facebook rss twitter

Former RARE employee: The Xbox 360's GPU could save parallel computing researchers thousands

by Steven Williamson on 14 September 2009, 09:38

Tags: Microsoft (NASDAQ:MSFT)

Quick Link: HEXUS.net/qatxe

Add to My Vault: x

A new study by a University of Warwick researcher has demonstrated that researchers trying to model a range of processes could use the power and capabilities of a particular XBox chip as a much cheaper alternative to other forms of parallel processing hardware.

Dr Simon Scarle, a researcher in the University of Warwick's WMG Digital Laboratory, wished to model how electrical excitations in the heart moved around damaged cardiac cells in order to investigate or even predict cardiac arrhythmias (abnormal electrical activity in the heart which can lead to a heart attack). To conduct these simulations using traditional CPU based processing one would normally need to book time on a dedicated parallel processing computer or spend thousands on a parallel network of PCs.

Dr Scarle however also had a background in the computer games industry as he had been a Software Engineer at the Warwickshire firm Rare Ltd, part of Microsoft Games Studios. His time there made him very aware of the parallel processing power of Graphical Processing Unit (GPU) of the XBox 360. He was convinced that this chip could, for a few hundred pounds, be employed to conduct much the same scientific modelling as several thousand pounds of parallel network PCs.

The results of his work have just been published in the journal Computational Biology and Chemistry under the title of "Implications of the Turing completeness of reaction-diffusion models, informed by GPGPU simulations on an XBox 360: Cardiac arrhythmias, re-entry and the Halting problem". The good news is that his hunch was right and the XBox 360 GPU can indeed be used by researchers in exactly the money saving way he envisaged. Simon Scarle said:

"This is a highly effective way of carrying out high end parallel computing on "domestic" hardware for cardiac simulations. Although major reworking of any previous code framework is required, the Xbox 360 is a very easy platform to develop for and this cost can easily be outweighed by the benefits in gained computational power and speed, as well as the relative ease of visualization of the system."

However his research does have some bad news for a particular set of cardiac researchers in that his study demonstrates that it is impossible to predict the rise of certain dangerous arrhythmias, as he has shown that cardiac cell models are affected by a specific limitation of computational systems known as the Halting problem.


HEXUS Forums :: 7 Comments

Login with Forum Account

Don't have an account? Register today!
Not meaning to be rude to the guy (it sounds like good medical research), but why is it a surprise that certain calculations can be done very fast on a GPU these days? Since there are already two GPU-programming frameworks out there, do we really need a third one just for 2005-era ATI GPU computing on an xbox?

My guess is that events have overtaken this guy a bit since he first started out on what in ~2006 must have been a fairly original idea, and now more convenient mechanisms exist with which to do GPU-based science… and on much more powerful graphics hardware.
…and how long has the folding client been out for the PS3 now?
darkghost
…and how long has the folding client been out for the PS3 now?

Ah… but isn't that subtly different, as it makes use of the cell processor (i.e. main CPU) rather than the PS3 GPU? Or am I wrong about that?
Fraz
Ah… but isn't that subtly different, as it makes use of the cell processor (i.e. main CPU) rather than the PS3 GPU? Or am I wrong about that?

Not sure - I think my point was kinda geared more towards the fact that a folding client has been out on the PS3 for a while now and there has been nothing for the Xbox. So RARE are hyping about doing something in parallel that nobody has even bothered to do using a single thread of the 360 yet. The whole point of this research approach is the user base that makes up the numbers!
all sounds really good and way too intelligent for me! but lets hope those gpu's dont red ring eh!