(per Folding@Home Stats page)

3051 machines = 92 teraflops

teraflop / machine will likely drop in the future as people don't crunch 24/7, but instead crunch part time...

At 30 gigaflops per machine, this puts the PS3 solidly ahead of any CPU crunchers, especially on a performance per dollar spent (including electric bill).

GPU protein folding is still more powerful and economical (raw performance and performance per dollar spent), but it is not as user-friendly as the PS3 folder.

ALSO, the PS3 and GPU folders cannot work on 7/8ths of the datasets for FAH because they are not as complex as CPU folders... so we still need CPU folders, of course. We still need Rosetta folders too!

---

251 Tera Flops, 10238 PS3s and rising.

"The most powerful distributed computing project was SETI@Home at 280 teraflops.
Now it is Folding@Home, which is at 492 teraflops and rising." (is this quote right about SETI?)

Anyhow, this news is very exciting for science!

---

I think its important to keep posting here and at Rosetta, as I do not know how frequent these two projects would communicate. The stated numbers are impressive indeed, and should trigger some response from the BOINC side, especially since today is the day it came out worldwide, including forgotten Europe (higher price, only software emulation for PS2 games and later start date).

---

It will be interesting to see where the PS3s level out at as far as # of teraflops. Gigaflop per machine will certainly drop as more gamers adopt PS3 folding; this is because if they have it crunching 16 hours or 8 hours a day (instead of 24), it will drag down the average.

If there were 40,000 dedicated PS3 folders, it would be 1 petaflop!

The fastest supercomputer "Blue Gene" has a peak of 360 teraflops. Folding@Home does not measure teraflops by using the peak values, they use average throughput for each WU.

---

SO, anyone knows, when we can expect Boinc and seti@home on PS3 linux?

---

--
Get up, stand up! Don\'t give up the fight!
Credits will make everybody feel high! ;-)

SO, anyone knows, when we can expect Boinc and seti@home on PS3 linux?

Folding@Home donators tried to run it in Linux on the PS3 but it has to be coded properly because they got 1/10th of the performance they were expecting. It was not utilizing the system. Unfortunately, it's not that easy.

---

Early Saturday morning update:

PS3s: 465 Teraflops, 18992 machines.
F@H Total: 716 Teraflops.

The PS3s by themselves surpassed IBM's "Blue Gene" (#1 supercomputer in da world).

... AND F@H calculates flops very conservatively. A square root of a number can easily be counted as 10-20 flops, but F@H counts it as 1-2 flops.

Watchers of the project are hoping that F@H's total reaches 1 Petaflop. It could happen by Sunday or Monday (unless there are many people who haven't bothered with the firmware update until getting off work Friday).

---

I just signed up my ps3 to folding at home and as a seti and Einstein host I went to their leader boards and kind of funny the "default" built into the ps3 is called ps3 and now 10000 people who have not entered a name into their console are top 60 on their board after 15 hours, will take top spot within 2 days.

http://www.boincstats.com/signature/user_178540_project-1.gif

---

Early Saturday morning update:
PS3s: 465 Teraflops, 18992 machines.
F@H Total: 716 Teraflops.

Getting there!
http://fah-web.stanford.edu/cgi-bin/main.py?qtype=osstats

As of Sat, 24 Mar 2007 18:28:50 (no timezone specified!!!)

PS3 668 Teraflops, 27260 active machines
Total 922 Teraflops

Well, that just about wraps it up... just these few machines and their spotty owners are now wielding much more power than the whole of the BOINC infrastructure combined with its 425,578 active machines cranking out 537 teraflops and consuming 120 Gigawatts. (60 full size power stations!)

That is an average processing efficiency ratio of 425578/27260 * 668/537 = ~20:1 over conventional machines, and even with a max power consumption of 380W, a probable 15:1 improvement in Watt/flop ratio.

Overnight, the PS3 has just about turned the whole of BOINC into a 120 Gigawatt irrelevance! Perhaps climate change could be better mitigated now by just switching the dedicated dinosaurs off, replacing with PS3s or finding another hobby instead?

But first have to get BOINC running on it... Time to get a PS3 and do it...

Overnight, the PS3 has just about turned the whole of BOINC into a 120 Gigawatt irrelevance!

Oh, I don't know.

The fact that they can crunch so fast is cool, but I just can't get interested in "folding" -- and I think that means something.

