Forum > GPU crunching
AK V8 + CUDA MB team work mod
Mach:
Having a small problem...I'm pretty sure I'm using v8b (the one called 'more clever autokill mod'?). Anyway, for the most part, everything runs great. However, when boinc downloads newer wu's with a more immediate date, boinc switches out one of the currently running wu's with the new one in high priority mode (in my case, this seems to almost always be the wu currently in GPU). When this happens, the new wu is not placed in the GPU and I end up having all 3 wu's processed by my 2 CPU cores. :o I can remedy the situation by shutting down boinc and restarting, but this isn't always possible (ie. the middle of the night). Any ideas on a possible fix?
*Edit* Just remembered seeing the 'ignore boinc suspend request' patch...gonna give that a try and see if it makes a difference.
Raistmer:
--- Quote from: Duke of Guinness on 20 Feb 2009, 11:12:11 am ---Thanks Raistmer!!!
--- End quote ---
Welcome :)
--- Quote from: Mach on 20 Feb 2009, 02:15:57 pm ---Having a small problem...I'm pretty sure I'm using v8b (the one called 'more clever autokill mod'?). Anyway, for the most part, everything runs great. However, when boinc downloads newer wu's with a more immediate date, boinc switches out one of the currently running wu's with the new one in high priority mode (in my case, this seems to almost always be the wu currently in GPU). When this happens, the new wu is not placed in the GPU and I end up having all 3 wu's processed by my 2 CPU cores. :o I can remedy the situation by shutting down boinc and restarting, but this isn't always possible (ie. the middle of the night). Any ideas on a possible fix?
--- End quote ---
Check what CPU based app you use, check what checkpoint interval you use (CPU app will check if GPU is idle only after checkpoint completion so big checkpoint nterval will delay idle GPU detection).
Mach:
--- Quote ---Check what CPU based app you use, check what checkpoint interval you use (CPU app will check if GPU is idle only after checkpoint completion so big checkpoint nterval will delay idle GPU detection).
--- End quote ---
I'm using the app_info.xml file supplied with your v8a package...as far as I can tell, it calls AK_V8b_win_SSE3_AMD_GPU_CPU_team.exe which then distributes to either the ap_5.00r103_SSE3.exe or MB_6.08_mod_CPU_team_CUDA.exe based on GPU availability. The one exception is when it calls ap_5.00r103_SSE3.exe directly for version 500 AP WUs.
I can't find the checkpoint interval, but the GPU usually completes MB WUs in approx 5 mins and picks up the next WU without problem. When the problem occurs, all 3 WUs continue on CPU until one is completed (not 100% sure if the next WU starts on the GPU correctly or not as I usually end up shutting down and restarting boinc). Definitely continues longer than 10-20 minutes without switching to the GPU in any case.
As I mentioned in my edited post above, I've added the 'ignore boinc suspend' patch. Will give an updat as to whether this works after I have some time to monitor behavior.
Raistmer:
Look this thread (few pages back) for CPU app update (only for SSSE3 version) named GPU_watch_update_for_SSSE3.rar
Mach:
That mod sounds exactly what I need in this case. Unfortunately, my CPU only supports the SSE3 instruction set and not SSSE3 :( Oh well, guess I'll just have to live with the occasional idle GPU for the time being. Thanks for the prompt responses above, Raistmer ;D
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version