Advanced search

Message boards : Number crunching : Bad batch of tasks?

Author Message
Richard Haselgrove
Send message
Joined: 11 Jul 09
Posts: 1620
Credit: 8,866,435,925
RAC: 20,013,489
Level
Tyr
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 48010 - Posted: 19 Oct 2017 | 12:39:42 UTC

One of my machines has had two recently from

ADRIA_FOLDUBQ80_crystal_ss_contacts_50_ubiquitin

which have failed on all machines they've been sent to. Error tasks for host 43404

Erich56
Send message
Joined: 1 Jan 15
Posts: 1132
Credit: 10,256,332,676
RAC: 29,209,412
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwat
Message 48012 - Posted: 19 Oct 2017 | 16:38:03 UTC
Last modified: 19 Oct 2017 | 16:38:26 UTC

same here with
e1s83_ubiquitin_50ns_16-ADRIA_FOLDUBQ80_crystal_ss_contacts_50_ubiquitin_0-0-1-RND4530_7
which errored out after 1.09 seconds :-(

http://gpugrid.net/result.php?resultid=16601415

these tasks should be removed from the batch!

Erich56
Send message
Joined: 1 Jan 15
Posts: 1132
Credit: 10,256,332,676
RAC: 29,209,412
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwat
Message 48014 - Posted: 20 Oct 2017 | 9:39:13 UTC

On the Project Status Page, at the moment this type of task shows an error rate of 68.69% !

Erich56
Send message
Joined: 1 Jan 15
Posts: 1132
Credit: 10,256,332,676
RAC: 29,209,412
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwat
Message 48015 - Posted: 20 Oct 2017 | 19:25:52 UTC

just a minute ago, another task from this batch errored out after 35 seconds.
The error rate on the Project Status page has meanwhile exceeded 70% !

I don't understand why this batch of faulty tasks has not been removed yet :-(

Erich56
Send message
Joined: 1 Jan 15
Posts: 1132
Credit: 10,256,332,676
RAC: 29,209,412
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwat
Message 48016 - Posted: 21 Oct 2017 | 6:14:51 UTC

Interestingly enough, on the Project Status Page this batch is now listed under "short runs" - why so?

On the other hand, that's where these tasks belong to, anyway, after they run only for very short time, i.e. a few seconds :-)

Betting Slip
Send message
Joined: 5 Jan 09
Posts: 670
Credit: 2,498,095,550
RAC: 0
Level
Phe
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 48017 - Posted: 21 Oct 2017 | 10:00:39 UTC - in response to Message 48016.

Have I caught you talking to yourself again Erich?

Erich56
Send message
Joined: 1 Jan 15
Posts: 1132
Credit: 10,256,332,676
RAC: 29,209,412
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwat
Message 48022 - Posted: 22 Oct 2017 | 7:03:45 UTC - in response to Message 48017.

Have I caught you talking to yourself again Erich?

in absence of any reaction/answer/comment from the GPUGRID people this may look like talking to myself, yes :-(

Bedrich Hajek
Send message
Joined: 28 Mar 09
Posts: 485
Credit: 11,109,083,435
RAC: 15,460,691
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwatwat
Message 48023 - Posted: 22 Oct 2017 | 10:09:53 UTC - in response to Message 48015.

just a minute ago, another task from this batch errored out after 35 seconds.
The error rate on the Project Status page has meanwhile exceeded 70% !

I don't understand why this batch of faulty tasks has not been removed yet :-(



It is actually better to let these units run the course, and become a "too many errors (may have a bug)" units, which means they will eventually disappear from your computer's task page. If they are canceled before that, they will stay with you forever. It is a bug in the system! This is mentioned in another thread, somewhere.

Here's another question. Where are the new IT people or person, that were or was supposed to be hired earlier this year to fix this and other problems?


Erich56
Send message
Joined: 1 Jan 15
Posts: 1132
Credit: 10,256,332,676
RAC: 29,209,412
Level
Trp
Scientific publications
watwatwatwatwatwatwatwatwat
Message 48025 - Posted: 22 Oct 2017 | 11:58:43 UTC - in response to Message 48023.

Here's another question. Where are the new IT people or person, that were or was supposed to be hired earlier this year to fix this and other problems?

I suspect there is no such person yet. Unless someone from GPUGRID can tell us the contrary.

Post to thread

Message boards : Number crunching : Bad batch of tasks?

//