Claimed vs Granted Credits

fadedrose
fadedrose
Joined: 6 Apr 13
Posts: 263
Credit: 316405
RAC: 0
Topic 196913

In my tasks, the column that shows "Valid" tasks, those that are completed and granted credit, is confusing.

To give an example, in one column headed "Claimed Credit," it shows 89.52, but in the "Granted Credit" column, it shows 251.19. All of the granted credits are unseemingly high compared to the claimed credit for the task. Also, who claimed the lower numbers? I know I didn't..and who gave me the higher credits?

Question is: Why are my granted credits so much higher than the claimed? Please make the reply as simple as possible so I could understand it.

Thanks

9.53 251.19

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284700169
RAC: 113726

Claimed vs Granted Credits

Quote:

In my tasks, the column that shows "Valid" tasks, those that are completed and granted credit, is confusing.

To give an example, in one column headed "Claimed Credit," it shows 89.52, but in the "Granted Credit" column, it shows 251.19. All of the granted credits are unseemingly high compared to the claimed credit for the task. Also, who claimed the lower numbers? I know I didn't..and who gave me the higher credits?

Question is: Why are my granted credits so much higher than the claimed? Please make the reply as simple as possible so I could understand it.

Thanks


Claimed credit is a BOINC feature but we don't use it here at E@H where all credit is assigned by the project.

This comes from the days when some users would build their own BOINC client - the program on the host side that does the transactions with all the projects that you have enrolled for. It's open source you see. This had the unfortunate effect of 'optimising' ( ahem .... ) credit calculations ie. allow the client ( read : user clever enough to rewrite the client code ) to decide upon the merit of the work done on their rigs. To avoid 'credit inflation', at least for E@H, the simplest answer was to award credit server side.

I'm not even sure if anyone is bothering with such 'optimised' clients anywhere anymore -> Ageless ??

Cheers, Mike.

( edit ) I personally am only interested in credit as a technical measure of performance. But I do respect others who wish to assign other significance. After all if you crunch for E@H then the merit of that for the project is a constant with respect to motivations, the science value is the same. In any event having some users 'choose their own price' over others is a damaging trend regardless of which impetus you highlight.

I'd really like someone suitably skilled in sociological techniques to examine this credit business, as I feel it is a key/pivot point for many people's involvement. Whatever they find I hope you could bottle it! Anyone want to co-author a study with me ??? ;-) :=)

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Jord
Joined: 26 Jan 05
Posts: 2952
Credit: 5779100
RAC: 0

RE: This comes from the

Quote:

This comes from the days when some users would build their own BOINC client - the program on the host side that does the transactions with all the projects that you have enrolled for. It's open source you see. This had the unfortunate effect of 'optimising' ( ahem .... ) credit calculations ie. allow the client ( read : user clever enough to rewrite the client code ) to decide upon the merit of the work done on their rigs. To avoid 'credit inflation', at least for E@H, the simplest answer was to award credit server side.

I'm not even sure if anyone is bothering with such 'optimised' clients anywhere anymore -> Ageless ??


Well no, the claimed credit comes from the days that credit was calculated from the run time of the task times the integer benchmark value. have two computers with similar enough credit claims and both would get that value. By artificially increasing your integer value, you would be able to claim astronomical credit values. Which on a project with a quorum of 1, you would get.

Einstein has never had a quorum of 1, so here your claim would always have been checked against the claim of another computer, and if your claim was exuberant, it would go out to a third computer, which would just conclude that the high claim was BS and the two matching would get credit, the one not matching wouldn't.

Of course there was the chance that there were two computers claiming the astronomical value and that they would get it. Which is why the Einstein project went and set up static credit.

The newer forum software no longer shows the claimed credit column. Instead it shows:
[pre]Task Work unit Computer Sent Time reported or deadline Status Run time (sec) CPU time (sec) Credit Application
[/pre]
So, just go with the flow and update the software. (or we'll wait for the Drupal pages...) ;-)

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284700169
RAC: 113726

RE: Einstein has never had

Quote:

Einstein has never had a quorum of 1, so here your claim would always have been checked against the claim of another computer, and if your claim was exuberant, it would go out to a third computer, which would just conclude that the high claim was BS and the two matching would get credit, the one not matching wouldn't.

Of course there was the chance that there were two computers claiming the astronomical value and that they would get it. Which is why the Einstein project went and set up static credit.


Thanks Jord. Now I remember. Yes, optimised counts were not a problem provided they remained rare as the quorum would sort that. Of course that begged to increase with time and take-up of said clients.

IIRC - I think the related issue of the number of ( valid ) results per quorum was settled at LSC level around that time too. This is important as regards the post-E@H processing load ie. reducing 'spurious finds'.

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.