Message boards :
Number crunching :
lots of " Completed, validation inconclusive "
Message board moderation
Author | Message |
---|---|
Send message Joined: 2 Oct 10 Posts: 74 Credit: 18,362,557 RAC: 0 |
Here is a snapshot for the problem http://www.m5zn.com/uploads/2011/5/26/photo/052611160504o1cc4atp8cuwzcl.png Is it ok? any solution? Whe is it happening ? |
Send message Joined: 24 Feb 09 Posts: 620 Credit: 100,587,625 RAC: 0 |
All is fine. If you click on "Account - Computers - Task - Pendings", then look at the "Work Unit" column, click on any blue number where it shows "inconclusive", and you will see a second copy sent out to another cruncher. You will have to wait for that to be returned by that Cruncher before validation is complete and you get the credits. The term is "wait for the Wingman" Validation is done in a number of ways. One of them is a check against the results you have produced by comparing them with someone else, if they differ, a third one is sent out as "arbitor". A number of other things happen (mainly off line that we dont see), but some rough checks are done before credit is granted, and the WU is moved out to the Project's next process (another form of validation). Selection of that method of checking is done at random (as such) depending on previous results. In any event, even if perfect on previous ones, there is still a random check carried out. All normal, no worries lurking. Once the Wingman returns their copy - and assuming it checks against yours - credit is granted. It can mean a wait of a couple of weeks sometimes, depends how fast the wingman gets on with it - or sometimes waiting for their's to timeout and get sent out again. Its a standard BOINC process that many Projects use when the nature of the data crunched means it need's validation. EDIT : Just noticed, you have done quite a lot at Collatz - the validation process is similar, except at Collatz every WU is validated, here its a random selection, not all of them. Regards Zy |
Send message Joined: 2 Oct 10 Posts: 74 Credit: 18,362,557 RAC: 0 |
Thanks Zy for your help |
Send message Joined: 22 Mar 09 Posts: 99 Credit: 503,422,495 RAC: 0 |
I also have a lot of "Completed, validation inconclusive"-GPU-WU. My findings are that the problem for me is the 6.12.26-Client. After roll-back to 6.12.22 nearly all tasks validate immediatly. Tests of newer clients then 6.12.26 are pending. |
Send message Joined: 22 Mar 09 Posts: 99 Credit: 503,422,495 RAC: 0 |
I have tested 6.12.28. It works fine. |
Send message Joined: 15 Jul 08 Posts: 383 Credit: 729,293,740 RAC: 0 |
I also have a lot of "Completed, validation inconclusive"-GPU-WU. My findings are that the problem for me is the 6.12.26-Client. After roll-back to 6.12.22 nearly all tasks validate immediatly. That is not related to the BOINC client. Using 6.12.26 and 6.12.28 here with no problems on any of a wide variety of machines. |
Send message Joined: 22 Mar 09 Posts: 99 Credit: 503,422,495 RAC: 0 |
On my system it seems to be related to the client: My findings with 2x5850: (Catalyst 11.3) 6.12.22: about 210000 credits/17 h 6.12.26: about 130000 credits/17 h 6.12.28: about 210000 credits/17 h Nothing changed but the client, so I think, it is a scientific evidence that the client makes problems on my system. But I have no idea why. :-) |
Send message Joined: 15 Jul 08 Posts: 383 Credit: 729,293,740 RAC: 0 |
On my system it seems to be related to the client: What that's really telling you is either that the validator was a bit behind during that period or your wingmen were slower on returning WUs. Validation inconclusive has nothing to do with the BOINC client. The only reason I'm correcting this idea is so others won't be misled. Regards/Beyond. |
Send message Joined: 15 Jul 11 Posts: 1 Credit: 3,228,580 RAC: 0 |
Thank you for the information. I also was getting worried that something was not copacetic and I saw this post. |
Send message Joined: 25 Dec 22 Posts: 1 Credit: 36,641 RAC: 0 |
thanks a lot I am allright now then.... |
©2024 Astroinformatics Group