GPUs on Grid5000

Purpose

This page presents how to use GPUs on Grid'5000 and how to install your own NVIDIA drivers and CUDA installation.
In this tutorial, we will first compile and use CUDA examples and in the second part, we will install NVIDIA drivers and compile CUDA 5 from a simple wheezy-x64-base environment.

Pre-requisite

• A basic knowledge of Grid'5000 is require, we suggest you to read Getting Started tutorial first.
• Information about hardware information and GPUs availability can be found on Special:G5KHardware.

Compiling examples

GPU are available in Grenoble, Lyon and Lille. NVIDIA drivers 304.54 and CUDA 5.0 are installed by default on nodes.
You can reserve a node with GPU using OAR GPU property. For Grenoble and Lyon:

frontend:

oarsub -I -p "GPU='YES'"

or for Lille:

lille:

oarsub -I -p "GPU='SHARED'"

	Warning
	Please note that Lille GPU are shared between nodes and the oarsub command will differ from Grenoble or Lyon. When using GPUs at Lille, you may encounterer some trouble, you can read more about Lille GPU on Lille:GPU.

We will then download CUDA 5.0 samples and install them.

frontend:

wget http://git.grid5000.fr/sources/cuda-samples_5.0.35_linux.run -P /tmp/ && cd /tmp

frontend:

sh cuda-samples_5.0.35_linux.run -cudaprefix=/usr/local/cuda-5.0/

You will be prompt to accept the EULA and for a installation path, we suggest you to use /tmp/samples/.
Then you can go to installation path, in our case /tmp/samples/. If you list the folder, you will see a lot of folder starting with a number, these are CUDA examples. CUDA examples are describe in this document. You might also want to have a look to "doc" directory.