| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: [resolved] Boinc 7 not requesting tasks |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 5
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi,
----------------------------------------this is a 4-thread Windows 7 32-bit machine with Boinc 7.0.64 and BoincTasks. Boinc is right now running one Einstein GPU task (0.2 CPUs +1 NVIDIA) plus one WUProp (non-CPU-intensive). CPU slots have all run dry. I have set WCG, Albert@home and LHC@home to no new tasks for now. WUProp, Einstein@home and Constellation are set to fetch. Constellation is set to a profile with no available work. Any idea why Boinc won't request new CPU work for Einstein ? I don't know if Boinc is playing tricks on me or i've gotten entirely stupid. 1 06-09-2013 06:20 No config file found - using defaults 2 06-09-2013 06:20 Starting BOINC client version 7.0.64 for windows_intelx86 3 06-09-2013 06:20 log flags: file_xfer, sched_ops, task 4 06-09-2013 06:20 Libraries: libcurl/7.25.0 OpenSSL/1.0.1 zlib/1.2.6 5 06-09-2013 06:20 Data directory: Q:\BOINCD 7 06-09-2013 06:20 Processor: 4 GenuineIntel Intel(R) Core(TM)2 Quad CPU @ 2.40GHz [Family 6 Model 15 Stepping 7] 8 06-09-2013 06:20 Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss htt tm pni ssse3 cx16 nx lm vmx tm2 pbe 9 06-09-2013 06:20 OS: Microsoft Windows 7: Home Premium x86 Edition, Service Pack 1, (06.01.7601.00) 10 06-09-2013 06:20 Memory: 2.00 GB physical, 8.00 GB virtual 11 06-09-2013 06:20 Disk: 19.52 GB total, 18.90 GB free 12 06-09-2013 06:20 Local time is UTC +2 hours 13 06-09-2013 06:20 CUDA: NVIDIA GPU 0: GeForce 8800 GTX (driver version 320.78, CUDA version 5.50, compute capability 1.0, 768MB, 677MB available, 518 GFLOPS peak) 14 06-09-2013 06:20 OpenCL: NVIDIA GPU 0: GeForce 8800 GTX (driver version 320.78, device version OpenCL 1.0 CUDA, 768MB, 677MB available, 518 GFLOPS peak) 15 06-09-2013 06:20 Config: GUI RPCs allowed from: 16 06-09-2013 06:20 192.168.2.100 17 06-09-2013 06:20 192.168.2.101 18 Einstein@Home 06-09-2013 06:20 URL http://einstein.phys.uwm.edu/; Computer ID 8670409; resource share 100 19 Albert@Home 06-09-2013 06:20 URL http://albert.phys.uwm.edu/; Computer ID 8733; resource share 100 20 WUProp@Home 06-09-2013 06:20 URL http://wuprop.boinc-af.org/; Computer ID 57501; resource share 100 21 World Community Grid 06-09-2013 06:20 URL http://www.worldcommunitygrid.org/; Computer ID 2445941; resource share 100 22 Constellation 06-09-2013 06:20 URL http://aerospaceresearch.net/constellation/; Computer ID 35560; resource share 100 23 LHC@home 1.0 06-09-2013 06:20 URL http://lhcathomeclassic.cern.ch/sixtrack/; Computer ID 10298249; resource share 100 24 World Community Grid 06-09-2013 06:20 General prefs: from World Community Grid (last modified 12-Jul-2012 21:43:37) 25 World Community Grid 06-09-2013 06:20 Computer location: work 26 06-09-2013 06:20 General prefs: using separate prefs for work 27 06-09-2013 06:20 Reading preferences override file 28 06-09-2013 06:20 Preferences: 29 06-09-2013 06:20 max memory usage when active: 1842.41MB 30 06-09-2013 06:20 max memory usage when idle: 1842.41MB 31 06-09-2013 06:20 max disk usage: 19.32GB 32 06-09-2013 06:20 (to change preferences, visit a project web site or select Preferences in the Manager) 33 06-09-2013 06:20 Not using a proxy ... 1035 Einstein@Home 06-09-2013 15:21 update requested by user 1036 Einstein@Home 06-09-2013 15:21 Sending scheduler request: Requested by user. 1037 Einstein@Home 06-09-2013 15:21 Not requesting tasks 1038 Einstein@Home 06-09-2013 15:21 Scheduler request completed 1039 Constellation 06-09-2013 15:21 Sending scheduler request: To fetch work. 1040 Constellation 06-09-2013 15:21 Requesting new tasks for CPU 1041 Constellation 06-09-2013 15:21 Scheduler request completed: got 0 new tasks 1042 Constellation 06-09-2013 15:21 No tasks sent 1043 Constellation 06-09-2013 15:21 No tasks are available for TrackJack Beta Test 1044 Constellation 06-09-2013 15:21 No tasks are available for Tycho 1045 Constellation 06-09-2013 15:21 No tasks are available for Tycho Betatest 1046 Constellation 06-09-2013 15:21 No tasks are available for Scilab Alpha Test 1047 Constellation 06-09-2013 15:21 No tasks are available for the applications you have selected. ... 1210 World Community Grid 06-09-2013 16:07 work fetch resumed by user 1211 World Community Grid 06-09-2013 16:07 update requested by user 1212 World Community Grid 06-09-2013 16:07 Sending scheduler request: Requested by user. 1213 World Community Grid 06-09-2013 16:07 Requesting new tasks for CPU 1214 World Community Grid 06-09-2013 16:07 Scheduler request completed: got 4 new tasks ... 1237 Einstein@Home 06-09-2013 16:07 Sending scheduler request: To fetch work. 1238 Einstein@Home 06-09-2013 16:07 Not requesting tasks 1239 Einstein@Home 06-09-2013 16:07 Scheduler request completed [Edit 1 times, last edit by Former Member at Sep 6, 2013 3:51:44 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
First you say
"I have set WCG, Albert@home and LHC@home to no new tasks for now." Then you bold WCG had a fetch of 4, so did you really do what you said? Why the client would not ask work of Einstein when it said it's going out to fetch work there, but then further responds with "Not requesting work", you'd have to ask at Einstein I suppose, after checking your profile there as to what work is allowed to be fetched for what resource. There's Work Fetch debug flag too. Set that and it will tell more. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Thanks for your reply SekeRob,
it's working again after increasing the minimum cache from 0.11 to 0.50 days. I turned WCG work fetch on again at line 1210. Before that WCG work fetch had been suspended for 12 hours. I thought Boinc was doing something wrong, no matter how i change Einstein's settings, Boinc doesn't even try to fetch CPU jobs when all other CPU projects are turned off. Yesterday, i had a message from Einstein not getting work because the project is not the highest priority. Using the work fetch flag i found this: 1589 Einstein@Home 06-09-2013 17:25 [work_fetch]: not requesting work for CPU: Constellation has higher priority I guess the prefs for Einstein hadn't been updated using "update" button, making Boinc think i only want GPU tasks. The prefs were finally updated after setting the cache high enough to request a second GPU wu. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Cache full had crossed my mind, but then there's other messages on top of 'not highest priority', such as 'No need' [There are ~14 specific messages why not]. Guessed the WTF flag would be the revealer, where of course GPU resource loading [Constellation is GPU?] has priority over CPU scheduling.
As for profile updates, don't think the project selections portions get transfered to the client, only those that you would [almost all] see in the local preferences. When a client connects, it mere tells how much it wants, for what resource, and the server scheduler then seeing what it's got for you. If you 'filter out' pieces of log from your post, then of course it becomes an even greater guess of reading between the lines, but roundabout you're on your way again. Next please [This one to be handled by Scribe, as I'm off] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
For 1000+ lines of log events i had to do some filtering. The first lines were just supposed to show client information and connected projects. The second segment is what i thought to be the problem. Part 3 shows that the client fetches work from WCG as it should (I put "work fetch resumed by user" in bold to make it clear what i did there). The last segment was repeated over and over again.
I found some similar topic (i think it was in Berkeley's forum) that Boinc 7.x is sometimes fooled when fetching work from more than 2 projects. However, i think that Einstein might also be the problem. On their Beta project, Albert@home, i got work units from applications i'd opted out from. To be more clear, this is what was running: - 1 NVIDIA GPU with ca. 5 hours time remaining - 1 WUProp (nci) (1 wu per client app that runs alongside on CPU without reserving its own thread; as these tasks are often time-overrated, could be it had 24 hours or more of est. remaining time). - zero CPU tasks (i.e. all 4 CPU threads dry) - no waiting to run/ready to start tasks All projects have the same prio (100). In any case, WCG is working as it should :) Enjoy your weekend, Sek. |
||
|
|
|