| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Mapping Cancer Markers Forum
Thread: CPU Time 0.00 |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 19
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I am not running any additional BOINC projects at this time.
/proc/[pid]/stat for one of the WUs gives me 15441 (wcgrid_mcm1_7.2) R 17850 17849 3343 34820 17849 4219136 1613217 0 0 0 3896105 2045 0 0 41 19 2 0 42263375 97947648 13212 18446744073709551615 4194304 6586118 140736130512336 140736130503144 4223384 0 41607712 0 1073755240 18446744073709551615 0 0 17 7 0 3 0 0 0 8683520 8692456 12480512 140736130518884 140736130519038 140736130519038 140736130523049 0 jagaer 15441 # ps ax | grep boinc 17849 pts/4 S+ 0:00 /bin/bash ./02-boincclient 17850 pts/4 Sl+ 9:47 /usr/bin/boinc_client 17986 pts/3 Sl 108:20 /usr/bin/boincmgr 18362 pts/3 S+ 0:00 /bin/bash ./04-saveboincram 19964 pts/2 S+ 0:00 grep --colour=auto boinc If I am reading this right: pid: 15441 comm: wcgrid_mcm1_7.2 state: R parent pid: 17850 process group: 17849 session: 3343 tty: 34820 terminal group: 17849 flags: 4219136 minor faults: 1613217 child minor faults: 0 major faults: 0 child major faults: 0 user time: 3896105 kernel time: 2045 child user time: 0 child kernel time: 0 priority: 41 nice: 19 ... So it appears that the kernel is identifying/tracking process time correctly, but BOINC is either not able to read from /proc or is processing it incorrectly. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi Jagaer,
 Just wanted to add my applause. You are being very helpful.  Lawrence |
||
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: 
|
The two projects that we run that use the proc filesystem to determine cpu time are running correctly on your machine and showing proper cpu time, these are cep2 and faah vina. The issue you are having with faah autodock and mcm1 use the linux system call getrusage to get the cpu time and this is having issues but I am not sure why.
Thanks, armstrdj |
||
|
|
Lavalamp
Cruncher Joined: Jan 4, 2013 Post Count: 2 Status: Offline |
There was a time keeping bug fix checked in for the BOINC client. Can't remember what platform it was directed at.
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
It appears getrusage is part of the glibc package.
----------------------------------------I am currently using glibc 2.17 sys-libs/glibc-2.17:2.2 USE="(multilib) -debug -gd (-hardened) -nscd -profile (-selinux) -suid -systemtap -vanilla" This was updated Jan 14 and may be causing the problem. I'll try re-installing glibc and associated libraries. [Edit 1 times, last edit by Former Member at Mar 6, 2014 1:01:55 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I believe the issue was actually my nvidia-drivers interfering with libtool
----------------------------------------Was using Nvidia-Drivers 331.17. Upgraded to 334.21. Additional detail here: http://forums.gentoo.org/viewtopic-t-979296-start-0.html The part that caused me to upgrade the driver There is a known bug in one of the currently used versions of the nVidia driver that mangles signal masks, which often manifests as weird time problems for build systems [Edit 1 times, last edit by Former Member at Mar 6, 2014 1:36:17 PM] |
||
|
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges:
|
Jagaer,
Did the upgraded drivers fix the issue with CPU time for the FAAH AutoDock and MCM1 projects? Thanks, armstrdj |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
armstrdj,
The upgraded nvidia-drivers did fix the CPU Time 0.00 issue with AutoDock and MCM projects. It did not fix the issue with CPU Time being *higher* than Elapsed time, however from other threads, this seems to be due to the 2xInt Core + 1xFP Core specific to AMD Bulldozer-series CPUs. Thanks for all your help, I wouldn't have been able to track down the issue without it. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
As always, your information is very helpful. Lawrence |
||
|
|
|