backtop


Print 32 comment(s) - last by randomwalk16.. on Sep 24 at 9:55 PM

PlayStation 3 helps Folding@home with further disease research

Aside from powering some of this holiday season’s most impressive games and providing playback of high-definition movies, the Cell Broadband Engine is also exceptionally adept at medical and scientific applications requiring massive amounts of floating-point computational power.
 
Since joining Stanford University’s Folding@home program in March, the PlayStation 3 has led all processors in sheer productivity numbers. It should come as no surprise that Sony announced this week that PlayStation 3 consoles all over the world have helped the Folding@home project to reach a petaflop, a milestone never before reached on a distributed computing network.
 
"The recent inclusion of PS3 as part of the Folding@home program has afforded our research group with computing power that goes far beyond what we initially hoped," said Vijay Pande, Associate Professor of Chemistry at Stanford University and Folding@home project lead. "Thanks to PS3, we are now essentially able to fast-forward several aspects of our research by a decade, which will greatly help us make more discoveries and advancements in our studies of several different diseases."
 
"When we introduced PS3, we knew its incredible processing power would allow for a great deal of innovation and creativity," said Jack Tretton, president and CEO of SCEA. "It's extremely rewarding to see that the scientific community has found a way to harness PS3 technology for humanitarian purposes and we continue to be amazed at what gamers and the Folding@home community have been able to accomplish in such a short amount of time."
 
The Folding@home program runs simulations in protein folding, helping scientists understand – and hopefully curing – diseases such as Alzheimer's, Parkinson's and certain forms of cancer. That’s not all the PS3’s CPU is able to do for the medical community, though, as the Cell Broadband Engine is also helping doctors at Mayo Clinic with medical imaging.



Comments     Threshold


This article is over a month old, voting and posting comments is disabled

This is a great use of technology
By Orbs on 9/21/2007 1:54:23 PM , Rating: 5
It would be great to see AMD/ATI & nVidia help get Folding@Home working on their full line of GPU's (as far as I know only a select few ATI cards run F@H). Given that they're even more efficient than PS3 and have a much larger user base if their full lines of GPUs were enabled, just think of how this could help this important research!




RE: This is a great use of technology
By AlphaVirus on 9/21/2007 2:00:10 PM , Rating: 1
ATI does contribute with some of their GPU models but you have to remember, if you take a low end card such as an ATI9600 or something under an X800 series you barely have enough power to run your own apps. Their choice of which gpus can run the F@H is wise, they are trying to not hinder the weaker computer users.


By Operandi on 9/21/2007 2:50:49 PM , Rating: 2
The same would be said for any GPU, folding is going use whatever processing power you have available. I don't have Folding@Home on my machines but I belive the idea is you fold when the machine is idle.

Not that it would matter though, even if you could get it work a X800 would probably be pretty limited in the types of calculations it could do to be at all useful, not to mention a 9600. Modern GPUs like the 1900 and 8800 are far more open in flexible.


By alancabler on 9/23/2007 3:05:21 PM , Rating: 2
ATI cards pre-x16xx series are not capable of running the FAH GPU client due to architectural inadequacies. x1xxx series of lesser power than x1600 might work, but can't meet FAH deadlines.
NVidia G80 series are allegedly capable, but nVidia has been unable to produce FAH-capable drivers, so far. Many FAH donors w/G80s are anxiously awaiting the next move from nVidia.

A FAH GPU client for the x2 series ATI cards is in the works.


RE: This is a great use of technology
By therealnickdanger on 9/21/2007 2:01:27 PM , Rating: 5
I have to wonder, can you write off the additional money you spend on electricity when running F@H as a charitable donation?


By ghost101 on 9/21/2007 3:22:36 PM , Rating: 2
You mean get a tax rebate? Because i dont see private electricity companies subsidising folding@home costs anytime soon.


RE: This is a great use of technology
By falc0ne on 9/21/07, Rating: -1
By Dactyl on 9/22/2007 12:52:45 AM , Rating: 2
How is it selfish to want a tax write-off for giving to charity?

The government allows people to write off charitable contributions in order to encourage charity.

A write-off just means that more of the money goes to the charity and less to the government. There is nothing selfish about it. I don't think you understand how our tax system works.

Maybe I am willing to give $25 to charity, but if you allow me to write it off, then I save $5 on my taxes, so I am willing to give $30 to charity. The end result: the charity gets $5 more, the government gets $5 less, I have the same amount of money in my wallet--no greed involved at all.


By PrinceGaz on 9/22/2007 1:28:53 PM , Rating: 2
Rare? I thought Alzheimer's was quite common in the elderly. Allegedlly, over five million Americans have it, which is anything but rare:

http://www.alz.org/alzheimers_disease_alzheimer_st...


By randomwalk16 on 9/24/2007 9:55:55 PM , Rating: 2
The last comment was probably tongue-in-cheek, but it got me curious.

The F@H website says their in-house ps3's eat about 200W of energy when running the program.

Figure about 8 hrs per work-unit (also taken from the site), and that comes to about 1.6KWhr. Assume 13cents per kilowatt-hour (approximate cost in New Jersey) and you get about approximately 21cents per work unit.

Therefore, if you did a work unit every night, it would be about 75$ a year tacked onto your electricity bill. Not exactly negligible, but most people aren't doing a work unit a night, and even if you are, thats not going to break the bank.

Also, many other states have significantly different costs of energy....


