Researchers at the Georgia Institute of Technology Research Institute (GTRI) are developing a programming tool that will enable defense industry engineers to use graphics processing units (GPUs) without having to learn to program them. "As radar systems and other sensor systems get more complicated, the computational requirements are becoming a bottleneck," says GTRI researcher Daniel Campbell. "We are capitalizing on the ability of GPUs to process radar, infrared sensor, and video data faster than a typical computer and at a much lower cost and power than a computing cluster."
Georgia Tech professor Mark Richards is working with Campbell and graduate student Andrew Kerr to rewrite common signal-processing commands to run on a GPU. The researchers are writing functions defined in the Vector, Signal, and Image Processing Library (VSIPL), an open standard. Currently, the researchers are writing the functions in Nvidia's CUDATM language, but the underlying principles can be applied to GPUs developed by other companies, Campbell says. The resulting GPU VSIPL will enable engineers to use high-level functions in C programs to perform linear algebra and signal-processing operations, and recompile with GPU VSIPL to capitalize on the speed of GPUs.
In the future, the researchers plan to develop other defense-related GPU function libraries and design programming tools to allow for the use of other highly-efficient processors, such as the Cell broadband engine processor used in the Playstation 3 game console.
From Georgia Tech Research Institute
View Full Article
Abstracts Copyright © 2009 Information Inc., Bethesda, Maryland, USA
No entries found