An understanding of Alzheimer’s disease might be on its way soon thanks to an unlikely source: hardcore gaming enthusiasts and their shining new Sony PlayStation 3 (PS3) consoles. Scientists at Stanford University are using the consoles as a big part of their project to understand a complicated biological process called “protein folding.”

The Folding@Home project, launched in 2000 by Stanford Professor Vijay Pande, uses distributed computing to simulate the complex folding process by which proteins attain their final shape and how “misfolds” can cause diseases like Alzheimer’s and Parkinson’s.

Using spare time from computer users worldwide, the project computes what they say would otherwise take years to compute. And thousands of computer users have been contributing to the computations by downloading the free software available at You must login or register to view this content. and letting it run when they are not using CPU time.

However, there has been a quantum jump in computation speed with the entry of the PS3’s new Cell microprocessor, Dr. Pande says. In a recent paper, he and his team have shown that the code optimised for the PS3 performs up to 40 times faster than other popular codes written for the single-core Intel Xeon 2.66 GHz processor.

The reason is the Cell Broadband Engine Architecture developed by Sony, Toshiba and IBM (the “STI alliance”) for an estimated $400 million. The developers rethought many of the basic architecture ideas for the PowerPC architecture used in Apple’s Macintoshes and created a Graphics Processing Unit with a dual-threaded Power Processing Element and eight Synergistic Processing Element.

These units are used to carry out floating point operations for graphics processing (the main computations required in gaming) and are hence also suited for scientific computations. Another big change brought in by the team was to use a high-bandwidth Element Interconnect Bus relying mainly on Direct Memory Access for memory transfer operations between the different processing elements and interfaces, speeding up the computations tremendously.


You must login or register to view this content.


“I think the Cell [inside the PS3] is a very interesting chip and could power scientific applications for years to come, as it is natural how to improve the chip in time [for example, by adding more SPEs],” Dr. Pande says.

In fact, supercomputers based on the Cell processor are already being built by IBM and others, achieving petaflop (1 petaflop = 1 million billion floating point operations per second) performance. The Folding@Home project, in the meanwhile, has already crossed 4.1 pFLOPS and PS3 users are among the biggest contributors to the speed, Dr. Pande’s website shows.

While GPUs released by NVIDIA are seen as contributing more FLOPS than PS3s, the Frequently Asked Questions page on the site explains that the PS3 provides a middle ground between GPUs (extremely fast but only some kinds of calculations are possible) and computer CPUs (extreme flexibility in types of calculations but limited speed). The PS3 (in common with GPUs) also allows for real-time monitoring of progress, making visualisation of the entire process possible in 3D, Dr. Pande’s paper says.

But more than the speed, he says the Cell architecture is a step forward in general. As he says in the conclusion to his paper, “new architectures such as the Cell may…open doors to calculations which we do not even consider due to their high computational expense on CPUs.”

Some gadgets are designed cool, some attain coolness through their applications. While the PS3 gaming console has staked claims for both, it seems coolness is also being thrust on it by scientists like the ones in the Pande group.