| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: GPU Support Forum
Thread: BOINC is fetching way too much work for GPU |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 5
|
|
| Author |
|
|
wujj123456
Cruncher Joined: Jun 9, 2010 Post Count: 38 Status: Offline Project Badges: 
|
I recently switched to 7.0.31 (downloaded from BOINC website) due to a bug in earlier clients that prevents idle detection under Linux. Finally, I am crunching GPU workloads from WCG while my computer is idle.
However, my BOINC client downloaded too many jobs. I only have one GPU, and a dual-core CPU. I set my work buffer to be 0.05 days, and each of my GPU job takes ~0.25hr. Even though I am pretty sure the client is using local preference, I verified my preference on WCG as well (it's actually 0.01 days of work buffer). Currently I have >65 GPU jobs sitting in the queue. That's ~16hrs worth of work... CPU part is as good as before. Here is a snippet of relevant logs. I removed download/upload messages for individual workloads. There aren't any preference changes during these periods. The beginning part is still sane, until ~7PM EST ish, server started bombing me with jobs... Thu 13 Dec 2012 02:24:04 AM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 02:24:04 AM EST | World Community Grid | Reporting 1 completed tasks Thu 13 Dec 2012 02:24:04 AM EST | World Community Grid | Requesting new tasks for CPU Thu 13 Dec 2012 02:24:06 AM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 02:24:30 AM EST | World Community Grid | Computation for task SN2S_AAW88546_0000236_0270_0 finished Thu 13 Dec 2012 03:27:31 AM EST | World Community Grid | Computation for task DSFL_00040-13_0000025_1136_0 finished Thu 13 Dec 2012 03:54:40 AM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 03:54:40 AM EST | World Community Grid | Reporting 2 completed tasks Thu 13 Dec 2012 03:54:40 AM EST | World Community Grid | Requesting new tasks for CPU Thu 13 Dec 2012 03:54:42 AM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 04:43:02 AM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 04:43:02 AM EST | World Community Grid | Requesting new tasks for CPU Thu 13 Dec 2012 04:43:04 AM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 05:20:12 AM EST | World Community Grid | Computation for task SN2S_AAW88546_0000239_0119_0 finished Thu 13 Dec 2012 06:04:29 AM EST | World Community Grid | Computation for task faah37030_ZINC06250958_xh2_xtal_00_0 finished Thu 13 Dec 2012 07:11:44 AM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 07:11:44 AM EST | World Community Grid | Reporting 2 completed tasks Thu 13 Dec 2012 07:11:44 AM EST | World Community Grid | Requesting new tasks for CPU Thu 13 Dec 2012 07:11:46 AM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 07:14:56 AM EST | World Community Grid | Computation for task GFAM_x2JKF_PfPrfln_0056652_0141_0 finished Thu 13 Dec 2012 07:33:03 AM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 07:33:03 AM EST | World Community Grid | Reporting 1 completed tasks Thu 13 Dec 2012 07:33:03 AM EST | World Community Grid | Requesting new tasks for CPU Thu 13 Dec 2012 07:33:04 AM EST | World Community Grid | Scheduler request completed: got 2 new tasks Thu 13 Dec 2012 08:59:41 AM EST | World Community Grid | Computation for task DSFL_00040-13_0000044_0078_0 finished Thu 13 Dec 2012 09:14:00 AM EST | World Community Grid | Computation for task GFAM_x2JKF_PfPrfln_0056674_0142_0 finished Thu 13 Dec 2012 10:30:13 