/configurate-deeplearning-ibm-ppc64le

Deep learning server configurate IBM POWER8 (ppc64le)

MIT LicenseMIT

Deep learning server configurate IBM POWER8 (ppc64le)

Bring more value to your organization’s data by developing with an entirely new approach to problems: deep learning, machine learning, and artificial intelligence. Unlock hidden potential and patterns in data organically — without you having to know the patterns, networks, or be an algorithmic expert. IBM is making deep learning easier and more performant for you with an enterprise software distribution of the most popular open-frameworks, in PowerAI.

In this tutorial we show a step to step for install and configure a deep learning server in IBM POWER8.

ibmcom/powerai

  • powerai

      >>> docker pull ibmcom/powerai
  >>> nvidia-docker run --name [name-pwai]  --ipc=host --env ACTIVATE=pytorch --env LICENSE=yes ibmcom/powerai:1.5.3-all-ubuntu16.04-py3 tail -f /dev/null

Nvidia Docker for POWER8

For docker installation

sudo echo 'deb http://ftp.unicamp.br/pub/ppc64el/ubuntu/16_04/docker-17.06.1-ce-ppc64el/ xenial main' >> /etc/apt/sources.list
sudo apt-get update && apt-get install docker-ce
sudo usermod -aG docker your-username

For Nvidia-docker installation

git clone https://github.com/NVIDIA/nvidia-docker.git
git checkout 1.0
git fetch --all
cd nvidia-docker
make deb
cd dist
sudo dpkg -i nvidia-docker_1.0.0~rc.3-1_ppc64el.deb

Pulling Nvidia Cuda 8.0 with CuDNN 6

nvidia-docker pull nvidia/cuda-ppc64le:8.0-cudnn6-devel-ubuntu16.04
#Creating a container
nvidia-docker run --name DL -v /home/projects:/data -d nvidia/cuda-ppc64le:8.0-cudnn6-devel-ubuntu16.04 tail -f /dev/null
#Accesing container bash
nvidia-docker exec -ti DL bash

Anaconda3 for POWER8

wget https://repo.continuum.io/archive/Anaconda3-5.0.0-Linux-ppc64le.sh

./Anaconda3-5.0.0-Linux-ppc64le.sh

Pytorch with Anaconda

# https://developer.ibm.com/code/howtos/install-pytorch-on-power    
# https://github.com/ppc64le/build-scripts/blob/master/pytorch/ubuntu_build_intructions.txt

export CMAKE_PREFIX_PATH="$(dirname $(which conda))/../"
conda install numpy pyyaml setuptools cmake cffi
git clone --recursive https://github.com/pytorch/pytorch
cd pytorch
python setup.py install

Pytorch vision

git clone https://github.com/pytorch/vision.git    
cd vision    
python setup.py install

OpenCV for Power 8

wget https://repo.continuum.io/pkgs/free/linux-ppc64le/opencv-3.1.0-np112py36_2.tar.bz2
mkdir opencv-3.1.0-py36
cd opencv-3.1.0-py36
tar -jxvf ../opencv-3.1.0-np112py36_2.tar.bz2
export PYTHONPATH=$PYTHONPATH:/home/<your-username>/opencv-3.1.0-py36/lib/python3.6/site-packages

PowerAI release 4 provides software packages for several Deep Learning frameworks and tools

https://public.dhe.ibm.com/software/server/POWER/Linux/mldl/ubuntu/README.html

  • Bazel
  • Caffe - BVLC, IBM, and NVIDIA variants
  • Chainer
  • DIGITS
  • NCCL
  • OpenBLAS
  • OpenMPI - with CUDA enablement
  • TensorFlow
  • Theano
  • Torch

Blogs