It certainly isn't relevant enough for me to buy a PS3.

---

why f@h is not part of the BOINC group?

---

why f@h is not part of the BOINC group?

They're working on a Boinc client, as well as a GPU client, and other clients...
More info at Folding@Home high performance client FAQ and Frequently Asked Questions (FAQ)

---

"The limits of the possible can only be defined by going beyond them into the impossible"
Arthur C. Clarke

"Science may set limits to knowledge, but should not set limits to imagination."
Bertrand Russel

"April 2006 We have updated much of the code, but are now in the midst of dealing with staff turnover in the BOINC part of the development team, which has slowed development." <-- from f@h website
lameeeeee

---

"April 2006 We have updated much of the code, but are now in the midst of dealing with staff turnover in the BOINC part of the development team, which has slowed development." <-- from f@h website
lameeeeee

they had a BOINC version for some time. But I guess they are just celebrating the PS3 impact on their numbers. It's quite surprising for me to see that only Einstein@Home thinks about porting their code. Loads of potential to tab, but I guess Folding has had the important first step. There are probably many people crunching for folding that have never been involved with distributed computing and just attached their ps3. they won't be easy to convince to switch to boinc. but at least some people could and seeing this console's crunching power, it seems to be worth giving it a go for BOINC too.

---

Here's an interesting article about PS3's and Folding@Home: An Idle Computer Is the Lord's Workshop

....
....
....
....
....Further, a computer considered "obsolete" by almost any other standard can still provide CPU cycles so long as there's an Internet hook-up. If you can play solitaire on your machine, you can help cure Alzheimer's.

Working out a partnership with Sony was fine, but for Folding@home and the other companies that use distributed contributing the real goal must be enlisting more computer users. Currently Folding@home has a base of 200,000 computers, which is to say there over 800 million computers not in its base - although other projects are using home computers for distributed computing for everything from advancing nanotechnology to detecting extraterrestrial intelligence.

Companies that rely on distributed computing need to make a real advertising push, as do CPU makers like Intel, IBM, and AMD. They, too, could do well by doing good in encouraging people to buy more powerful (and expensive) processors.

---

Join TeamACC

Sometimes I think we are alone in the universe, and sometimes I think we are not. In either case the idea is quite staggering.

the PS3 power has flopped as most owners and installers switch their PS3s off or use them for gaming. PS3 power is at 400 teraflops which still is more powerful than IBM's "Blue Gene"

---

One problem that would concern me was at least with the earlier PSx's the optical drives didn't like baking away inside the enclosure. I had a couple fail when I used to leave them on all for extended periods (read as I that didn't care for the limitations for certain games when you saved to the card).

Once I stopped doing that, I stopped having trouble with them.

Of course now that they're hawking the other things you can use the console for, you would think they addressed that in PS3.

Alinator

Hi folks,

just a note:

NEW: Version 3.2 alpha, Cell support, MPI

http://www.fftw.org/

So whoever did the test build of the s@h app for PS3 running yellow dog might try again ;)

---

Join BOINC United now!

Actually doesn't look so impressive, compared with the same graph for P4 3600!
Why are the flops so much higher for very large arrays?

I think recoding boinc and the seti app will be far from trivial, mainly because each spu is limited to 256k program code optimised for simd, so I don't think it could run 7 instances of the app.

I think it more likely that it would run as 2 instances one the PPE, with those instances sharing available spus in parallel. I got the Cell SDK yesterday and emulator, and hope to find some time to play with it.

Anyone who can crack S@H on the PS3/QS20 will get worldwide fame, everlasting adulation and a lifetime supply of nubile girls... :-)

Actually doesn't look so impressive, compared with the same graph for P4 3600!
Why are the flops so much higher for very large arrays?

I suspect the reason that the FLOPS are higher for very large arrays is because the shorter FFTs don't have enough work involved to distribute effectively across the SPEs. Interestingly enough, very large FFTs tend to be where conventional desktop processor performance drops, since the data falls out of L2 cache once the arrays get too large.

In any case, performance on large arrays (i.e. those with 16384-131072 elements) is what makes a larger difference to SETI. Here, FFTW on Cell is about 3x faster than a P4 and 50% faster than a single 3.0 GHz Woodcrest core. Still, these numbers don't seem as good as the ones that I recall seeing in other Cell literature for hand-tuned FFT implementations, so perhaps there's more room for improvement.