During this Tech Breifing, Professor Vijay S. Pande (Chemistry, and by courtesy, Structural Biology and Computer Science) and Alex Chekholko and Phil Reese of IT Services detailed the general history and methodology behind the distributed computing model, how this is a paradigm shift from the mainframe/supercomputer model, and how the Folding@ Home project utilizes this novel technique.
Using the CPU power and communications abilities of unattended desktop computers throughout the world, the Folding@ Home project studies protein folding and misfolding. In this method of computer processing, known as distributed computing, different parts of a computer program run simultaneously on two or more computers that are communicating with each other over a network.
By harnessing the power of many machines, researchers are able to analyze far more data than they might have been able to do so otherwise. Indeed, Folding@home was the first machine to pass the petaflop scale and is now is at almost 10 petaflops in performance. The Folding @ Home project runs on any modern computer, including Playstation 3s, and runs only when the computer is otherwise not being used.
This session also demoed StarCluster, an open source cluster-computing toolkit for Amazon’s Elastic Compute Cloud (EC2). By watching this video, you will also learn about cloud based applications available to the campus, such as Amazon and others.