| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: One Wonders...... |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 52
|
|
| Author |
|
|
knreed
Former World Community Grid Tech Joined: Nov 8, 2004 Post Count: 4504 Status: Offline Project Badges: 
|
Pretty sure the 6870 is not double point, but time will tell. Correct - when I was testing other projects on my gpu I was not allowed at Milkyway. However, I could run HCC1 so you do not need double point. |
||
|
|
nanoprobe
Master Cruncher Classified Joined: Aug 29, 2008 Post Count: 2998 Status: Offline Project Badges: 
|
Correct - when I was testing other projects on my gpu I was not allowed at Milkyway. What possible reason could they have for that? I thought we were all in this together.
In 1969 I took an oath to defend and protect the U S Constitution against all enemies, both foreign and Domestic. There was no expiration date.
----------------------------------------  [Edit 1 times, last edit by nanoprobe at Mar 9, 2012 10:45:11 PM] |
||
|
|
mecole
Advanced Cruncher Joined: Jun 18, 2011 Post Count: 55 Status: Offline |
Correct - when I was testing other projects on my gpu I was not allowed at Milkyway. What possible reason could they have for that? I thought we were all in this together. Milkyway science requires double precision and the GPU Ken was testing with does not support it. The good news is that the HCC1 does not require double precision so a lot more different cards will be able to run this project. <thumbsup>  |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges:
|
One concern I have is whether you can keep a dedicated GPU busy. For example, I run BOINC/WCG on the CPU of a dedicated PC, but run folding@home on the Nvidia GPUs. I don't do anything else on that PC, and it runs 24/7. If I were to devote a whole GPU to WGC, can I ensure that there is always a project running that uses the GPU? If it just sits there idle at times, I would arguably be better off folding with it.
----------------------------------------That is, I would like to still use some cores for the CPU-only projects, but ensure that one core always has a GPU project running. Maybe this is premature, and the answers will be become evident, but someone should start thinking about it. [Edit 1 times, last edit by Jim1348 at Mar 10, 2012 3:51:06 AM] |
||
|
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges: 
|
Cheers WCG
---------------------------------------- Finally a great excuse to upgrade this 9500GT and buy a decent graphics card  Looking through the Gflops/W stats here http://en.wikipedia.org/wiki/Comparison_of_Nv..._units#GeForce_500_Series It looks like the GTX560 or the TI version are currently the most efficient NVidia Cards for crunching Will the new code run on both Nvidia and ATI? It's been a while since I had a half decent 3DMark score  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
One concern I have is whether you can keep a dedicated GPU busy. For example, I run BOINC/WCG on the CPU of a dedicated PC, but run folding@home on the Nvidia GPUs. I don't do anything else on that PC, and it runs 24/7. If I were to devote a whole GPU to WGC, can I ensure that there is always a project running that uses the GPU? If it just sits there idle at times, I would arguably be better off folding with it. That is, I would like to still use some cores for the CPU-only projects, but ensure that one core always has a GPU project running. Maybe this is premature, and the answers will be become evident, but someone should start thinking about it. Based on the thought of everyone doing HCC converting to doing this science on GPU only, I'd think we'll be done by end of April, if go-live were April 1. At CPU 70 minutes / GPU 7 minutes it would be going 10 times faster. That alone dictates there's huge volumes of work units required [for a GPU suitable science app] to even contemplate hosting any new GPU based project at WCG if adoption were big. HCC on GPU is certainly going to be a first reference of how fast we get this research now done. --//-- edit: spell-check [Edit 1 times, last edit by Former Member at Mar 10, 2012 9:52:06 AM] |
||
|
|
RT
Master Cruncher USA - Texas - DFW Joined: Dec 22, 2004 Post Count: 2636 Status: Offline Project Badges: 
|
Cheers WCG Finally a great excuse to upgrade this 9500GT and buy a decent graphics card Looking through the Gflops/W stats here http://en.wikipedia.org/wiki/Comparison_of_Nv..._units#GeForce_500_Series It looks like the GTX560 or the TI version are currently the most efficient NVidia Cards for crunching Will the new code run on both Nvidia and ATI? It's been a while since I had a half decent 3DMark score I would also like to hear any performance information (however sketchy) about the NVIDIA GTX5xx series cards. ---------------------------------------- [Edit 1 times, last edit by RT at Mar 10, 2012 2:54:12 PM] |
||
|
|
David Autumns
Ace Cruncher UK Joined: Nov 16, 2004 Post Count: 11062 Status: Offline Project Badges:
|
RT just rubbing my eyes in disbelief
---------------------------------------- That's a Blue Snail ..... already !!!! (Give me 2 Months  maybe less  ) |
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges:
|
Based on the thought of everyone doing HCC converting to doing this science on GPU only, I'd think we'll be done by end of April, if go-live were April 1. At CPU 70 minutes / GPU 7 minutes it would be going 10 times faster. That alone dictates there's huge volumes of work units required [for a GPU suitable science app] to even contemplate hosting any new GPU based project at WCG if adoption were big. HCC on GPU is certainly going to be a first reference of how fast we get this research now done. Thanks. I think you are saying that there will soon be enough GPU projects to keep my GPU busy, and so I don't need to worry about it. But that depends on the statistics of the adoption rate, verses the randomness of the assignments in the present scheme. So in case there aren't enough GPU projects to always keep my GPU busy (considering the randomness of the assignments), I think they should provide some option to always provide you a GPU project in preference to a CPU project. Maybe that is not politically-correct at WCG? But it would help maximize the science output. And it would free up your CPU to be used with CPU-only projects. So it is really a win-win, if they can do it. [Edit 1 times, last edit by Jim1348 at Mar 10, 2012 7:34:30 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Just bought a EVGA 570. It.screams through work on GPUgrid . Also EVGA haslifetime warranty, and their precision tool for ocing is quite nice. While booted (no restarts). Although if you have two+ slots multiple 560s would be.better price/performance.
Just.remember, these things consume a lot.of electricity and heat. Proper ventilation and PSU are.required!! |
||
|
|
|