AM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 10:30:13 AM EST | World Community Grid | Reporting 2 completed tasks Thu 13 Dec 2012 10:30:13 AM EST | World Community Grid | Requesting new tasks for CPU Thu 13 Dec 2012 10:30:15 AM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 11:01:34 AM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 11:01:34 AM EST | World Community Grid | Requesting new tasks for CPU Thu 13 Dec 2012 11:01:36 AM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 11:20:19 AM EST | World Community Grid | Restarting task X0900085170149200703061001_0 using hcc1 version 708 (nvidia_hcc1) in slot 2 Thu 13 Dec 2012 11:26:15 AM EST | World Community Grid | Computation for task faah37036_ZINC12776639_xh2_xtal_03_0 finished Thu 13 Dec 2012 11:26:15 AM EST | World Community Grid | Restarting task SN2S_AAW88546_0000269_0362_0 using sn2s version 620 in slot 1 Thu 13 Dec 2012 11:33:07 AM EST | World Community Grid | Computation for task X0900085170149200703061001_0 finished Thu 13 Dec 2012 11:45:20 AM EST | World Community Grid | Computation for task X0900085170137200703061002_0 finished Thu 13 Dec 2012 11:58:01 AM EST | World Community Grid | Restarting task X0900085170136200703061002_0 using hcc1 version 708 (nvidia_hcc1) in slot 0 Thu 13 Dec 2012 12:09:34 PM EST | World Community Grid | Computation for task X0900085170136200703061002_0 finished Thu 13 Dec 2012 12:17:50 PM EST | World Community Grid | Computation for task SN2S_AAW88546_0000269_0362_0 finished Thu 13 Dec 2012 12:17:50 PM EST | World Community Grid | Restarting task DSFL_00040-14_0000008_0713_1 using dsfl version 625 in slot 2 Thu 13 Dec 2012 12:21:08 PM EST | World Community Grid | Computation for task X0900085170138200703061002_0 finished Thu 13 Dec 2012 12:32:46 PM EST | World Community Grid | Computation for task X0900085170115200703061002_0 finished Thu 13 Dec 2012 12:45:09 PM EST | World Community Grid | Computation for task X0900085170148200703061001_0 finished Thu 13 Dec 2012 12:57:59 PM EST | World Community Grid | Computation for task X0900085170107200703061002_1 finished Thu 13 Dec 2012 01:10:18 PM EST | World Community Grid | Computation for task X0900085170104200703061002_0 finished Thu 13 Dec 2012 01:15:30 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 01:15:30 PM EST | World Community Grid | Reporting 10 completed tasks Thu 13 Dec 2012 01:15:30 PM EST | World Community Grid | Requesting new tasks for CPU and NVIDIA Thu 13 Dec 2012 01:15:34 PM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 01:21:46 PM EST | World Community Grid | Computation for task X0900085170128200703061002_0 finished Thu 13 Dec 2012 01:34:06 PM EST | World Community Grid | Computation for task X0900085170150200703061001_0 finished Thu 13 Dec 2012 01:46:20 PM EST | World Community Grid | Computation for task X0900085170126200703061002_1 finished Thu 13 Dec 2012 01:56:33 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 01:56:33 PM EST | World Community Grid | Reporting 3 completed tasks Thu 13 Dec 2012 01:56:33 PM EST | World Community Grid | Requesting new tasks for CPU and NVIDIA Thu 13 Dec 2012 01:56:39 PM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 01:56:54 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 01:56:54 PM EST | World Community Grid | Requesting new tasks for NVIDIA Thu 13 Dec 2012 01:56:56 PM EST | World Community Grid | Scheduler request completed: got 6 new tasks Thu 13 Dec 2012 01:56:56 PM EST | World Community Grid | Your preferences allow tasks