LHC@home

I have 20 outstanding Tunscana WU's in my pending credits plus numerous other shorter run WU's & I'm sure other people are in the same boat as me. I checked every one of the Tunascan WU's and every one of them has at least 3 or 4 people with pending credits on them, some of them have as many as 10 people that have turned in the results and are awaiting credit on them.

How many times are these WU's going to be sent out so somebody else can waste their time crunching them and sit there waiting for the credits to be granted.

It's not so bad with the shorter run WU's, at least you didn't spend a lot of time running them, but the Tunascana can take me anywheres from 9 hr's to 12 hr's to run them in HT Mode so you spend a lot of time running them and it would be nice to either get the credit for them or wipe them out so at least you don't have to wonder about them...IMO

I had a tunescan that finished on 25 October. It took until
yesterday to finally get the result removed from pending
credits.

The solution is, as far as I can tell, patience for things
to happen.

I think you missed my whole point dbernat, it has nothing to do with patience, my contention is why are 4 or 6 and even as many as 10 people turning in results on the same WU but nobody is getting credit for it & it just keeps getting sent back out for somebody else to do ...

Look at THIS WU,
where there is one person in progess now though there have been 3 good returns, a -zero credit- return, and 5 failures. Ive seen several with this profile, its alsmot as if they system isnt looking for a set number of good returns, but instead a porportion. Perhaps it grants credit for the WU when say 2/3 of ALL the returns (not just successes) approx agree on credit amt. I have seen several like this that have lead me to this conclusion. Has anyone else?


> my contention is why are 4 or 6 and even as many as 10 people turning in
> results on the same WU but nobody is getting credit for it & it just keeps
> getting sent back out for somebody else to do ...
>

This is how it works:

The validator requires three returns with identical values. It initially issues four Wus to have a reasonable chance that this will happen.

If it doesn't happen (values are different, WUs aren't returned in time, etc.,) it issues new WUs. if the quorum cannot be satisfied after ten WUs are returned (note - ten must be returned) then it gives up, issues credit for the work and moves on.

This means that for a problematic WU that doesn't generate consistent results at least ten WUs must be issued to get ten returns. If some aren't returned then the WU will be issued more than ten times.

This is all covered in the FAQ

Giskard - the first telepathic robot.

Footnote: The term identical value has nothing to with credits or matching values of credits. It refers to the end result of the calculations of the WU. This end result, whatever it may be, is not made available for the participants to view.