Message boards :
Number crunching :
Just deleting the flops in app_info.xml file and fine?
Message board moderation
Author | Message |
---|---|
Sutaru Tsureku Send message Joined: 6 Apr 07 Posts: 7105 Credit: 147,663,825 RAC: 5 |
I have an app_info.xml file with flops entries. This values are the same which are displayed at 'average processing rate' on the 'application details' overview. E.g. above mentioned overview: 9.86 GFLOPS flops entry in app_info.xml file: 9.86e09 I can just delete all flops entries in the app_info.xml file and it's fine? Or the estimated times for the already downloaded WUs in BOINC will go 'crazy' (increase or decrease?)? Finally the WUs will end with ('-177') errors? Thanks. |
Josef W. Segur Send message Joined: 30 Oct 99 Posts: 4504 Credit: 1,414,761 RAC: 0 |
I have an app_info.xml file with flops entries. If you delete the flops, the BOINC client will substitute flops based on your CPU Whetstone benchmark. Your host currenly shows "Measured floating point speed 1800 million ops/sec", and that 1.8 GFLOPS will be used for CPU app versions. For GPU app versions the flops will be 10x that, 18 GFLOPS. The flops value you gave was for SETI@home v7 (anonymous platform, CPU), so removing it would make any such tasks already on the system have estimates 5.478 times longer than before the change. There's nothing dangerous about that, it would just mean the host wouldn't want to download any more CPU tasks until most of those were finished. Your Application details currently show these APR values: SETI@home v7 (anonymous platform, CPU) Average processing rate 9.86 GFLOPS SETI@home v7 (anonymous platform, NVIDIA GPU) Average processing rate 20.54 GFLOPS SETI@home v7 (anonymous platform, Intel GPU) Average processing rate 7.55 GFLOPS AstroPulse v7 (anonymous platform, CPU) Average processing rate 22.70 GFLOPS AstroPulse v7 (anonymous platform, NVIDIA GPU) Average processing rate 75.91 GFLOPS AstroPulse v7 (anonymous platform, Intel GPU) Average processing rate 27.08 GFLOPS The shift from 22.70 GFLOPS to 1.8 GFLOPS for the AstroPulse v7 (anonymous platform, CPU) tasks you have would probably cause those tasks to be processed at high priority. The host doesn't show any SETI@home v7 (anonymous platform, Intel GPU) tasks in progress, that's the only case where the change in flops would reduce the allowed maximum runtime. But the maximum allowed is 20x the estimate for SETI@home v7 tasks, so even if there were some the 7.55/18 ~= 0.42x factor change in the allowed maximum would be unlikely to cause an error. Joe |
jason_gee Send message Joined: 24 Nov 06 Posts: 7489 Credit: 91,093,184 RAC: 0 |
That's definitely Joe's department :D "Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions. |
©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.