Message boards :
Number crunching :
UV reionization tasks wont finish
Message board moderation
Author | Message |
---|---|
Send message Joined: 22 Feb 15 Posts: 24 Credit: 250,365,396 RAC: 0 |
I have a few UV reionization tasks WUs running under Win64 that have been running 1.5 days, at 100% progress but won't finish. Let them run or kill them? This host...http://universeathome.pl/universe/results.php?hostid=2114 |
Send message Joined: 4 Feb 15 Posts: 847 Credit: 144,180,465 RAC: 0 |
Kill them... Krzysztof 'krzyszp' Piszczek Member of Radioactive@Home team My Patreon profile Universe@Home on YT |
Send message Joined: 22 Feb 15 Posts: 12 Credit: 16,678,134 RAC: 0 |
I've many compute errors with the UV tasks. |
Send message Joined: 28 Feb 15 Posts: 253 Credit: 200,562,581 RAC: 0 |
I've many compute errors with the UV tasks. You may be confusing errors with server aborts (code 202). I have seen a lot of the aborts on the recent batch. But that just means that the server has cancelled them because they have already been returned by someone else, not because they errored out on your PC. |
Send message Joined: 26 Mar 16 Posts: 2 Credit: 3,444,700 RAC: 0 |
... why does the server distribute the same/identical WU to different users in the first place ... ? Or am I missing out on something? I've been getting (since yesterday) "computation errors" and "aborts" and "cancellations" for the ultraviolet reionization WU en masse (means: many)! Something is going wrong on this project ?? |
Send message Joined: 4 Feb 15 Posts: 49 Credit: 15,956,546 RAC: 0 |
... why does the server distribute the same/identical WU to different users in the first place ... ? I think it has something to do with Krzyszp doing some manual intervention to get work flow going at a higher rate due to lots of complaints about not being able to get work. A lot of work that had been sent out and returned was found to have not had a second work unit sent out a number of days later so validation was not getting done. In the case of all the server cancellations (I also have had at least 22 today), it is due to 2 work units being sent out to get the original work unit to validate. As only a quorum of 2 is needed to validate a work unit the first back gets credit and the second work unit is then cancelled as not needed. It is unfortunate that BOINC calls these cancelled work units "errors" when they are not in fact errors at all. That is my take on it anyway. Conan |
Send message Joined: 4 Feb 15 Posts: 847 Credit: 144,180,465 RAC: 0 |
Exactly as you wrote. Also, due to high number of UV tasks in progress server are bit delayed with validations at the moment. Krzysztof 'krzyszp' Piszczek Member of Radioactive@Home team My Patreon profile Universe@Home on YT |