!"
	
	quit("yes") # CTRL(or CMD) + Z