View Full Version : Stanford University’s uses ATI stream to squeeze research

10 October 2006, 08:22 AM
Stanford University’s Folding@home uses ATI stream computing to squeeze three years worth of disease research into just one month

ATI Technologies Inc.,a Canada-based $2 .2 billion world leader in the design and manufacture of innovative 3D graphics and digital media silicon solutions (TSX:ATY, NASDAQ:ATYT)recently announced that Stanford University has released software to utilize ATI processors for the Stanford’s Folding@home distributed computing project, yielding an enormous increase in the speed of disease research. By enabling Folding@home participants to make use of the powerful processing architectures found in ATI Radeon® X1900 and Radeon® X1950 graphics cards, Stanford expects to gain new capabilities that have never been possible before, allowing researchers to address questions that were previously considered impossible to tackle computationally. With this new processing power now tapped, researchers expect to discover even more about the science of protein folding and folding-related diseases, with potentially dramatic effect for studies into Alzheimer’s, cancer, Huntington’s and Parkinson’s diseases among others.

“The quest to understand diseases like cancer and Alzheimer’s is the reason why the Folding@home program exists, and with the phenomenal acceleration provided by ATI processors, we can truly say this is life-changing work,” said Vijay Pande, associate professor of chemistry, and director, Folding@Home distributed computing project, Stanford University.

What is Folding@home?

Stanford University’s Folding@home program studies the ways in which proteins fold, assembling themselves in order to carry out important functions in the body. While protein folding is critical and fundamental to virtually all of biology, today the concept is still considered to be a mystery. When proteins do not fold correctly, consequences include a number of diseases, so to better understand this area, Stanford’s Folding@home program groups hundreds of thousands of home computers together worldwide, studying folding by taking advantage of each computer’s processing power to form a distributed supercomputer. Today the program consists of approximately 200,000 computers, using various computer processors.

With support for ATI processors, Folding@home participants and disease researchers are tapping into the new capabilities offered by these advanced architectures, with a Radeon X1900 card performing scientific calculations at 20x to 40x the speed of traditional computer processors.2 At that rate, using the processing power of one ATI Radeon X1900, three years worth of research can now be done in just one month.3 The processing power of just 5,000 ATI processors is also enough to rival that of the existing 200,000 computers currently involved in the Folding@home project; and it is estimated that if a mere 10,000 computers were to each use an ATI processor to conduct folding research, that the Folding@home program would effectively perform faster than the fastest supercomputer in existence today, surpassing the 1 petaFLOP level.

Folding@home is made up of a number of engine cores – software designed to calculate various aspects of molecular dynamics associated with protein folding. Stanford’s Gromacs software core, introduced in May 2003, remains the fastest and one of the most optimized molecular dynamics codes in the world. Through the work of Stanford University researchers, the Gromacs software was tuned to take advantage of ATI’s Radeon X1900 and Radeon X1950 products, delivering the massive speed up in processing referenced above. However, as versatile as the Gromacs code is when combined with ATI processors, some protein folding calculations cannot be performed on ATI hardware, and require the use of computer processors, and other engine cores.

CGTalk Moderation
10 October 2006, 08:22 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.