Welcome to MilkyWay@home

Completed, validation inconclusive, credit pending

Message boards : Number crunching : Completed, validation inconclusive, credit pending
Message board moderation

To post messages, you must log in.

AuthorMessage
Marek Majewski

Send message
Joined: 27 Mar 15
Posts: 13
Credit: 17,932
RAC: 0
Message 63300 - Posted: 30 Mar 2015, 21:21:18 UTC

Can somebody enlighten me what is the actual difference between a task labeled

1. VALIDATION PENDING (this I think I do understand), and

2. COMPLETED, VALIDATION INCONCLUSIVE, CREDIT PENDING (this I do know I don't understand).

Thanks,

-mm-
ID: 63300 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
swiftmallard
Avatar

Send message
Joined: 18 Jul 09
Posts: 300
Credit: 303,562,776
RAC: 0
Message 63301 - Posted: 30 Mar 2015, 21:32:01 UTC - in response to Message 63300.  

Work units are sent to more than one cruncher to process. After you return results that are somewhere near what the project expects them to be, you are now waiting for another cruncher to return the same results you did, plus or minus an acceptable margin of error. Validation inconclusive WUs are the units waiting for another cruncher to report results.
ID: 63301 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Marek Majewski

Send message
Joined: 27 Mar 15
Posts: 13
Credit: 17,932
RAC: 0
Message 63302 - Posted: 30 Mar 2015, 21:56:54 UTC - in response to Message 63301.  

Work units are sent to more than one cruncher to process. After you return results that are somewhere near what the project expects them to be, you are now waiting for another cruncher to return the same results you did, plus or minus an acceptable margin of error. Validation inconclusive WUs are the units waiting for another cruncher to report results.


I thought this is VALIDATION PENDING.

I have several units in either state on one of my computers, but I have never seen VALIDATION INCONCLUSIVE coming from the another one.

Can you elaborate, as I am sill confused...

Thanks for your time.

-mm-
ID: 63302 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Richard Haselgrove

Send message
Joined: 4 Sep 12
Posts: 219
Credit: 456,474
RAC: 0
Message 63303 - Posted: 30 Mar 2015, 22:52:50 UTC - in response to Message 63302.  

See my reference to Adaptive Replication
ID: 63303 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Marek Majewski

Send message
Joined: 27 Mar 15
Posts: 13
Credit: 17,932
RAC: 0
Message 63304 - Posted: 30 Mar 2015, 23:13:12 UTC

See my reference to Adaptive Replication


Richard,

thank you. I have read that, and unfortunately I am not the wiser.

Is there a user-friendly explanation on how VALIDATION INCONCLUSIVE varies from VALIDATION PENDING, or did I just read it in the link you provided and I am denser than Ununoctium? It is quite possible I should be added to the periodic table with the atomic number of 119. :-)

-mm-
ID: 63304 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 8 May 09
Posts: 3315
Credit: 519,939,926
RAC: 22,722
Message 63305 - Posted: 31 Mar 2015, 12:03:23 UTC - in response to Message 63304.  

See my reference to Adaptive Replication


Richard,

thank you. I have read that, and unfortunately I am not the wiser.

Is there a user-friendly explanation on how VALIDATION INCONCLUSIVE varies from VALIDATION PENDING, or did I just read it in the link you provided and I am denser than Ununoctium? It is quite possible I should be added to the periodic table with the atomic number of 119. :-)

-mm-


Validation pending is when someone returns a result and is waiting for the wingman, ie other pc, to return the same unit. Validation inconclusive means that one of the results is outside the expected range and the unit needs further crunching. Each unit is run thru a program that gives the project a quickie idea of what they should expect from it, it is not always right, but more often than not it is pretty close. Units could have a ton of stuff in them that the quickie look didn't see, so if one pc does see that and another one doesn't then one of those is wrong, that makes both results inconclusive until another pc checks the unit. Too many inconclusive results on any single unit and the project itself will run the unit to figure out exactly what is going on.

This is why some projects have multiple pc's crunching the same unit, the results aren't always predictable. Some projects units are less variable so a single pc crunching it is okay. And some projects start out with at least two pc's crunching a unit, but then as time goes by and a pc sends in the predicted result most of the time the project will consider that pc stable enough to stop using wingmen on that pc's workunits. They do periodic tests to spot check, but essentially once a pc becomes 'trusted' it is no longer waiting for a wingman to validate it's results.

Here at MW units are sent out like this:
minimum quorum 2
initial replication 3
max # of error/total/success tasks 3, 9, 6

That means each unit is sent out to three pc's, but only two are needed for a "quorum" saying it's okay. ALL three pc's get credit for the unit, unless it is returned past the deadline then it gets complicated.

The last line means each unit can be sent to a max of 9 pc's with a total of 6 of them getting the expected result and a max of 3 pc's having problems with the unit. After that the Project takes the unit out of circulation and crunches it itself. My understanding is the project also crunches some units as a 'spot check' just to make sure things are going as they expect, remember this is about Scientific Results, so accuracy is critical for the data to be used when writing papers etc.
ID: 63305 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Marek Majewski

Send message
Joined: 27 Mar 15
Posts: 13
Credit: 17,932
RAC: 0
Message 63307 - Posted: 31 Mar 2015, 21:38:55 UTC - in response to Message 63305.  

Thanks, now I understand.

-mm-

P.S. Your answer should be pinned as it fully explains the difference. Thanks again.
ID: 63307 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile mikey
Avatar

Send message
Joined: 8 May 09
Posts: 3315
Credit: 519,939,926
RAC: 22,722
Message 63308 - Posted: 1 Apr 2015, 11:12:56 UTC - in response to Message 63307.  

Thanks, now I understand.

-mm-

P.S. Your answer should be pinned as it fully explains the difference. Thanks again.


I've just been crunching awhile. once you get some more time in at other projects you will know this stuff too. While the Seti forums are informative they don't cover everything people see in Boinc. Personally I'm glad you are branching out, there's lots of good projects out here.
ID: 63308 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote

Message boards : Number crunching : Completed, validation inconclusive, credit pending

©2024 Astroinformatics Group