Message boards :
Number crunching :
Problem can't download cpu wu's, gpu's wu's are fine though[solved]
Message board moderation
Author | Message |
---|---|
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13961 Credit: 208,696,464 RAC: 304 ![]() ![]() |
I disabled the settings for gpus(ati & nvidia) and I'm now getting cpu work and I also got some gpu work too, strange. When a lot of shorties start coming through, the manager notices the GPU is way short of work to do, so it keeps trying for GPU work only. After a while the amount of CPU work ready to run declines, and it's work requirement priority increases so it asks for both CPU & GPU work. After a while both the CPU & GPU have similar amounts of shorties ready to run, so the work fetch priorities are similar & it will maintain both levels at the serverside limits. Setting a lower number of days to cache figure also reduces the effect of requesting GPOU work only when lots of shorties start going through the system. Grant Darwin NT |
![]() ![]() Send message Joined: 26 May 99 Posts: 9959 Credit: 103,452,613 RAC: 328 ![]() ![]() |
Is it the new 580 you are using? If so how quickly does it crunch tasks? How many instances are you running My less powerful not overclocked or water cooled GTX560 is doing one every 8-10 minutes, running two instances. So as you get you 100 GPU units it is already finishing and asking for more. Personally I don't have any problems but am running Boinc 7.0.52 PS anonymous platform is supported,. |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13961 Credit: 208,696,464 RAC: 304 ![]() ![]() |
I'm still running v6.10.58 Same here- 6.10.58 on one system & 6.10.60 on the other. Grant Darwin NT |
Grant (SSSF) Send message Joined: 19 Aug 99 Posts: 13961 Credit: 208,696,464 RAC: 304 ![]() ![]() |
My less powerful not overclocked or water cooled GTX560 is doing one every 8-10 minutes, running two instances. My factory overclock GTX-560Ti knocks over a shortie in 2:50-3:10 (depending on the WU) running 2 at a time. Grant Darwin NT |
![]() Send message Joined: 22 Jul 99 Posts: 768 Credit: 24,140,697 RAC: 0 ![]() |
You might want to run with <sched_op_debug> Details of scheduler RPCs; also shows deferral intervals and other low info. New in 5.10.24 Syntax is: <sched_op_debug>1</sched_op_debug> (1 is on, 0 is off) This flag is more useful to me than the work fetch debug one which you tried. I run with it all the time - shows just how much work is being requested. Another Fred Support SETI@home when you search the Web with GoodSearch or shop online with GoodShop. ![]() |
![]() ![]() Send message Joined: 26 May 99 Posts: 9959 Credit: 103,452,613 RAC: 328 ![]() ![]() |
Seem to be quite short at the moment my last 2 on the 560 were 2.33 |
bill Send message Joined: 16 Jun 99 Posts: 861 Credit: 29,352,955 RAC: 0 ![]() |
Seems this issue is solved. But you have not said which of the offered solutions was the fix for your problem. It would be of immense value to the next person looking for an answer to the problem if they're suffering from it. |
©2025 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.