Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
![]() |
World Community Grid Forums
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 73
|
![]() |
Author |
|
BladeD
Ace Cruncher USA Joined: Nov 17, 2004 Post Count: 28976 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
If click on the link in the OP, you would see that a ATI 5770 was used in the test and it lacks double precision.
---------------------------------------- |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Here is a nice quick reference for Boinc Compatible GPUs.
The ones in bold are Double Precision GPUs. ![]() |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
This is great news. Hopefully we get good support for our gpus and single precision gpus would be supported. I have 5 gpus and if gpu app would be stable and make good use of my gpus i would dedicate all of them to this project.
|
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
See the previous comment on the development platform used, which was not double precision capable. A very long beta is to be expected as this will set the WCG base for anything gpu ported in the future.
--//-- |
||
|
Platoon
Advanced Cruncher Russia Joined: Jun 28, 2006 Post Count: 62 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Any ideas if SLI/Crossfire configurations will be supported?
----------------------------------------
" forever forge ahead and keep the dream in sight!"
![]() |
||
|
Bearcat
Master Cruncher USA Joined: Jan 6, 2007 Post Count: 2803 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Will we have to select to opt in or will WCG send wu's if the GPU is detected to run?
----------------------------------------
Crunching for humanity since 2007!
![]() |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Any ideas if SLI/Crossfire configurations will be supported? I'm curious, what grid project support this now? |
||
|
Tapiot
Cruncher Joined: Dec 8, 2009 Post Count: 15 Status: Offline |
Any ideas if SLI/Crossfire configurations will be supported? Why you need SLI / CF support? When the project does not support SLI / CF mode, you can calculate the two different work units at the same time without slowing down the calculation, even if the other graphics card are slower than the other. Example: http://moowrap.net/forum_thread.php?id=138 |
||
|
kashie
Cruncher Joined: Mar 7, 2007 Post Count: 46 Status: Offline Project Badges: ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
Dual cards such as HD 5970 have Crossfire enabled internally and it cannot be turned off. This sometimes causes problems with OpenCL applications using 2 full CPU cores due to each GPU core using a full CPU core. It's kind of a "busy wait" thing.
The PrimeGrid ATI PPSieve OpenCL application suffers from this problem. You can assign both GPU cores to a single CPU core with programs such as Process Lasso but it's a messy workaround. Or you can use an app_info.xml file with avg_ncpus value of 1 to leave 2 CPU cores free to support OpenCL GPU processing if you have 8 CPU "cores". Those with only 4 core computers with no hyperthreading and those with 2 dual GPUs do not have the same flexibility though. If you don't do anything and the default avg_ncpus value is too low it causes your CPU to become overcommitted if you are also processing BOINC CPU tasks. This is why you need Crossfire support, to be able to run an OpenCL task on each core of dual GPUs without eating up your CPU resources. |
||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Great news! I'm looking forward to dedicating a GPU or two to the effort once you guys are ready.
----------------------------------------But I was wondering, will there be disconnect between runtime (and therefore badges awarded) and points awarded for these GPU work units vs the equivalent CPU workunits? In other words, if it takes a CPU thread 6hrs to complete a task - but it only takes a GPU 6min to complete the "same" task - will the points awarded be the same? What about runtime? It seems to me that if there's any discrepancies between CPU/GPU on runtime, it might throw the whole badging system out of whack (or am I missing something?) [Edit 1 times, last edit by Former Member at Nov 17, 2011 8:33:06 PM] |
||
|
|
![]() |