Hi.

This figures was taken from Folding@home a couple of minutes ago:

*** COPY ***
Client statistics by OS


OS Type Native TFLOPS* x86 TFLOPS* Active CPUs Total CPUs
Windows 239 239 251423 2742051
Mac OS X/PowerPC 4 4 5315 128592
Mac OS X/Intel 21 21 6856 91988
Linux 51 51 30160 399605
ATI GPU 925 976 9067 70269
NVIDIA GPU 1796 3790 15096 121800
PLAYSTATION®3 893 1884 31666 811818
Total 3929 6965 349583 4366123

*** COPY ***

Can this be true???

Rosetta@home: 86 TFLOPS
Folding@home: 6965 TFLOPS

Obviously, the main difference is the GPU and PS3 performance... but with normal processors they have more than 300 TFLOPS as well!.

This is the link: http://fah-web.stanford.edu/cgi-bin/main.py?qtype=osstats

Too much computing power I think... :S

---

Folding@Home is as old as SETI Classic (If I remember right)

---

Yep, when it comes to processing power, Folding@home is hard to beat.

But from what I know of both projects, I still prefer the science of Rosetta@home (which is why I crunch here).

You can also look at it this way: A CPU has a higher marginal utility on Rosetta@home because they have less total processing power (that's true unless we're talking about GPUs or PS3s, which Rosetta@home can't harness (yet?)). In other words, each additional cruncher contributes a bigger slice of the pie at Rosetta@home than at folding@home, so each new cruncher makes a bigger difference to the project.

---

From the link I see they report roughly 300,000 active CPU's which produce 220 TFLOPS. This does not include PS3's and GPU's. Rosetta can't touch that. Rosetta only has 80,000 active hosts(CPU'S) Another slight difference is that Folding@home defines an active cpu as a CPU which has returned work within 50 days. For Rosetta that period is 30 days.

id like to know if there is any speculation what could be accomplished with 800,000 hosts or 8 million and in what time frame ........ folding@home has been in the pflops for some time now.. when i started with Rosetta it was 27-29 tflops (2006)

From the link I see they report roughly 300,000 active CPU's which produce 220 TFLOPS. This does not include PS3's and GPU's. Rosetta can't touch that. Rosetta only has 80,000 active hosts(CPU'S) Another slight difference is that Folding@home defines an active cpu as a CPU which has returned work within 50 days. For Rosetta that period is 30 days.

id like to know if there is any speculation what could be accomplished with 800,000 hosts or 8 million and in what time frame ........ folding@home has been in the pflops for some time now.. when i started with Rosetta it was 27-29 tflops (2006)

I think alot of that would depend on if the project could support all those new machines or not. A while back Seti had sooo many users they were actually reissuing the exact same units over and over and over just to keep the crunchers happy. There was pandemonium when people found out they were just wasting there time! Seti lost alot of credibility and crunchers because of it. And I am talking SINCE Boinc, not just before Boinc when they did it ALOT!!!

---

From the link I see they report roughly 300,000 active CPU's which produce 220 TFLOPS. This does not include PS3's and GPU's. Rosetta can't touch that. Rosetta only has 80,000 active hosts(CPU'S) Another slight difference is that Folding@home defines an active cpu as a CPU which has returned work within 50 days. For Rosetta that period is 30 days.

id like to know if there is any speculation what could be accomplished with 800,000 hosts or 8 million and in what time frame ........ folding@home has been in the pflops for some time now.. when i started with Rosetta it was 27-29 tflops (2006)

I think alot of that would depend on if the project could support all those new machines or not. A while back Seti had sooo many users they were actually reissuing the exact same units over and over and over just to keep the crunchers happy. There was pandemonium when people found out they were just wasting there time! Seti lost alot of credibility and crunchers because of it. And I am talking SINCE Boinc, not just before Boinc when they did it ALOT!!!

well yes assuming the project can handle the hosts.....i wonder how many
it can handle now

From the link I see they report roughly 300,000 active CPU's which produce 220 TFLOPS. This does not include PS3's and GPU's. Rosetta can't touch that. Rosetta only has 80,000 active hosts(CPU'S) Another slight difference is that Folding@home defines an active cpu as a CPU which has returned work within 50 days. For Rosetta that period is 30 days.

id like to know if there is any speculation what could be accomplished with 800,000 hosts or 8 million and in what time frame ........ folding@home has been in the pflops for some time now.. when i started with Rosetta it was 27-29 tflops (2006)

I think alot of that would depend on if the project could support all those new machines or not. A while back Seti had sooo many users they were actually reissuing the exact same units over and over and over just to keep the crunchers happy. There was pandemonium when people found out they were just wasting there time! Seti lost alot of credibility and crunchers because of it. And I am talking SINCE Boinc, not just before Boinc when they did it ALOT!!!

well yes assuming the project can handle the hosts.....i wonder how many
it can handle now

I wish I had the resources to find out!

---

I currently use my cpu for Rosetta, and my gpu for F@H. Seems to work out ok. Right now I'm running my cpu's for folding@home, because I can't get any work for my cpu's. Finished my projects yesterday and never got any more. Oh well.

Hi.

Can this be true???

Rosetta@home: 86 TFLOPS
Folding@home: 6965 TFLOPS

Obviously, the main difference is the GPU and PS3 performance... but with normal processors they have more than 300 TFLOPS as well!.

