Install PetaVision Ubuntu modified by Brian Broom-Peltz

Brian Broom-Peltz — Thu, 30 Apr 2015 20:59:39 -0000

Ubuntu Installation

This is a tutorial on getting PetaVision and running on Ubuntu-14.04 with GPU support, assuming you have a clean installation of Ubuntu.

Requirements

An NVIDIA card with compute capabilities 3.0 or above. Check at https://developer.nvidia.com/cuda-gpus to see if your video card is supported
Ubuntu 14.04
sudo access

Initialization

sudo apt-get update
sudo apt-get upgrade
sudo apt-get install cmake cmake-curses-gui gcc g++ gdb subversion

OpenMPI

sudo apt-get install libopenmpi-dev openmpi-bin

GDAL

sudo apt-get install libgdal-dev gdal-bin

CUDA and NVIDIA Driver

Go to http://www.nvidia.com/Download/index.aspx?lang=en-us and find your video card. Download the linux RUN file. Note the location.
Go to https://developer.nvidia.com/cuda-downloads.
Click the RUN file for your archetecture. Make sure if you have a 64 bit version of ubuntu that you download the 64 bit version cuda. Note the location.

NVIDIA drivers may not be installed unless X is not running. Here's how to install the drivers.

Reboot Ubuntu. At the GRUB menu, select Acvanced options for Ubuntu.
Boot Ubuntu into recovery mode.
Select root to get into root command line.
mount -o remount,rw #This remounts your drive into read/write mode
cd path/to/cuda.run
chmod u+x path/to/cuda/driver.run
./path/to/cuda/driver.ruh

Follow the onscreen instructions.
When it asks if you want to install NVIDIA driver, select yes. This version is outdated, but cuda needs this driver to work.
Make sure to install CUDA 6.5 toolkit into the default location. Add a symbolic link to /usr/local/cuda when asked.

cd path/to/driver.run
chmod u+x path/to/driver.run
./path/to/driver.run

Follow the instructions on the screen.
When asked to uninstall the driver installed by cuda, select yes.
When asked to update your Xorg, select yes. This will use this driver as your GUI display.
Run nvidia-smi and make sure you see your video card.
Reboot.

CUDNN

Go to https://developer.nvidia.com/cuDNN and click Download at the bottom.
Register with NVIDIA developers if need be, and wait for confirmation.
Download the Linux version of CUDNN
(Optional: if you have access to NMC's compuneuro, grab it from here: /nh/compneuro/Data/cuDNN)

Finalization

Do one final update/upgrade for everything
sudo apt-get update
sudo apt-get upgrade

Checking Out and Installing PetaVision

cd ${HOME}
mkdir workspace
cd workspace
svn co https://<useranme>@svn.code.sf.net/p/petavision/code/trunk PetaVision
svn co https://<username>@svn.code.sf.net/p/petavision/code/PVSystemTests PVSystemTests
#You can download your sandbox here
cp PetaVision/docs/cmake/CMakeLists.txt .
#Here, you would add your sandbox to the end of CMakeLists
ccmake .

CMake options:
CMAKE_BUILD_TYPE = Release #Optimiztaion for CPU
CUDA_GPU = True #Depending on if you want to use GPUS
CUDA_RELEASE = True #Optimization for Cuda
CUDNN = True #If you're using CUDNN
CUDNN_PATH = /path/to/cudnn/folder #The path to the cuDNN folder you downloaded/copied
OPEN_MP_THREADS = True #If we use threads or not

If some of these options do not show up on ccmake, fill the ones you can, press c to configure, and look again for variables
Press G to generate when avaliable.

To Test:

cd ~/workspace/PVSystemTest/GPUSystemTest
make -j 8
Release/GPUSystemTest -p input/postTest.params -t

Optional: Python

Ubuntu comes with python 2.7. We suggest installing anaconda python for analysis tools.
- Go to https://store.continuum.io/cshop/anaconda/
- Enter your email
- Click I WANT PYTHON 3.4
- Click the linux penguin
- Click the download link and save it

chmod u+x path/to/python.sh
./path/to/python.sh

Follow the directions. Make sure to specify yes when it asks you if you want anaconda in your path.

source ~/.bashrc

To ensure it's working, type python and make sure the text splash says Python 3.4.1 |Anaconda 2.1

Optional: Octave

sudo apt-get install octave liboctave-dev
octave

In octave:
pkg install -forge general
pkg install -forge control
pkg install -forge signal
pkg install -forge image
pkg install -forge parallel
exit

echo pkg load all >> ~/.octaverc

Recent changes to Install PetaVision Ubuntu