from applications other than those selected Thu 13 Dec 2012 01:56:56 PM EST | World Community Grid | Sending tasks from other applications Thu 13 Dec 2012 01:58:35 PM EST | World Community Grid | Computation for task X0900085170119200703061002_0 finished Thu 13 Dec 2012 02:09:57 PM EST | World Community Grid | Computation for task DSFL_00040-14_0000008_0713_1 finished Thu 13 Dec 2012 02:10:12 PM EST | World Community Grid | Computation for task X0900085170116200703061002_1 finished Thu 13 Dec 2012 02:22:31 PM EST | World Community Grid | Computation for task X0900085170135200703061002_1 finished Thu 13 Dec 2012 02:35:11 PM EST | World Community Grid | Computation for task X0900085170102200703061002_0 finished Thu 13 Dec 2012 02:47:56 PM EST | World Community Grid | Computation for task X0900085170152200703061001_1 finished Thu 13 Dec 2012 02:59:58 PM EST | World Community Grid | Computation for task X0900085170140200703061002_0 finished Thu 13 Dec 2012 03:01:07 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 03:01:07 PM EST | World Community Grid | Reporting 7 completed tasks Thu 13 Dec 2012 03:01:07 PM EST | World Community Grid | Requesting new tasks for CPU and NVIDIA Thu 13 Dec 2012 03:01:09 PM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 03:01:25 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 03:01:25 PM EST | World Community Grid | Requesting new tasks for CPU and NVIDIA Thu 13 Dec 2012 03:01:27 PM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 03:01:42 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 03:01:42 PM EST | World Community Grid | Requesting new tasks for NVIDIA Thu 13 Dec 2012 03:01:44 PM EST | World Community Grid | Scheduler request completed: got 6 new tasks Thu 13 Dec 2012 03:01:44 PM EST | World Community Grid | Your preferences allow tasks from applications other than those selected Thu 13 Dec 2012 03:01:44 PM EST | World Community Grid | Sending tasks from other applications Thu 13 Dec 2012 03:12:45 PM EST | World Community Grid | Computation for task X0960085660874200704041056_0 finished Thu 13 Dec 2012 03:26:03 PM EST | World Community Grid | Computation for task X0960085660860200704041057_0 finished Thu 13 Dec 2012 03:40:08 PM EST | World Community Grid | Computation for task X0960085660648200704041100_1 finished Thu 13 Dec 2012 03:53:05 PM EST | World Community Grid | Computation for task X0960085660729200704041058_0 finished Thu 13 Dec 2012 03:54:51 PM EST | World Community Grid | Computation for task GFAM_x2JKF_PfPrfln_0056736_0132_2 finished Thu 13 Dec 2012 04:05:52 PM EST | World Community Grid | Computation for task X0960085660868200704041056_0 finished Thu 13 Dec 2012 04:18:05 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 04:18:05 PM EST | World Community Grid | Reporting 6 completed tasks Thu 13 Dec 2012 04:18:05 PM EST | World Community Grid | Requesting new tasks for NVIDIA Thu 13 Dec 2012 04:18:07 PM EST | World Community Grid | Scheduler request completed: got 6 new tasks Thu 13 Dec 2012 04:18:07 PM EST | World Community Grid | Your preferences allow tasks from applications other than those selected Thu 13 Dec 2012 04:18:07 PM EST | World Community Grid | Sending tasks from other applications Thu 13 Dec 2012 04:18:13 PM EST | World Community Grid | Computation for task X0960085660459200704041103_0 finished Thu 13 Dec 2012 04:30:54 PM EST | World Community Grid | Computation for task X0960085660861200704041057_1 finished Thu 13 Dec 2012 04:42:50 PM EST | World Community Grid | Computation for task X0960085660864200704041057_1 finished Thu 13 Dec 2012 04:55:53 PM EST | World Community Grid | Computation for task X0960085660837200704041056_1 finished Thu 13 Dec 2012 05:08:58 PM EST | World Community Grid | Computation for task X0960085660530200704041101_0 finished Thu 13 Dec 2012 05:26:09 PM EST | World Community Grid | Restarting task X0960085660825200704041056_1 using hcc1 version 708 (nvidia_hcc1) in slot 0 Thu 13 Dec 2012 05:28:01 PM EST | World Community Grid | Restarting task SN2S_AAW88547_0000045_0088_0 using sn2s version 620 in slot 2 Thu 13 Dec 2012 05:46:48 PM EST | World Community Grid | Computation for task CMD2_2502-1NCJ_A.clustersOccur-2V5Y_A.clustersOccur_42_517104_529415_523461_524652_524085_524274_1 finished Thu 13 Dec 2012 06:21:05 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 06:21:05 PM EST | World Community Grid | Reporting 6 completed tasks Thu 13 Dec 2012 06:21:05 PM EST | World Community Grid | Requesting new tasks for CPU and NVIDIA Thu 13 Dec 2012 06:21:08 PM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 06:40:52 PM EST | World Community Grid | Computation for task SN2S_AAW88547_0000047_0433_0 finished Thu 13 Dec 2012 06:54:58 PM EST | World Community Grid | Restarting task X0960085660825200704041056_1 using hcc1 version 708 (nvidia_hcc1) in slot 0 Thu 13 Dec 2012 07:08:33 PM EST | World Community Grid | Computation for task X0960085660825200704041056_1 finished Thu 13 Dec 2012 07:14:47 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 07:14:47 PM EST | World Community Grid | Reporting 2 completed tasks Thu 13 Dec 2012 07:14:47 PM EST | World Community Grid | Requesting new tasks for CPU and NVIDIA Thu 13 Dec 2012 07:14:49 PM EST | World Community Grid | Scheduler request completed: got 1 new tasks Thu 13 Dec 2012 07:21:28 PM EST | World Community Grid | Computation for task X0960085660845200704041057_0 finished Thu 13 Dec 2012 07:24:38 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 07:24:38 PM EST | World Community Grid | Reporting 1 completed tasks Thu 13 Dec 2012 07:24:38 PM EST | World Community Grid | Requesting new tasks for NVIDIA Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | Scheduler request completed: got 16 new tasks Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | Your preferences allow tasks from applications other than those selected Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | Sending tasks from other applications Thu 13 Dec 2012 07:24:57 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 07:24:57 PM EST | World Community Grid | Requesting new tasks for NVIDIA Thu 13 Dec 2012 07:24:59 PM EST | World Community Grid | Scheduler request completed: got 30 new tasks Thu 13 Dec 2012 07:24:59 PM EST | World Community Grid | Your preferences allow tasks from applications other than those selected Thu 13 Dec 2012 07:24:59 PM EST | World Community Grid | Sending tasks from other applications Thu 13 Dec 2012 07:35:47 PM EST | World Community Grid | Computation for task X0960085660598200704041101_1 finished Thu 13 Dec 2012 07:41:46 PM EST | World Community Grid | Computation for task SN2S_AAW88547_0000045_0088_0 finished Thu 13 Dec 2012 07:41:46 PM EST | World Community Grid | Restarting task DSFL_00040-14_0000028_0548_1 using dsfl version 625 in slot 1 |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi, there are a number of threads on this and we thought it was actually cured with 7.0.31 and up but, I'm not sure if it is the same issue.
----------------------------------------Seeing that the client is reporting results in groups of 6 and 7, such as this sample Thu 13 Dec 2012 01:58:35 PM EST | World Community Grid | Computation for task X0900085170119200703061002_0 finished Thu 13 Dec 2012 02:09:57 PM EST | World Community Grid | Computation for task DSFL_00040-14_0000008_0713_1 finished Thu 13 Dec 2012 02:10:12 PM EST | World Community Grid | Computation for task X0900085170116200703061002_1 finished Thu 13 Dec 2012 02:22:31 PM EST | World Community Grid | Computation for task X0900085170135200703061002_1 finished Thu 13 Dec 2012 02:35:11 PM EST | World Community Grid | Computation for task X0900085170102200703061002_0 finished Thu 13 Dec 2012 02:47:56 PM EST | World Community Grid | Computation for task X0900085170152200703061001_1 finished Thu 13 Dec 2012 02:59:58 PM EST | World Community Grid | Computation for task X0900085170140200703061002_0 finished Thu 13 Dec 2012 03:01:07 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 03:01:07 PM EST | World Community Grid | Reporting 7 completed tasks It only fetching new work one by one, 25-30 tasks being completed until the large assignments starts [don't know if there was a shortage before, as your config log flag prefs do not specify to print how much work is really asked for]. Then this line appears which I can't remember having seen before, user friendly no doubt: Thu 13 Dec 2012 07:24:59 PM EST | World Community Grid | Your preferences allow tasks from applications other than those selected This one suggests that there was in fact a shortage of your regular project selection, so something else is send, the "if there is no work for your selections, send something else", at which point there are these task floods, getting 16 and 30 [when the official "per assignment" limit is 15 per work fetch]. The only reason I can think of is that the Duration Correction Factor has gone off. Open BOINC Manager in the project tab, select WCG, and hit the properties button on left. In that screen there is a scheduling section. Tell us what the values in that section. Mind you, 16 hours computing time buffer is still small compared to what others have experienced. They got so much work that the client could never finish it all in time. Now, the good news is, that an up and coming client currently actively tested, is going having additional controls on request of WCG, which hopefully are going to stop these floods for sure. For the moment no solution to your and many's a problem, but it is coming. [Edit 1 times, last edit by Former Member at Dec 14, 2012 9:52:12 AM] |
||
|
|
wujj123456
Cruncher Joined: Jun 9, 2010 Post Count: 38 Status: Offline Project Badges:
|
Thanks for your info. That's very useful. Indeed, WCG reports that no tasks are available for other projects. But those projects don't have GPU workload at all... I guess I shouldn't have truncated this part. Same thing happened with the 30 job batch and another 19 job batch last night.
Thu 13 Dec 2012 07:24:38 PM EST | World Community Grid | Sending scheduler request: To fetch work. Thu 13 Dec 2012 07:24:38 PM EST | World Community Grid | Reporting 1 completed tasks Thu 13 Dec 2012 07:24:38 PM EST | World Community Grid | Requesting new tasks for NVIDIA Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | Scheduler request completed: got 16 new tasks Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | No tasks are available for the applications you have selected Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | No tasks are available for Say No to Schistosoma Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | No tasks are available for GO Fight Against Malaria Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | No tasks are available for Drug Search for Leishmaniasis Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | No tasks are available for Help Cure Muscular Dystrophy - Phase 2 Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | No tasks are available for Help Fight Childhood Cancer Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | No tasks are available for FightAIDS@Home Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | Your preferences allow tasks from applications other than those selected Thu 13 Dec 2012 07:24:41 PM EST | World Community Grid | Sending tasks from other applications I indeed have "If there is no work available for my computer for the projects I have selected above, please send me work from another project" selected in my preference. From your description, looks like I should turn it off for now. I subscribe to most of the projects, and it's not likely that I will be short of tasks. Will that affect scheduler behavior in this case? Btw, my Duration Correction Factor is 1.000. Seems to be sane. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
A DCF of 1.0 is crazy sane. The only time this value would exist is after a clean install, else a chance in a million or more.
Deselecting the "If there is no..." and having enough other CPU sciences and a little bit of a buffer to bridge the moments there is no work for the choice projects, you are 99+% of the times fine. I've not ticked the box, usually running 4-5 way mix. The log you posted actually tried to fetch GPU work for all active sciences you've selected, except of course there's no other but HCC. It's not clear though as HCC is not mentioned. |
||
|
|
wujj123456
Cruncher Joined: Jun 9, 2010 Post Count: 38 Status: Offline Project Badges:
|
Yes, the client is freshly installed two days ago. Deselecting "If there is no..." option seems to fix the problem. I no longer get tons of GPU jobs.
I didn't post those download messages for HCC GPU workload. All fetched work in those logs were HCC GPU work. Thank you very much for your help. I am very happy with the current result. Hopefully next release will thorouly fix this problem so that others with that option selected won't be flooded with GPU jobs. |
||
|
|
|