pending but no credit

Eyespy27
Eyespy27
Joined: 19 Feb 05
Posts: 4
Credit: 1129
RAC: 0

Thanks! So far, all the

Thanks! So far, all the results I returned were "valid," so I guess everything's working ok. I was just making sure my crunching wasn't for nought.

Thanks!

Jim Baize
Jim Baize
Joined: 22 Jan 05
Posts: 116
Credit: 582144
RAC: 0

I am going to try to answer

Message 3793 in response to message 3790

I am going to try to answer your question just a little bit differently, Eyespy.

If I am reading your question right, you are not asking about the claim / granted credits, but rather, you are asking what happens to the work that you turn in for evaluation.

> I'm a complete neewbie, but just want to clarify what happens with the data I
> crunch. Personally, I couldn't care less about credits, but I can also see
> where others could have a lot of fun with them. Either way, what about the
> bottom line, data integrity?

The credits are a measure of how much work was done. It does not directly reflect how much science you actually accomplished, although there is a correlation.

>
> Do credits and the scoring system have anything to do with the actual
> usefullness of the packets I'm crunching? Are the results of the WU thrown out
> if for whatever reason, my packet is the high or low one?

When the quorum of results are returned, the results are examined and compared. The results must lie within a certain tolerance of each other to be considered valid. I have seen some discussion on the accuracy of this method, so it may change in the future, but for now this is how it works. It is assumed that if multiple results lie within the given tolerance then the result must be valid. Therefore one result is picked out of the quorum and that is the result that is used for any further investigations. So, even if your result is not inserted into the DB for further eval, you are still doing useful work if for no other reason than just to confirm the accuracy of the given result. This process may vary a little from project to project, but this is the basic premise of the process.

>
> If a WU is discarded, is that just as helpful to the project as one that is
> not, or was it just a waste of 10 hours of my PC?

This question is a little more tricky. It depends upon why the WU was discarded. Some computers are overclocked or have hardware problems that cause innaccuracies. Sometimes a computer can just throw a random error for no apparent reason. If this is the case then yes, the WU was in vain.

HOWEVER, there are times, especially in these early days / weeks / months of the projects that a WU can be returned as invalid or because of an error but still be useful. It won't be directly useful for the science end of the project but what it does do is gives the devel team information and clues to help track down bugs in the code. I've seen this happen numerous times.

> If for some reason my claimed credit is the highest, does that mean that there
> is an error in my computations? If there are errors in the computations, is
> that because of a coding error for my machine/platform, or did my computer
> "make a mistake?"
>

You will probably bounce around from highest to lowest to somewhere in between. As long as your results are within the given tolerances there are no major errors. The differences in claimed credit have a lot to do with your hardware (amd vs Intel) and your OS (Windows vs Linux). Oh, I just about forgot, there is a problem with the claimed credit between Linux and Windows. There are other discussions that can explain it better than I can so I'm not even going to try. Look here and over on the SETI boards. There is a wealth of information over there. Another way to increase your CLAIMED credit is to "roll your own" core client. You can download the source code and compile it yourself on your machine with the optimization settins you want. This will help your claimed credit score some.

I hope I helped to clear things up a little.

Jim

wow, my fingers are tired now. LOL

PS... I also endorse Paul D. Buck's website. He has over 700 pages worth of infomation on Boinc, SETI, Einstein, and the other projects. Take a gander over there. You won't be disappointed. Make sure to drop a line to Paul telling him that you saw his site. he loves to get feed back on his site.

Jim

Comment viewing options

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