The PlayStation 3 has recently flexed its floating point
muscles by leading all
processors at Folding@home. Sony’s game machine has been performing more work than any
other system on helping to understanding protein folding, misfolding and
related diseases such as Alzheimer’s and cancer.
A topic up for debate between the console circles is how the
PlayStation 3’s arch-nemesis, Microsoft’s Xbox 360, would perform on
Folding@home. While there is no Folding@home client for the Xbox 360, Vijay
Pande, Associate Professor of Chemistry at Stanford University and Folding@home
program lead, told Pro-G that the
CPU inside Microsoft console would definitely lag behind the PS3’s Cell
“We are simulating key processes in protein folding and
misfolding in Alzheimer's Disease. PS3's are performing aspects of these
simulations, and doing so about 20 times faster than a typical PC,” Pande said.
When asked if the Xbox 360 could be of use to the Folding@home program, he
answered, “Possibly, although the cell processor in the PS3 is much more
powerful for our calculations than the CPU in the Xbox 360.”
Mind you, the comparison may only be strictly between the
two CPU architectures of each console. The Xbox 360 uses a custom triple-core
PowerPC-based CPU, while the PlayStation 3 uses the Cell Broadband Engine
composed of seven processor elements.
While the Xbox 360 CPU may not be tops in terms of
Folding@home, its ATI-developed GPU could be much better. ATI Radeon GPUs
running Folding@home are outperforming the PlayStation 3 on a per capita basis.
Though whether or not the Xbox 360 will ever get its crack at helping to cure a
disease remains to be seen.
quote: I suppose the Cell would probably be faster, being able to execute 8 threads at a time. Although the 360 can execute 6 at a time, hardly a slouch.
quote: This even means that instruction level parallelism is unarchievable on Xenon so the Hyperthreading got in there only to fill FPUs and ALUs which were unutilized.
quote: Is Sony that desparate that they have to pay for this type of article?
quote: let alone an article that is concerned with the best method of folding proteins to fight disease - be considered a waste of time?
quote: ...PS3's are performing aspects of these simulations, and doing so about 20 times faster than a typical PC... [emphasis added]
quote: The R580 (in the X1900XT, etc.) performs particularly well for molecular dynamics, due to its 48 pixel shaders. Currently, other cards (such as those from nVidia and other ATI cards) do not perform well enough for our calculations as they have fewer pixel shaders. Also, nVidia cards in general have some technical limitations beyond the number of pixel shaders which makes them perform poorly in our calculations.
quote: We balance the points based on both speed and the flexibility of the client. The GPU client is still the fastest, but it is the least flexible and can only run a very, very limited set of WU's. Thus, its points are not linearly proportional to the speed increase. 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). We have picked the PS3 as the natural benchmark machine for PS3 calculations and set its points per day to 900 to reflect this middle ground between speed (faster than CPU, but slower than GPU) and flexibility (more flexible than GPU, less than CPU).
quote: Since the 360's CPU is an in-order There's mostly no gain in adding execution units to the core since there would be no way to feed all of them, that's why the 360's cpu hyperthreads, kinda like the p4 but hopefully a lot more efficient. Besides that, not all execution units provide floating point performance, because there are different types of them inside the cpu.
quote: On a side note, IBM designed both cpu's, do you really think that they would be promoting cell as the ultimate processor architecture if xenon was more powerfull?
quote: Xenon only performs fast in very specific cases, and are even more specific than those for the cell, contrary to what you might think. Since the fact that you lack dynamic execution in that cpu, only in very specific streams of instruction do you get to utilize all the execution units. I really doubt anyone at Stanford would want to put time tuning fah to run fast on xenon.