| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Active Research
Forum: Africa Rainfall Project
Thread: Work Available |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 3596
|
|
| Author |
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7851 Status: Offline Project Badges: 
|
Hopefully I'll get some resends to keep me busy until then. Eventually got a resend, a _5 which may be a personal record going for the 6th attempt at a work unit.I seem to remember that the limit for retries is 6, of which you have the sixth one. After that the unit is either examined by the technicians or sent back to the researchers, because their is either something malformed, missing or otherwise out of whack. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
mikey
Veteran Cruncher Joined: May 10, 2009 Post Count: 826 Status: Offline Project Badges:
|
Hopefully I'll get some resends to keep me busy until then. Eventually got a resend, a _5 which may be a personal record going for the 6th attempt at a work unit.I seem to remember that the limit for retries is 6, of which you have the sixth one. After that the unit is either examined by the technicians or sent back to the researchers, because their is either something malformed, missing or otherwise out of whack. Cheers I wonder if they are just user aborted tasks or time outs if they will just put them back in the queue and see if they get done the 2nd time thru?   |
||
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12594 Status: Offline Project Badges: 
|
I believe that the limit is on the number errored- possibly 4. Since the restart there have been so many timed out that we have had some bg numbers of re-sends because only errors countfor the limit.
Mike |
||
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12594 Status: Offline Project Badges:
|
I don't know about rainfall. It is a parched desert here.
Mike |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7851 Status: Offline Project Badges:
|
Hopefully I'll get some resends to keep me busy until then. Eventually got a resend, a _5 which may be a personal record going for the 6th attempt at a work unit.I seem to remember that the limit for retries is 6, of which you have the sixth one. After that the unit is either examined by the technicians or sent back to the researchers, because their is either something malformed, missing or otherwise out of whack. Cheers I wonder if they are just user aborted tasks or time outs if they will just put them back in the queue and see if they get done the 2nd time thru? That is a good question about whether or not they counted the aborted units. Now I wonder about that too. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 1332 Status: Offline Project Badges:
|
On retries, or "Why are we seeing work units with seven or eight attempts?"
----------------------------------------Over in a thread in "Website Support" a couple of examples of such work units have been highlighted (including an example of an 8-attempt workunit!), and I've also been involved in a work unit that needed _6 to get it validated! As I understand it, an unmodified BOINC transitioner treats the "maximum number of attempts" as a hard limit on a per-application basis (with no adjustment for any type of "non-success" result); this implies that for ARP1 the limit has been lifted above the six it used to seem to be. Now, I wonder if IBM raised the limits during the pre-transfer run-down because there were more No Reply and "Not started by deadline" tasks because of the compressed deadlines -- if they wanted to minimize the amount of units getting failed because of non-computational issues it would make sense... Or perhaps Krembil have upped the limits?? Just a possibility... Cheers - Al. [Edit - added the "no adjustment" remark] [Edit 1 times, last edit by alanb1951 at Oct 11, 2022 9:00:40 PM] |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7851 Status: Offline Project Badges:
|
As I understand it, an unmodified BOINC transitioner treats the "maximum number of attempts" as a hard limit on a per-application basis (with no adjustment for any type of "non-success" result); this implies that for ARP1 the limit has been lifted above the six it used to seem to be. Now, I wonder if IBM raised the limits during the pre-transfer run-down because there were more No Reply and "Not started by deadline" tasks because of the compressed deadlines -- if they wanted to minimize the amount of units getting failed because of non-computational issues it would make sense... Or perhaps Krembil have upped the limits?? Interesting information. The last time I remember it even coming up for discussion was the Clean Energy Project a long time ago because of the heavy requirements for the work units. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
PMH_UK
Veteran Cruncher UK Joined: Apr 26, 2007 Post Count: 786 Status: Offline Project Badges: 
|
Below waited days to re-send.
----------------------------------------https://www.worldcommunitygrid.org/contribution/workunit/177231490 Very slow validation. Paul.
Paul.
|
||
|
|
Mike.Gibson
Ace Cruncher England Joined: Aug 23, 2007 Post Count: 12594 Status: Offline Project Badges:
|
Paul
That unit has more than 1 problem. -0 and -1 were sent out with 6 days to complete. -1 errored almost immediately so -2 was sent almost immediately. -2 also errored but after 1 day, so -3 was sent, again almost immediately. However, -3 was sent with only 3 days to complete, although it was still less than halfway through the initial period. There was then no reply from either -0 or -3 when their deadlines passed but no replacements sent for either of them for days. -4 & -5 were sent out together, days late, and both completed and validated. The 2 problems with the handling of this unit were that -3 was sent out with its deadline halved when it was less than halfway through the original deadline and that there was a holdup in the dispatch of replacements when the deadlines passed. Mike |
||
|
|
Unixchick
Veteran Cruncher Joined: Apr 16, 2020 Post Count: 1312 Status: Offline Project Badges: 
|
I'm still only getting the odd resend here and there from ARP. I'm surprised when I look at the scripts to see that there are still between 1000-2000 plots moving to the next iteration. Is this all resends? Do some people have a huge cache? Are new ARPs going out but just such a small amount, that I miss them?
If they are going to send out so few ARPs I really wish they would focus on the ones far behind. We could really use this slow time to make some good progress. Send each extreme to 5 people. When two come back that agree move to next gen. repeat. |
||
|
|
|