How to Set Up CUDA

  • Download CUDA driver, toolkit and SDK from this link. Choose version 2.0 or 2.1 (depending on whether the driver supports your card: most desktop cards are supported by the 2.1 driver; laptop cards are only supported by the 2.0 driver at the moment.
  • Install the driver, toolkit, and SDK you downloaded. It's recommended that you install everything to the default paths.
  • Launch CUDA SDK browser, and check to see if you can successfully run the examples included in the SDK.
  • If the program reports missing dll/library files, try to add "C:\CUDA\bin" and "C:\Program Files\NVIDIA Corporation\NVIDIA CUDA SDK\bin" (assuming you installed to the default paths) to your PATH environment variable. If you work with Linux or Mac, locate the corresponding paths and add them.
  • Go to "C:\Program Files\NVIDIA Corporation\NVIDIA CUDA SDK\projects" (or corresponding paths in Linux and Mac), and open the SDK project file. Note that CUDA 2.0 only works with Visual Studio 2005. Locate any project of your choice (say nbody), and choose to build it in Release mode. Identify and fix any problem you may encounter in building the project.
  • The SDK contains a project named simpleTemplates, which is a good starting point for you to write your own CUDA program.


  • You do not need to turn in anything for this assignment. However, by the class time on Tuesday, February 3rd, you should have completed all the steps above. Please report any unresolved problem.

Some Tips

Topic revision: r2 - 2009-01-28 - RuiwanG
This site is powered by the TWiki collaboration platform Powered by PerlCopyright © 2008-2018 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding UMass CS EdLab? Send feedback

mersin escort adana escort izmir escort gaziantep escort