backtop


Print 59 comment(s) - last by masher2.. on Feb 5 at 10:03 PM

PlayStation 3 gives strength to Stanford's Folding@home

Since last March, the PlayStation 3 has been one of the leading contributing technologies to the Folding@home effort.

Now, less than a year since release, more than one million users have taken part in Stanford University's Folding@home project. According to Sony, this equates to roughly 3,000 PS3 users registering for Folding@home per day or 2 new registered users every minute worldwide.

"Since partnering with SCEI, we have seen our research capabilities increase by leaps and bounds through the continued participation of Folding@home users," said Vijay Pande, Associate Professor of Chemistry at Stanford University and Folding@home project lead.  "Now we have over one million PS3 users registered for Folding@home, allowing us to address questions previously considered impossible to tackle computationally, with the goal of finding cures to some of the world's most life-threatening diseases.  We are grateful for the extraordinary worldwide participation by PS3 and PC users around the globe."

According to the Folding@home team, a network of roughly 10,000 PS3s can accomplish the same amount of work as a network of 100,000 PCs. It took just six months after PS3 joining Folding@home for the project to surpass a petaflops, a computing milestone that had never been reached before by a distributed computing network. On September 16, 2007, Folding@home was recognized by Guinness World Records as the world's most powerful distributed computing network.

Currently PS3 users make up approximately 74 percent of the total teraflop computing power of the Folding@home project.

The Folding@home program runs simulations in protein folding and misfolding, helping scientists understand – and hopefully curing – diseases such as Alzheimer's, Huntington'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

RE: GPUs are the kings though
By zpdixon on 2/5/2008 2:18:24 AM , Rating: 2
quote:

GPU < Cell < CPU in flexibility

I am sorry but this is BS. I develop, mostly in C & assembly, on the Cell on my PS3 running Linux. The SPU and PPU cores on the Cell are general purpose units , you can do anything with them. Sure the SPUs are highly-optimized for integer & single-precision SIMD workloads, and you have to constantly shuffle around your data between the 128 kB Local Store and the RAM to accomodate for the relatively small amount of directly addressable memory. But I repeat myself: technically the Cell, even if highly optimized for some workloads, is a general purpose CPU.

Sony, Toshiba, IBM are using and plan to use the Cell in numerous products: HPC servers, media devices, laptops (yes), game consoles, etc.


RE: GPUs are the kings though
By zpdixon on 2/5/2008 2:23:31 AM , Rating: 2
s/128/256/;


RE: GPUs are the kings though
By wien on 2/5/2008 5:47:56 AM , Rating: 3
GPUs are more or less general purpose these days too, but that doesn't mean they run normal general purpose code with any kind of efficiency. Same with the SPEs. Even though they can in theory run anything, they're not really efficient unless it's a certain type of workload. CPUs also have their weak points of course, but of the three they're the least picky about the workload they need to operate efficiently.

Hence "GPU < Cell < CPU in flexibility".


RE: GPUs are the kings though
By zpdixon on 2/5/2008 12:20:20 PM , Rating: 2
You are obviously not a Cell programmer either.

The Cell SPU ISA defines arithmetic, floating-point, branching, boolean , bit manipulation instructions. It is just as general-purpose as a CPU can be: they couldn't make the ISA more flexible .

Don't comment on what you don't know.


RE: GPUs are the kings though
By wien on 2/5/2008 2:41:05 PM , Rating: 2
I know it has a similar ISA to other processors. Even a DX10 level GPU can do most of that you mentioned there. The point is that it's still very dependant on having suitable workloads to take advantage of all those FLOPS. A GPU would absolutely croak if you for instance tried to compile on it because that kind of branch heavy workload just isn't suitable for the GPU's architecture.

The Cell SPE is similarity more specialised than a normal CPU core in terms of workloads it is good at, and that makes it less flexible in my book. The ISA has nothing to do with it.


RE: GPUs are the kings though
By zpdixon on 2/5/2008 3:12:10 PM , Rating: 2
I think a point of vocabulary is causing this disagreement between us. For me "flexible" means "able to run a wide range of workloads" (at average or above average performance levels). While it seems you use the word "flexible" as a synonym for "able to run a wide range of workloads in way that always has to be as efficient as possible" (so you require the presence of traditional CPU features which do not exist on the SPUs: caches, out-of-order execution, branch prediction, register renaming, etc).


"Nowadays, security guys break the Mac every single day. Every single day, they come out with a total exploit, your machine can be taken over totally. I dare anybody to do that once a month on the Windows machine." -- Bill Gates














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