Nvidia CUDA Helps Bring Parallel Programming to the Classroom -- ADTmag

Nvidia CUDA Helps Bring Parallel Programming to the Classroom

By David Nagel
November 26, 2007

Nvidia has released a public beta of CUDA 1.1, an update to the company's C-compiler and software development kit. CUDA stands for "Compute Unified Device Architecture." It's used for developing multicore and parallel processing applications on graphics processing units (GPUs), specifically Nvidia's 8-series GPUs and their successors.

While the 1.1 beta was originally released to developers on Nov. 12, an update was posted last week that includes new public beta Linux display drivers.

In addition, Nvidia announced that more than 20 universities around the world have adopted CUDA for multicore and parallel processing programming, with several more also exploring CUDA for inclusion in their parallel programming curricula.

"Perhaps the most important challenge facing the computing community is the move to parallel processing. As educators, teaching parallel hardware and software today are vital to giving our students the tools they need to build tomorrow's hardware and software," said John Owens, assistant professor in the department of computer engineering at University of California at Davis, in a statement released earlier this month to coincide with the SuperComputing 07 conference.

New in the 1.1 beta release is a CUDA dll for standard display drivers, support for 64-bit Windows XP, graphics interoperability across multiple GPUs, new async and stream management functionality, and additional source code examples, according to information posted at Nvidia's developer forums.

CUDA 1.1 presently supports only Linux and Windows XP environments. A Mac OS X release is expected in January, and a version for Windows Vista is also forthcoming. The 1.1 beta supports Windows XP (32- and 64-bit) and Red Hat Linux 4 and 5 (32- and 64-bit). The production release will support a full set of Linux distributions, according to the company.

More details on Nvidia's CUDA is available here.

About the Author

Dave Nagel is the executive editor for 1105 Media's educational technology online publications and electronic newsletters. He can be reached at [email protected].