de & ps_separation_09_2s_sample_2_ GPU tasks validate error
log in

Advanced search

Message boards : Number crunching : de & ps_separation_09_2s_sample_2_ GPU tasks validate error

Author Message
Profile Ray_GTI-R
Avatar
Send message
Joined: 5 Nov 10
Posts: 56
Credit: 10,052,813
RAC: 45,222
Message 55188 - Posted: 20 Jul 2012 | 23:40:00 UTC

Started up a PC today after leaving it off for a while. Have now got:-
116 (new) failed GPU WUs.
70-odd other "Pending. Outcome Success. Validation inconclusive. Checked, but no consensus yet".
3 more OK.
Previously this PC worked fine.
I have not updated anything (as it was switched off).

See http://milkyway.cs.rpi.edu/milkyway/result.php?resultid=260073423 as an example.

What?

Link
Avatar
Send message
Joined: 19 Jul 10
Posts: 277
Credit: 10,204,111
RAC: 7,129
Message 55193 - Posted: 21 Jul 2012 | 8:26:42 UTC - in response to Message 55188.

The numbers you got are quite different from the wingmen, but I don't know why they are validate errors instead of "marked as invalid".

Well, that can be anything, the first two I'd check are heat and voltages.
____________
.

Profile Ray_GTI-R
Avatar
Send message
Joined: 5 Nov 10
Posts: 56
Credit: 10,052,813
RAC: 45,222
Message 55198 - Posted: 22 Jul 2012 | 0:08:43 UTC - in response to Message 55193.

Upped voltage from 0.974 (yep way, way undervolted, my choice).
99% sure that's the problem... I checked some results and they dropped into "Pending. Outcome Success. Validation inconclusive. Checked, but no consensus yet".

Then drastically dropped GPU from std 669 & Mem from std 829 - just in case 'tho heat wasn't ever a problem at that voltage.

Re-checked & so far OK? Conclusion - might have a crippled card ...

Simple solutions - which seem to work pending more, manual cross-checking - I like.

Thanks, "Link" for the support and the "keep it simple" approach to problem solving - saved me hours of what-if/maybe answers

:-P

BulletMagnetEd
Send message
Joined: 8 Jul 10
Posts: 13
Credit: 7,258,640
RAC: 833
Message 55273 - Posted: 1 Aug 2012 | 6:05:23 UTC

I too had a bunch of failed WUs - all from this job batch, and all dated sent on 30 July 2012 22:42:57 UTC and after. Nothing has changed on my PC, so any further ideas would be swell.

Link
Avatar
Send message
Joined: 19 Jul 10
Posts: 277
Credit: 10,204,111
RAC: 7,129
Message 55274 - Posted: 1 Aug 2012 | 7:07:32 UTC - in response to Message 55273.

Tried rebooting?
____________
.

BulletMagnetEd
Send message
Joined: 8 Jul 10
Posts: 13
Credit: 7,258,640
RAC: 833
Message 55278 - Posted: 2 Aug 2012 | 1:27:58 UTC

Oh yeah.

Post to thread

Message boards : Number crunching : de & ps_separation_09_2s_sample_2_ GPU tasks validate error


Main page · Your account · Message boards


Copyright © 2013 AstroInformatics Group