Why so long for validation

Every workunit is validated on at least two computers. If you return your workunit first, you end up waiting on the other computer, colloquially known as your "wingman". If their computer is considerably slower, such as my dual Pentium II 333MHz server, it will be a while before it returns it's copy of the same workunit. If you return your copy after your wingman and the results match, then credit is granted almost immediately (but it was your wingman that was waiting on you).


Further slowing down the validation process is the fact that if the two computer don't match on their results, a third copy will be sent out to yet another computer as a "tie breaker". An example is this workunit where my dual Pentium II server is the third computer to receive that workunit as the first two results didn't match.


A third copy may also be sent out if the wingman doesn't return their copy of the workunit before the workunit deadline. The third copy of the workunit will then be given a new deadline even further out so as to give the computer enough time to finish the workunit and return the result.

It isn't unheard of for a result to be sent out up to 6 or 7 times, but it is rather rare.

The longest I have had to wait was about 9 months.

As "" stated, it takes at least two matching results for valid results to be declared.

A couple of minor items missed:

In SETI, it is possible for three results to match weakly - in which case all three get credit.

If your result is past deadline when it is returned, and the replacement has not been returned, you will get credit if it is valid. If your result is past deadline and it is returned after the replacement, and validation has already occurred, you will not get credit. In any case, if your task is returned by deadline and it is valid you will get credit - even if the result is replacing has been returned in the iterim.
____________


BOINC WIKI