Forum > Windows

Computation Error with CUDA

<< < (2/4) > >>

Frizz:

--- Quote from: Geek@Play on 10 Aug 2010, 06:27:31 pm ---Were thise work units moved from gpu to CPU???

--- End quote ---

No. Nothing rescheduled.


What I seriously don't understand is why the error happens AT THE EXACT SAME SECOND for all WUs!

Raistmer:

--- Quote from: Geek@Play on 10 Aug 2010, 06:27:31 pm ---See this.......

http://boincfaq.mundayweb.com/index.php?language=1&view=480

Also more info available when Seti forums come back online.  See the forums there.

Were thise work units moved from gpu to CPU???

--- End quote ---
aborting by elapsed time limit exceeding usually marked in stderr by additional info lines. here no such info...

Richard Haselgrove:

--- Quote from: Frizz23 on 10 Aug 2010, 06:07:33 pm ---Ah! Thanks ... now how to read this?

<stderr_txt>
setiathome_CUDA: Found 1 CUDA device(s):
... 

</stderr_txt>

--- End quote ---

Also, have a look just a couple of lines above <stderr.txt>. If Carola's right, you should see
<message>
 - exit code -12 (0xfffffff4)
</message>and a few lines above that, <exit_status>-12</exit_status>If you see any other number, let us know: I must say, my first instinct on reading your initial report that they happened at the exact same second was that it might be a return of the -177s, but only you can tell until the servers are back up.

Miep:

--- Quote from: Geek@Play on 10 Aug 2010, 06:27:31 pm ---See this.......

http://boincfaq.mundayweb.com/index.php?language=1&view=480

Also more info available when Seti forums come back online.  See the forums there.

Were thise work units moved from gpu to CPU???

--- End quote ---

... uh right... there's such things as standard error messages ::)
'exceeded maximum disk space' ? or a memory leak? modern systems rarely run out of space, but a quick check can't hurt. The other one - try rebooting perhaps? something stuck on the GPU?

Miep:

--- Quote from: Richard Haselgrove on 10 Aug 2010, 06:36:05 pm ---Also, have a look just a couple of lines above <stderr.txt>. If Carola's right, you should see
<message>
 - exit code -12 (0xfffffff4)
</message>and a few lines above that, <exit_status>-12</exit_status>If you see any other number, let us know: I must say, my first instinct on reading your initial report that they happened at the exact same second was that it might be a return of the -177s, but only you can tell until the servers are back up.


--- End quote ---

oh right, exit codes get parsed to somewhere else...
I was just guessing, Richard, as I assumed -12 would show up as such. I've yet to come across one for myself, so I don't know what they look like.
ah and max CPU time exceeded should be -177 from the description in the faq.

Navigation

[0] Message Index

[#] Next page

[*] Previous page