That is a lie misleading. They still haven't published the way they calculate their figures, and I'm afraid that that's because they are inflated. What they have said doesn't bode well - it's all estimation and fiddle-factors.

flops can be calculated in lots of different ways - and they just picked the method that yielded the biggest numbers (probably due to pressure from Sony). The result is that the numbers from the Folding@Home project can't be compared meaningfully to anything else - they exist in their own bubble of meaninglessness.

How are the FLOPS calculated?

People often use the number of Floating point operations per second (FLOPS) as a metric for the speed of a computer. One question that arises is how to compare machines with radically different architectures. In particular, what requires only a few operations (or even just a single operation) on one machine could require many operations on another. Classic examples are evaluations of functions like the exp(x) or sin(x). One GPU and Cell hardware, functions like this can often be calculated very quickly, say in one cycle, while this is often counted as 10-20 operations for other machines.

We take a conservative approach to FLOP calculation, rendering quantities such as exp(x) or sqrt(x) as a single FLOP, if the hardware supports it. This can significantly underestimate the FLOP count (as others would count an exp(x) as 10 or 20 FLOPS, for example). Others take a much less conservative approach and we are considering giving two counts, adding a more traditional (less conservative) count as well.

The ideal comparison would be to run Folding@Home on the supercomputer itself to test its speed. In this sort of comparison, Folding@Home would likely do very well, and we are investigating the best way to perform this benchmark, as we expect people would be interested.

So they've come up with a system that actually penalises their PC users? I'm afraid we will never have a fair basis for comparison, unless they manage to come up with a work unit that can be run on all platforms and give the same result. But that's not likely to happen.

I have reached the point where I simply don't believe what Folding@Home say, and that's really very depressing. Their credit calculations in particular are shockingly arbitrary.

And why don't I trust them? Because they keep publishing these petaflop claims and trying to directly compare them to general purpose architectures. It makes them look good, but it's simply not true. I think it hurts the distributed computing community more than it helps.

And every other week, someone comes here demanding a PS3 client from WCG. You think it hasn't been suggested before?!

I also am somewhat surprised that while IBM seems to be a huge sponsor of WCG, they did not think of using WCG to show of the power of the processor they helped creating.

I think that the limitations of the PS3 platform are scaring off most researchers.