• CUDA3.2 toolkit and latest release driver
  • CULA2.1 Premium, basic has extremely limited function.
  • PyCUDA v2011.1 or newer, preferable from Andreas’s git repo
  • scikits.cuda for cuBLAS and cuFFT

Installation can be done in two ways, manual, or an automated easy_install. If you wish to have all the source and the demo files, download the egg file from the python repo and follow the instructions in the README.INSTALL file.

Otherwise if you would rather just get to business, make sure you have the dependencies, then run:

sudo easy_install pycula

You should check your install by trying our tutorial!

Previous topic

Welcome to PyCULA’s documentation!

Next topic

Simple PyCULA Tutorial

This Page