Six native petaFLOPS
Posted: Thu Nov 10, 2011 10:55 pm
I noticed today that Folding@home finally has enough computing power to cross over the 6 native petaFLOP barrier. I find this quite impressive, and wanted to point it out. From the OS stats page I saw this:
This accomplishment really is remarkable considering that according to Dr. Pande's blog the 5 petaFLOP milestone was passed in February of 2009. And as you can see we have about 7.9 x86 petaFLOPS, from about 447,000 active CPUs, about 35,000 active GPUs, and about 6,000 active PS3s. I find it quite amazing that such a large, powerful, and diverse system can be voluntarily amassed and utilized.
For some FLOP-to-FLOP comparisons:
Other DC projects: As of today, the BOINC stats reports that the combined efforts of all distributed computing projects using the BOINC middleware add up to 5.262 petaFLOPS. It is my understanding that BOINC is mostly CPU-based, so it speaks well for them that they can assemble that much computing power. There are so many BOINC projects that are greatly assist researchers in many different scientific areas. For some examples, PrimeGrid itself currently sits at about 1.4 petaFLOPS, (they use GPUs too apparently) with Milkyway@home at 0.578 petaFLOPS, and Collatz Conjecture at 0.545 petaFLOPS, and there are many other BOINC-based projects that are focused on various problems. I'm glad that Folding@home ensures data integrity using methods other than the "quorum" approach, even though the client has to be closed-source to do so.
Supercomputers: Although DC projects have a vastly different architecture than supercomputers, as they were measured in June of 2011, Japan's K supercomputer stands at 8.162 petaFLOPS, and China's Tianhe-1A supercomputer runs at 2.566 petaFLOPS. Those numbers are also quite impressive. How amazing it is that we have this kind of computing power available!
I just wanted to throw in those numbers for a sort of comparison. I hope we stay above this level, and I'm confident that v7's default configuration of SMP+GPU will greatly assist both Folding@home's FLOP numbers and its overall scientific production. How great it is that we can be a part of this massive effort!
This accomplishment really is remarkable considering that according to Dr. Pande's blog the 5 petaFLOP milestone was passed in February of 2009. And as you can see we have about 7.9 x86 petaFLOPS, from about 447,000 active CPUs, about 35,000 active GPUs, and about 6,000 active PS3s. I find it quite amazing that such a large, powerful, and diverse system can be voluntarily amassed and utilized.
For some FLOP-to-FLOP comparisons:
Other DC projects: As of today, the BOINC stats reports that the combined efforts of all distributed computing projects using the BOINC middleware add up to 5.262 petaFLOPS. It is my understanding that BOINC is mostly CPU-based, so it speaks well for them that they can assemble that much computing power. There are so many BOINC projects that are greatly assist researchers in many different scientific areas. For some examples, PrimeGrid itself currently sits at about 1.4 petaFLOPS, (they use GPUs too apparently) with Milkyway@home at 0.578 petaFLOPS, and Collatz Conjecture at 0.545 petaFLOPS, and there are many other BOINC-based projects that are focused on various problems. I'm glad that Folding@home ensures data integrity using methods other than the "quorum" approach, even though the client has to be closed-source to do so.
Supercomputers: Although DC projects have a vastly different architecture than supercomputers, as they were measured in June of 2011, Japan's K supercomputer stands at 8.162 petaFLOPS, and China's Tianhe-1A supercomputer runs at 2.566 petaFLOPS. Those numbers are also quite impressive. How amazing it is that we have this kind of computing power available!
I just wanted to throw in those numbers for a sort of comparison. I hope we stay above this level, and I'm confident that v7's default configuration of SMP+GPU will greatly assist both Folding@home's FLOP numbers and its overall scientific production. How great it is that we can be a part of this massive effort!