Message boards :
Number crunching :
Folding@Home using GPUs
Message board moderation
Author | Message |
---|---|
[STS]LoB Send message Joined: 23 May 99 Posts: 53 Credit: 3,375,888 RAC: 144 |
http://folding.stanford.edu/FAQ-ATI.html In short, Folding@home uses ATI's X1900 GPUs to better fold proteins. (Increase of calculations by 20 or 30 times...) Would that technique be able to help SETI@home as well, or ar the calculations of a completely different nature? |
Marco Feindler Send message Joined: 10 Aug 00 Posts: 12 Credit: 2,602 RAC: 0 |
yes this would be a big improvement live fast, die young |
Bruno G. Olsen & ESEA @ greenholt Send message Joined: 15 May 99 Posts: 875 Credit: 4,386,984 RAC: 0 |
GPU's aren't as uniform as CPU, as far as I know, so you'd have to target a huge bunch of GPU's just to cover a small percentage of participants hosts. The man hours needed to create this code vs. the amount of angry users that won't be covered by this with the potential loss of support for the project becuase of this (we all know how little it takes for people here to get angry), makes the efford potentially not worth while |
Matthew Johnson Send message Joined: 11 Dec 05 Posts: 92 Credit: 5,482 RAC: 0 |
GPU's aren't as uniform as CPU, as far as I know, so you'd have to target a huge bunch of GPU's just to cover a small percentage of participants hosts. The man hours needed to create this code vs. the amount of angry users that won't be covered by this with the potential loss of support for the project becuase of this (we all know how little it takes for people here to get angry), makes the efford potentially not worth while this has been discussed before .. berkley dont have the money like standford do to pay coders to code for GPU;s.. and the maths isnt the same ... the interger unit is slightly out and rounds down numbers not working them out properly |
[STS]LoB Send message Joined: 23 May 99 Posts: 53 Credit: 3,375,888 RAC: 144 |
First of all: Sorry for posting this allthough it had allready been discussed. Didn't know that. Regarding coding effort and too few manpower: SETI@home has a huge userbase, each of these users have their talents, their should accidently be some who could (and perhaps would) code something like that in their freetime. I mean, people also compile own optimised versions etc... |
Benher Send message Joined: 25 Jul 99 Posts: 517 Credit: 465,152 RAC: 0 |
I've downloaded some open source FFT GPU code. I have emailed a bit with the authors. I may get to testing it in seti eventually. Regular optimization first. |
www.WowTattoos.com (≈ #275/1.1mil Users In Average Credit) Send message Joined: 16 Nov 00 Posts: 2 Credit: 16,811,496 RAC: 0 |
What about using something like the PPU (Physics Processing Unit) in an Ageia PhysX accelerator card. I doubt it rounds down numbers. And it seems like something dedicated like that might put out some serious Seti@Home work units. If someone could code something for that, my PhysX card would get some use while we wait for more games to come out that support it. I could provide hardware for a coder that's seriously interested. |
Benher Send message Joined: 25 Jul 99 Posts: 517 Credit: 465,152 RAC: 0 |
If you want the code written...just send your card to _____ _____ ___USA, 99999 |
-= Vyper =- Send message Joined: 5 Sep 99 Posts: 1652 Credit: 1,065,191,981 RAC: 2,537 |
If you want the code written...just send your card to Heh, but if that's possible i could donate spare time to test it 4 u.. The only problem is that i'm in a possesion of a 7900GT series GPU and isn't based on ATI's latest offering... Contact me at Simons (Chickens) place , i'm registered there.. Kind Regards Vyper _________________________________________________________________________ Addicted to SETI crunching! Founder of GPU Users Group |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.