RE: This is a great use of technology
By AmbroseAthan on 9/21/2007 2:02:30 PM , Rating: 2
I was actually about to ask/post something similiar to this. I had heard rumors of them working on enabling the NVidia's 8800's and also ATI's 2900's; does anyone know if this is still the case? I tried to go explore a bit in the Folding forums, but they are blocked at work. I would love if I was able to put my 8800GTS to use for this.


By falc0ne on 9/21/2007 10:45:19 PM , Rating: 2
me too, very much agree


RE: This is a great use of technology
By h0kiez on 9/21/2007 2:03:09 PM , Rating: 1
I thought the PS3 basically trounced even higher-end GPUs from both ATI and Nvidia.

Then again, I think I also remember reading something convoluted basically saying that the PS3s numbers were inflated because it was only assigned certain types of F@H threads because the hardware was so specialized...ah, hell, I don't even know.


By AmbroseAthan on 9/21/2007 2:15:19 PM , Rating: 2
No NVidia cards are supported that I know of, and the only "high-end" ATI cards supported are the 1900's; nothing in the 2-series yet.

But on that note, using only ATI's older cards cards:
quote:
When it comes to pure performance though, the PS3’s Cell Broadband Processor is still no match for ATI GPUs for protein folding. The GPUs on Folding@home sit at 41 current TFLOPS, which come from only 700 processors. If there were as many GPUs folding as there are PS3s on the network, it can be extrapolated that GPUs could reach 876 TFLOPS.


Stole that from the "led all processors" link of the article above. At the time, the PS3 did 367 TFLOPS over 14971 active processors.


RE: This is a great use of technology
By ajfink on 9/21/2007 2:24:34 PM , Rating: 2
No, per-unit GPUs still outclass PS3s. High-end GPUs are the basis for massive stream processing, after all. The highly-parallel architectures are far more aptly suited for this sort of number crunching than even the Cell processor. That being said, on a per-watt basis, the Cell probably beats out GPUs. Take an X1950~, though, and it would trounce a PS3 in overall performance in terms of F@H and several other applications.


By killerroach on 9/21/2007 2:43:43 PM , Rating: 2
Precisely. As per the most recent stats, it takes about 17 GPUs to provide 1 TFLOP of folding performance, while it takes about 40 Cell processors to reach that same performance level. That being said, there are about 35,000 PS3s actively folding, which is in stark contrast to the 737 GPU users. As a result, over three-quarters of the entire network performance of the F@H project is directly related to the PS3. Not too shabby.


RE: This is a great use of technology
By Dactyl on 9/21/2007 2:55:15 PM , Rating: 4
It's important to note that the x86 CPUs can do the most broad kinds of work. The PS3s can't do quite as much, but what they can do, they can do much faster than an x86 CPU.

Graphics cards are another step in that direction. Their abilities are even more limited in the types of calculations they can perform. However, they are even faster than the PS3s at performing those calculations.

So there are reasons to use all three in the F@H program. Your x86 CPUs are still very valuable.


RE: This is a great use of technology
By Dactyl on 9/21/2007 9:43:19 PM , Rating: 5
Since I've been downranked, I feel like I have to respond.
quote:
What type of calculations the PS3 client is capable of running?
The PS3 right now runs what are called implicit solvation calculations, including some simple ones (sigmodal dependent dielectric) and some more sophisticated ones (AGBNP, a type of Generalized Born method from Prof. Ron Levy's group at Rutgers). In this respect, the PS3 client is much like our GPU client. However, the PS3 client is more flexible, in that it can also run explicit solvent calculations as well, although not at the same speed increase relative to PC's. We are working to increase the speed of explicit solvent on the PS3 and would then run these calculations on the PS3 as well. In a nutshell, the PS3 takes the middle ground between GPU's (extreme speed, but at limited types of WU's) and CPU's (less speed, but more flexibility in types of WU's) .

That's straight from the Folding@Home project itself:
http://folding.stanford.edu/FAQ-PS3.html

What I wrote is absolutely correct, and I stand by it 100%.


By Ringold on 9/22/2007 12:05:09 AM , Rating: 2
Anybody familiar with F@H and has read much of what people in the Pande Group write knows what you mean. Never mind the down-mod gremlins. :P


RE: This is a great use of technology
By mars777 on 9/22/2007 1:10:04 AM , Rating: 3
I didn't downvote you but i argue that a PS3 can do any kind of calculation that an X86 machine can.

There are no types of calculations that a Turing machine can't do, it's just a matter of how much time you need to accomplish that.

And i stand 200% behind that.

That said, GPUs weren't a complete Turing machine until recently. Today you can accomplish on the GPU anything you can on your x86 CPU. These projects are called GPGPU.

Some freaks made an software audio effect processor, and audio acceleration using the GPU...

References:
http://en.wikipedia.org/wiki/Turing_machine
http://www.nforcershq.com/article1853.html
http://www.steve-lacey.com/blogarchives/2006/05/au...


By alancabler on 9/23/2007 3:26:37 PM , Rating: 2
quote:
it's just a matter of how much time you need to accomplish that.
And with that disclaimer, you've given part of the reason why the PS3 is less versatile for FAH calculations than a standard CPU. GPU's are even less versatile than the PS3.
A Pentium mmx rig is a Turing machine, but that doesn't mean it is FAH-capable.
The CPU client can perform all types of FAH calculations, while the PS3 is more limited, and the mighty GPU client is very much architecturally restricted in the types of FAH calculations it can perform... subtle arguments aside.

Discussions at FAH go into greater detail as to why the GPU and PS3 clients are limited in scope, despite their awesome power.


"It seems as though my state-funded math degree has failed me. Let the lashings commence." -- DailyTech Editor-in-Chief Kristopher Kubicki














botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki