I noticed recently that cpu time drops to near-zero periodically on all four cores while Boinc tasks are running in the background. I may be losing as much as one second in every five.

I am running Boinc 6.2.18 on Ubuntu Linux 9.04, kernel 2.6.29-14-generic, 32 bit.
Four CPUs, Intel(R) Core(TM)2 Quad CPU Q8200 @ 2.33GHz. 1.5GB RAM, 1.9GB swap.

The only way I know it pauses is from the system monitor "cpu history" graph. I originally had it sampling every 3 seconds, and saw an odd pattern. I changed it to sample every .25 seconds, and now see a clear 1-second "U" shape with all four processors dropping well below 40% every time. I did verify that a compute-bound program kept one processor graph up. I'm not fully convinced it's a Boinc bug, but it's clearly Boinc-related. The "waiting channel" for the Boinc tasks sometimes flashes "hrtimer_nanosleep", presumably in sync with the drop in cpu usage.

Mostly I want to know if anyone else has seen anything remotely like this. I wondered if it was configuration-dependent. Conceivably it might pause to check if the cpu is in use elsewhere by pausing. Either Boinc or my OS might have a "duty cycle" feature, possibly because of or to prevent overheating. I will submit config files if they appear to be relevant.


Here are the messages when I restarted the client:

Sun 17 Oct 2010 09:59:11 PM EDT||Starting BOINC client version 6.2.18 for i686-pc-linux-gnu
Sun 17 Oct 2010 09:59:11 PM EDT||log flags: task, file_xfer, sched_ops
Sun 17 Oct 2010 09:59:11 PM EDT||Libraries: libcurl/7.18.2 OpenSSL/0.9.8g zlib/1.2.3.3 libidn/1.10
Sun 17 Oct 2010 09:59:11 PM EDT||Data directory: /var/lib/boinc-client
Sun 17 Oct 2010 09:59:13 PM EDT||Processor: 4 GenuineIntel Intel(R) Core(TM)2 Quad  CPU   Q8200  @ 2.33GHz [Family 6 Model 23 Stepping 7]
Sun 17 Oct 2010 09:59:13 PM EDT||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 ht tm pbe lm constant_tsc arch_perfmon pebs bts pni dtes64 monitor ds_cpl est tm2 ssse3 cx16 xtpr pdcm sse4_1 lahf_lm
Sun 17 Oct 2010 09:59:13 PM EDT||OS: Linux: 2.6.28-14-generic
Sun 17 Oct 2010 09:59:13 PM EDT||Memory: 1.46 GB physical, 1.86 GB virtual
Sun 17 Oct 2010 09:59:13 PM EDT||Disk: 8.25 GB total, 1.31 GB free
Sun 17 Oct 2010 09:59:13 PM EDT||Local time is UTC -4 hours
Sun 17 Oct 2010 09:59:13 PM EDT||No coprocessors
Sun 17 Oct 2010 09:59:13 PM EDT|World Community Grid|URL: http://www.worldcommunitygrid.org/; Computer ID: 922063; location: (none); project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT|boincsimap|URL: http://boinc.bio.wzw.tum.de/boincsimap/; Computer ID: 144534; location: (none); project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT|Poem@Home|URL: http://boinc.fzk.de/poem/; Computer ID: 47818; location: (none); project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT|FreeHAL@home|URL: http://www.freehal.net/freehal_at_home/; Computer ID: 15772; location: home; project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT|rosetta@home|URL: http://boinc.bakerlab.org/rosetta/; Computer ID: 1056045; location: (none); project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT|QMC@HOME|URL: http://qah.uni-muenster.de/; Computer ID: 144725; location: home; project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT|orbit@home|URL: http://orbit.psi.edu/oah/; Computer ID: 19227; location: (none); project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT|lhcathome|URL: http://lhcathome.cern.ch/lhcathome/; Computer ID: 9849984; location: home; project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT|yoyo@home|URL: http://www.rechenkraft.net/yoyo/; Computer ID: 44880; location: (none); project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT|Einstein@Home|URL: http://einstein.phys.uwm.edu/; Computer ID: 1919156; location: home; project prefs: default
Sun 17 Oct 2010 09:59:13 PM EDT||General prefs: from http://milkyway.cs.rpi.edu/milkyway/ (last modified 06-Sep-2009 18:07:44)
Sun 17 Oct 2010 09:59:13 PM EDT||Host location: none
Sun 17 Oct 2010 09:59:13 PM EDT||General prefs: using your defaults
Sun 17 Oct 2010 09:59:13 PM EDT||Reading preferences override file
Sun 17 Oct 2010 09:59:13 PM EDT||Preferences limit memory usage when active to 599.53MB
Sun 17 Oct 2010 09:59:13 PM EDT||Preferences limit memory usage when idle to 1049.18MB
Sun 17 Oct 2010 09:59:13 PM EDT||Preferences limit disk usage to 0.67GB
Sun 17 Oct 2010 09:59:13 PM EDT|QMC@HOME|Restarting task qasino_b3lypqz-E10_iso34.7204_0 using qasinoAlpha version 501
Sun 17 Oct 2010 09:59:13 PM EDT|World Community Grid|Restarting task c4cw_target02_032156235_1 using c4cw version 613
Sun 17 Oct 2010 09:59:13 PM EDT|World Community Grid|Restarting task faah16358_ZINC04723609_WT2md01450CTP_01_1 using faah version 607
Sun 17 Oct 2010 09:59:13 PM EDT|QMC@HOME|Restarting task qasino_b3lypqz-E11_iso34.3886_0 using qasinoAlpha version 501
Sun 17 Oct 2010 10:00:20 PM EDT||General prefs: from http://milkyway.cs.rpi.edu/milkyway/ (last modified 06-Sep-2009 18:07:44)
Sun 17 Oct 2010 10:00:20 PM EDT||Host location: none
Sun 17 Oct 2010 10:00:20 PM EDT||General prefs: using your defaults
Sun 17 Oct 2010 10:00:20 PM EDT||Reading preferences override file
Sun 17 Oct 2010 10:00:20 PM EDT||Preferences limit memory usage when active to 599.53MB
Sun 17 Oct 2010 10:00:20 PM EDT||Preferences limit memory usage when idle to 1049.18MB
Sun 17 Oct 2010 10:00:20 PM EDT||Preferences limit disk usage to 0.67GB
Sun 17 Oct 2010 10:00:58 PM EDT||General prefs: from http://milkyway.cs.rpi.edu/milkyway/ (last modified 06-Sep-2009 18:07:44)
Sun 17 Oct 2010 10:00:58 PM EDT||Host location: none
Sun 17 Oct 2010 10:00:58 PM EDT||General prefs: using your defaults
Sun 17 Oct 2010 10:00:58 PM EDT||Reading preferences override file
Sun 17 Oct 2010 10:00:58 PM EDT||Preferences limit memory usage when active to 599.53MB
Sun 17 Oct 2010 10:00:58 PM EDT||Preferences limit memory usage when idle to 1049.18MB
Sun 17 Oct 2010 10:00:58 PM EDT||Preferences limit disk usage to 0.67GB

Obviously, you have set your computing preferences to:

Use at most 80% of CPU time
Can be used to reduce CPU heat

That's the BOINC way to reduce CPU usage. ;-)

Gruß,
Gundolf

---

Computer sind nicht alles im Leben. (Kleiner Scherz)

OK, problem solved. I get my prefs from milkyway, which I was disconnected from. I reconnected, went in and looked at the prefs, and it showed 100. Since there is nowhere to set it in the client, I didn't expect it to be overridden.

I looked at the .xml's. The prefs had 100, but the override had 80. Maybe I put it in there manually or a different boinc manager wrote it. Thanks a lot for pointing me to the right place.

Since there is nowhere to set it in the client, I didn't expect it to be overridden.

BOINC Manager->Advanced view->Advanced->Preferences.
Clear will clear them completely.
OK will save them unconditionally and BOINC will immediately start using them.

And it does say so in your messages:

Sun 17 Oct 2010 10:00:20 PM EDT||Reading preferences override file

---