Report problems with Rosetta version 5.36

Message boards : Number crunching : Report problems with Rosetta version 5.36

To post messages, you must log in.

Previous · 1 · 2 · 3 · 4 · 5

AuthorMessage
JohanDM

Send message
Joined: 22 Nov 05
Posts: 1
Credit: 219,288
RAC: 0
Message 31125 - Posted: 14 Nov 2006, 9:51:55 UTC
Last modified: 14 Nov 2006, 9:59:41 UTC

I've two WU's that reported OK but didn't get credit because of validate error:
46913918
46814172

ID: 31125 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
RC

Send message
Joined: 27 Sep 05
Posts: 13
Credit: 262,048
RAC: 0
Message 31130 - Posted: 14 Nov 2006, 13:07:22 UTC

I also have two WUs that failed with a validate error:
46966003
46965919

Both of these ran normally with no errors in the log. This is on a Linux box (AMD Athlon64 3000, no screensaver, non-optimized x86 BOINC v5.4.9).
ID: 31130 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Christoph Jansen
Avatar

Send message
Joined: 6 Jun 06
Posts: 248
Credit: 267,153
RAC: 0
Message 31134 - Posted: 14 Nov 2006, 16:34:19 UTC
Last modified: 14 Nov 2006, 16:39:30 UTC

And the same here, all "DOC" WUs seem to just not want to validate. And I noticed that the project's RAC has dropped from 38.5 to just 33.5 TeraFLOPs today. Probably it is the effect of all those WUs erroring out in validation?

[EDIT] Yes, I've browsed through the results on some of the top machines. All those WUs that errored out with "validate error" I saw since Sunday were "DOC" WUs. So it looks like somebody ought to look into that, as it is also lost capacity to the project if they are useless -or lost results if the validator just is wrong on those special WUs and results are worth looking at.

Could we please get an instruction of what to do with those WUs?[/EDIT]
ID: 31134 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Chu

Send message
Joined: 23 Feb 06
Posts: 120
Credit: 112,439
RAC: 0
Message 31156 - Posted: 15 Nov 2006, 0:42:51 UTC - in response to Message 31134.  
Last modified: 15 Nov 2006, 0:54:11 UTC

Please abort those WUs with names like "DOC_????_R061030_st_mode_??", but NOT "DOC_R061113_***_fa_relax_from_native_bound". The problem is related to something I had post in here

These WUs were added into the queue before 5.40 was updated and they work fine with 5.36. But we found later the new application has some backward compatibility issue with these WUs and since most of these WUs are still in the queue, we chose to cancel the whole batch. However, this generates the new problem for validating results from those WUs which are in the same batch and have been sent out to run with 5.36 already. We are still investigating why this happens and for the moment, please abort these WUs and we will try to come up with a plan to adjust credits for those validator errors later. Sorry for causing this mess
And the same here, all "DOC" WUs seem to just not want to validate. And I noticed that the project's RAC has dropped from 38.5 to just 33.5 TeraFLOPs today. Probably it is the effect of all those WUs erroring out in validation?

[EDIT] Yes, I've browsed through the results on some of the top machines. All those WUs that errored out with "validate error" I saw since Sunday were "DOC" WUs. So it looks like somebody ought to look into that, as it is also lost capacity to the project if they are useless -or lost results if the validator just is wrong on those special WUs and results are worth looking at.

Could we please get an instruction of what to do with those WUs?[/EDIT]


ID: 31156 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Chu

Send message
Joined: 23 Feb 06
Posts: 120
Credit: 112,439
RAC: 0
Message 31159 - Posted: 15 Nov 2006, 0:58:01 UTC

Please continue to report problems in 5.40 bug-report thread. Thanks.
ID: 31159 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Previous · 1 · 2 · 3 · 4 · 5

Message boards : Number crunching : Report problems with Rosetta version 5.36



©2022 University of Washington
https://www.bakerlab.org