Advanced search

Message boards : Graphics cards (GPUs) : Just joined, all WUs error

Author Message
poppageek
Avatar
Send message
Joined: 4 Jul 09
Posts: 76
Credit: 114,610,402
RAC: 0
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 10970 - Posted: 5 Jul 2009 | 2:34:27 UTC

Hi
Win7 64 bit 7100. Boinc 6.6.36. EVGA GTX 260 192. Nvidia 185.85 and 186.18.

All WUs so far error out:

Name 81-KASHIF_HIVPR_twomons_far_ba7-8-100-RND2602_0
Workunit 581505
Created 1 Jul 2009 18:43:08 UTC
Sent 4 Jul 2009 7:08:17 UTC
Received 4 Jul 2009 8:49:54 UTC
Server state Over
Outcome Client error
Client state Compute error
Exit status 1 (0x1)
Computer ID 42462
Report deadline 9 Jul 2009 7:08:17 UTC
CPU time 37.28424
stderr out

<core_client_version>6.6.36</core_client_version>
<![CDATA[
<message>
Incorrect function. (0x1) - exit code 1 (0x1)
</message>
<stderr_txt>
# Using CUDA device 0
# Device 0: "GeForce GTX 260"
# Clock rate: 1405000 kilohertz
# Total amount of global memory: 939524096 bytes
# Number of multiprocessors: 24
# Number of cores: 192
# Amber: readparm : Reading parm file parameters
# PARM file in AMBER 7 format
# Encounter 10-12 H-bond term
WARNING: parameters.cu, line 568: Found zero 10-12 H-bond term.
WARNING: parameters.cu, line 568: Found zero 10-12 H-bond term.
MDIO ERROR: cannot open file "restart.coor"
Cuda error: Kernel [PmeRealSpace_compute_forces] failed in file 'PmeRealSpace.cu' in line 172 : unknown error.

</stderr_txt>
]]>

Validate state Invalid
Claimed credit 4569.99074074074
Granted credit 0
application version 6.64


This GTX 260 has done F@H for 2 months using 185.85 drivers with no errors. Any other drivers error rate is 20-30%. Have run Fur and CUDAmemtest several times with no errors. All errors seem to be same as above.

Thanks!

poppageek
Avatar
Send message
Joined: 4 Jul 09
Posts: 76
Credit: 114,610,402
RAC: 0
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 10971 - Posted: 5 Jul 2009 | 5:10:10 UTC

Update:

Have installed 182.50 drivers and took OC off CPU. GPU is and was at stock clocks. So far it is 1 hour into a WU which is almost twice as any of the others so maybe........

poppageek
Avatar
Send message
Joined: 4 Jul 09
Posts: 76
Credit: 114,610,402
RAC: 0
Level
Cys
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 10976 - Posted: 6 Jul 2009 | 4:47:30 UTC

Another update.

Installed Vista 64, Nvidia driver 186.18 and Boinc 6.6.36. First WU errored. If second one does will try 185.50.

Running out of ideas here......

MarkJ
Volunteer moderator
Volunteer tester
Send message
Joined: 24 Dec 08
Posts: 738
Credit: 200,909,904
RAC: 0
Level
Leu
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 10979 - Posted: 6 Jul 2009 | 7:51:47 UTC - in response to Message 10976.

Another update.

Installed Vista 64, Nvidia driver 186.18 and Boinc 6.6.36. First WU errored. If second one does will try 185.50.

Running out of ideas here......


I'd suggest 182.50 drivers. An oldie but a goodie. We've had various issues with 185.xx and 186.xx drivers on GPUgrid. Remember you have to uninstall the old drivers before reinstalling. Also finish off any cuda work (regardless of project) and make sure you report it before doing the driver reinstall.
____________
BOINC blog

Post to thread

Message boards : Graphics cards (GPUs) : Just joined, all WUs error

//