As KWSN tried not to say credits are an EXTREMELY sore subject for MANY people!! Some projects use the lure of giving lots of credits to get THEIR work done faster, it works but can put a large demand on their hardware resources as people come and go as often as the wind changes! LOTS of projects just try to keep giving a happy medium of credits that keeps most people crunching for their project but there are always some people will leave and some people who will join up. Keeping a happy medium keeps the hardware requirements fairly stable but as I said that is not every projects desire. SOME projects actually make a profit off of the data, or at least say they do, so for them more is better. The Server side of the Boinc software IS being tweaked to minimize the wide variations but in order to keep the dozens of Boinc Projects all happy some tweaking of the credits MUST be allowed by the individual projects. There is a Boinc Developers Mailing List you can subscribe to that has hashed this over and over and over again many, MANY times. In short it is what it is, some projects give lots, some give very little, most are somewhere in between! If you crunch for the credits find one that gives lots and have fun, if you don't crunch for the credits find one that interests you and have fun there. Here is the current list of the Active Distributed Computing Projects, the Boinc ones are noted:
http://www.distributedcomputing.info/projects.html

Hello StopSocialExclusion,

I have avoided this thread, hoping that the people jumping in would come up with an answer. However, the thread seems to be wandering off into a discussion about the existential meaninglessness of points, which is unfortunately close to deploring the shallow attitude of crunchers who think that they deserve their fair share of points.

So I had better jump in with my estimate of the situation. From the very beginning, I have thought that you were misjudging the nature of the problem. Here are some figures that explain my estimate. Going to Statistics - By Projects I see the following values for Yesterday:

Project Years Points (millions)
ALL ---- 526 -- 583.8
SN2S --- 75 --- 89.2
GOFAM - 54 --- 60.9
DSFL --- 18 --- 21.2
C4CW -- 84 --- 92.8
CEP2 --- 24 --- 27.3

Looking at these values, it seems to me that the projects are staying reasonably close together awarding points, so any large variation is probably caused by a specific computer system. There is infinite ammunition for arguing about methods of awarding points, but instead we need to drill down to specific details about your computer and what you are running on it. For a start, please post the start of your Messages tab after a reboot. Which projects are you running? What does Results Status say about your recent CEP2 units (run time / credits)? How many do you run at a time?


Lawrence

The big picture, once more: http://bit.ly/WCGCPH

No any complicated abacus math needed from the project stats pages.

--//--

"I don't know exactly how points are calculated but I think that it should be the same for all projects at WCG so people don't chose projects after how much points they give. That it's different to non WCG projects I guess is out of WCG's control."

Some WCG projects (cfsw) run >50% faster on some operating systems (even W7x64 vs XPx64). Ideally (IMO) tasks get the same credit no matter how long they take. This encourages people to crunch such tasks on the faster systems, but not necessarily on other systems (say XP in this case), which can run other tasks.

I can't tell if any matching is being done, but in terms of run time, these two systems don't match up very well:
cfsw_ 0291_ 00291262_ 1-- 605 Valid 24/04/12 02:30:12 24/04/12 09:49:00 4.77 77.4 / 92.9
cfsw_ 0291_ 00291262_ 0-- 605 Valid 24/04/12 01:38:27 26/04/12 13:17:28 28.32 108.5 / 92.9

When only a few tasks are run the system tends to grant very different credit. This tends to equilibrate when plenty of tasks have run,
cfsw_ 0291_ 00291262_ 1-- sys1 Valid 24/04/12 02:30:12 24/04/12 09:49:00 4.77 77.4 / 92.9
cfsw_ 0016_ 00016704_ 0-- sys1 Valid 18/04/12 00:02:53 18/04/12 10:21:56 4.74 139.1 / 168.8

As for claimed credit, I guess this is also part of the new credit system, otherwise how could one system with the same setup and config ask for half the credit, for the same runtime?

I have not seen a general decrease in credit for CEP2, but Linux_x64 grants better credit than Win_x64 (33/h vs 28/h on similar systems).
I also see quite a variation in runtime (2h to 9h).

On-line or In-Line?. The credit division seems to follow the pro-ration of past based on number of jobs completed.

Wont allow myself to get exited over this... but the susceptible will do their cherry picking, inside WCG we hope. It has to be managed tight, as it does go by flavor of the day atm.

http://bit.ly/WCGCPH
http://bit.ly/WCGART
http://bit.ly/WCGFPY

*ATM* [with emphasis] on ATM, HFCC gives lowest credit per year computed, and CEP2 highest. The micrograph indicates a steep drop, the star of stars HCMD2 (for those who like hunting down the last few tasks ;>)

--//--