Advanced search

Message boards : Number crunching : output file ... absent

Author Message
Profile Stephen Uitti
Send message
Joined: 17 Mar 14
Posts: 4
Credit: 77,427,636
RAC: 0
Level
Thr
Scientific publications
watwatwatwatwatwatwatwat
Message 37237 - Posted: 6 Jul 2014 | 15:17:59 UTC

I don't look at this much. But i saw that GPUGRID had run a couple units, yet i got no credit. Under "tasks", i have a number of units that say Error while computing. I looked at stdoutdae.txt, and the most recent GPUGRID entry is this:


06-Jul-2014 09:20:17 [GPUGRID] Sending scheduler request: To report completed tasks.
06-Jul-2014 09:20:17 [GPUGRID] Reporting 1 completed tasks
06-Jul-2014 09:20:17 [GPUGRID] Not requesting tasks: don't need
06-Jul-2014 09:20:29 [GPUGRID] Scheduler request completed
06-Jul-2014 09:21:46 [GPUGRID] Finished download of potx1x533-NOELIA_INSP-5-par_file
06-Jul-2014 09:21:46 [GPUGRID] Started download of potx1x533-NOELIA_INSP-5-conf_file_enc
06-Jul-2014 09:21:49 [GPUGRID] Finished download of potx1x533-NOELIA_INSP-5-conf_file_enc
06-Jul-2014 09:21:49 [GPUGRID] Started download of potx1x533-NOELIA_INSP-5-metainp_file
06-Jul-2014 09:21:50 [GPUGRID] Finished download of potx1x533-NOELIA_INSP-5-metainp_file
06-Jul-2014 09:21:50 [GPUGRID] Started download of potx1x533-NOELIA_INSP-5-potx1x533-NOELIA_INSP-4-13-RND4008_7
06-Jul-2014 09:21:52 [GPUGRID] Finished download of potx1x533-NOELIA_INSP-5-psf_file
06-Jul-2014 09:21:52 [GPUGRID] Finished download of potx1x533-NOELIA_INSP-5-potx1x533-NOELIA_INSP-4-13-RND4008_7
06-Jul-2014 09:21:52 [GPUGRID] Started download of potx1x533-NOELIA_INSP-5-potx1x533-NOELIA_INSP-4-13-RND4008_10
06-Jul-2014 09:21:53 [GPUGRID] Finished download of potx1x533-NOELIA_INSP-5-potx1x533-NOELIA_INSP-4-13-RND4008_10
06-Jul-2014 09:21:53 [GPUGRID] Starting task potx1x533-NOELIA_INSP-5-13-RND4008_0
06-Jul-2014 09:21:54 [GPUGRID] Computation for task potx1x533-NOELIA_INSP-5-13-RND4008_0 finished
06-Jul-2014 09:21:54 [GPUGRID] Output file potx1x533-NOELIA_INSP-5-13-RND4008_0_0 for task potx1x533-NOELIA_INSP-5-13-RND4008_0 absent
06-Jul-2014 09:21:54 [GPUGRID] Output file potx1x533-NOELIA_INSP-5-13-RND4008_0_1 for task potx1x533-NOELIA_INSP-5-13-RND4008_0 absent
06-Jul-2014 09:21:54 [GPUGRID] Output file potx1x533-NOELIA_INSP-5-13-RND4008_0_2 for task potx1x533-NOELIA_INSP-5-13-RND4008_0 absent

I'm on Linux, so i searched for 'absent' in stdoutdae.txt, and got 415 lines like that since 18 May 2014. I've gotten plenty of credit from GPUGRID over these months, so not everything could be broken. There seem to be four files per unit, so that's over 100 units. Ouch. I looked to see if i have any disk free issues, and i have 4.5 GB free on the partition where these files go. The BOINC manager agrees on disk space. It's the most recent BOINC manager - 7.2.42 (x64). Some recent applications include these:


15-Jun-2014 21:34:00 [GPUGRID] Output file I14R106-SDOERR_BARNA5-20-100-RND0625_1_0 for task I14R106-SDOERR_BARNA5-20-100-RND0625_1 absent

15-Jun-2014 21:37:57 [GPUGRID] Output file A2ARNUL_adapt4x04x40-GERARD_A2ARNUL_adapt4-16-17-RND6194_0_0 for task A2ARNUL_adapt4x04x40-GERARD_A2ARNUL_adapt4-16-17-RND6194_0 absent

18-Jun-2014 21:21:25 [GPUGRID] Output file e2s261_e1s77f308-SANTI_marsalWTbound2-5-32-RND2368_1_0 for task e2s261_e1s77f308-SANTI_marsalWTbound2-5-32-RND2368_1 absent

03-Jul-2014 21:30:19 [GPUGRID] Output file e2s4_e1s67f139-SANTI_marsalWTbound2-6-32-RND5456_1_0 for task e2s4_e1s67f139-SANTI_marsalWTbound2-6-32-RND5456_1 absent

06-Jul-2014 09:18:24 [GPUGRID] Output file I906-NATHAN_CMYBKIX_run3-13-250-RND7145_2_0 for task I906-NATHAN_CMYBKIX_run3-13-250-RND7145_2 absent



Computer ID 170278

CPU and run time are showing zero for lots of these failed units. So, maybe these units didn't actually run.

Profile Retvari Zoltan
Avatar
Send message
Joined: 20 Jan 09
Posts: 2343
Credit: 16,201,255,749
RAC: 851
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 37238 - Posted: 6 Jul 2014 | 18:08:46 UTC - in response to Message 37237.

It seems to me as a driver issue, as only the CUDA 6.0 tasks failing on your host, the CUDA 4.2 tasks working fine.
See the Important news for Linux Crunchers thread.

Post to thread

Message boards : Number crunching : output file ... absent

//