This is the link: http://fah-web.stanford.edu/cgi-bin/main.py?qtype=osstats

Too much computing power I think... :S

Folding@home has more computing power probably than every other DC project combined as well as being the best funded. It is important though to remember that it is also trying to solve a very important computationally intensive problem as well. IMHO both projects are among the most worthly projects out there and deserve computing power. If it was not for the stupid Berkeley Stanford rivalry F@H would probably even offer BOINC credits (that would rule). I would recommend using any gpu and ps3 resources you have on F@H as the F@H is light years ahead of BOINC using these resources effectively. Again IMHO rosetta deserves all the general cpu (regular computer) you can give it. Personally I dedicate a 8800GT to F@H, about %33 of my computer cores to Rosetta, about %50 to WCG (having many projects to help gets the nod slightly) and the rest to Einstein (probably less likely to help mankind in short term but gravity wave astronomy will be a major revolution to science when figured out).

http://boincstats.com/stats/project_graph.php?pr=rosetta

Hi.

Can this be true???

Rosetta@home: 86 TFLOPS
Folding@home: 6965 TFLOPS

Obviously, the main difference is the GPU and PS3 performance... but with normal processors they have more than 300 TFLOPS as well!.

This is the link: http://fah-web.stanford.edu/cgi-bin/main.py?qtype=osstats

Too much computing power I think... :S

Folding@home has more computing power probably than every other DC project combined as well as being the best funded. It is important though to remember that it is also trying to solve a very important computationally intensive problem as well. IMHO both projects are among the most worthly projects out there and deserve computing power. If it was not for the stupid Berkeley Stanford rivalry F@H would probably even offer BOINC credits (that would rule). I would recommend using any gpu and ps3 resources you have on F@H as the F@H is light years ahead of BOINC using these resources effectively. Again IMHO rosetta deserves all the general cpu (regular computer) you can give it. Personally I dedicate a 8800GT to F@H, about %33 of my computer cores to Rosetta, about %50 to WCG (having many projects to help gets the nod slightly) and the rest to Einstein (probably less likely to help mankind in short term but gravity wave astronomy will be a major revolution to science when figured out).

I currently use my cpu for Rosetta, and my gpu for F@H. Seems to work out ok. Right now I'm running my cpu's for folding@home, because I can't get any work for my cpu's. Finished my projects yesterday and never got any more. Oh well.

I decided to try this out, just to play with it. And keep in mind that my gpu is an onboard unit. Now I know that a lot of people will think that it will take away from my Rosetta crunching, but I have found the opposite to be true. The only time Rosetta is not running full bore, other than browsing, movie watching, ect., is when I bring up the Folding@home viewer/display. I installed the high performance GPU client that has a viewer for Nvidia cuda devices. Here are some screen shots of my task manager showing the differences.
The first screenshot is how it looks most of the time. Folding is taking up memory room, but not imposing on the main cpu at all. The second shot shows Folding using 1% of cpu power, and that only lasts for a second. The first screenshot is what is most dominant. The third screenshot is what happens when I put the Folding@home viewer/display up. And I have checked this thing over time, the Folding work is progressing, even without the display up. What I have concluded is that Folding can run on my onboard gpu, which is a cuda device, and won't bother my cpu as long as I'm not displaying the graphics. And if anyone ever wondered, displaying the Rosetta graphics uses about 8% of available cpu power. (That's why I never used any of the grapics as screensavers. Too wasteful. Screensavers are silly alternative to just turning the monitor off.) The only graphics issues I have seen had to do with running the Asus PC Probe. My son was able to watch his Pokemon cartoons online with no problem.

---

I currently use my cpu for Rosetta, and my gpu for F@H. Seems to work out ok. Right now I'm running my cpu's for folding@home, because I can't get any work for my cpu's. Finished my projects yesterday and never got any more. Oh well.

I decided to try this out, just to play with it. And keep in mind that my gpu is an onboard unit. Now I know that a lot of people will think that it will take away from my Rosetta crunching, but I have found the opposite to be true. The only time Rosetta is not running full bore, other than browsing, movie watching, ect., is when I bring up the Folding@home viewer/display. I installed the high performance GPU client that has a viewer for Nvidia cuda devices. Here are some screen shots of my task manager showing the differences.
The first screenshot is how it looks most of the time. Folding is taking up memory room, but not imposing on the main cpu at all. The second shot shows Folding using 1% of cpu power, and that only lasts for a second. The first screenshot is what is most dominant. The third screenshot is what happens when I put the Folding@home viewer/display up. And I have checked this thing over time, the Folding work is progressing, even without the display up. What I have concluded is that Folding can run on my onboard gpu, which is a cuda device, and won't bother my cpu as long as I'm not displaying the graphics. And if anyone ever wondered, displaying the Rosetta graphics uses about 8% of available cpu power. (That's why I never used any of the grapics as screensavers. Too wasteful. Screensavers are silly alternative to just turning the monitor off.) The only graphics issues I have seen had to do with running the Asus PC Probe. My son was able to watch his Pokemon cartoons online with no problem.

Lucky you, man. There is not a Mac or Linux version... :(

---

Sorry dude.

I have found that glitchiness occurs when I try to watch "Miami Vice" online, small player or fullscreen. I'm going to see what I can do to overclock my gpu, but we are talking about an onboard one here. In the future if I do build the computer I want, with an Phenom II X4 955(make you jealous Chilean) I will get a dedicated video card, maybe even a dual gpu one like one of those Nvidia GTX 295's. (If you compare all the stats, stream processors per core, price...the GTX 295, while pricey, is definately a better deal than the Nvidia Quadro's and Tesla's. Just my humble opinion.)

---