I dont know if you have enough information from projects to calculate that : Credit by computer time used (for each project of a computer) by day.

It is difficult to know, when multi project on a same machine, if projects are just additive or antagonist. With one computer it's possible and difficult but with multiple configuration and age.

Thank s for stats

Philippe
PS : My english is not very good, but i read it better.

Here is the cross-project credit comparison, if that's the kind of thing you're interested in:

http://boincstats.com/stats/project_cpcs.php

It doesn't give you information on your computer's performance, but it does evaluate how much credit is given for a similar amount of work.

Thank s a lot, but not exactly the mind. Crossproject are interesting to compare the "production" of credit between projects.

But for a type of processor and a specific association of project, the crossproject is not enough.
Many times when i add a project, i could not know if the result is optimised. Why sometimes the credit by day falls (and stay at a lower step after a delay of observation).

It seems that some projects associated in a same computer interacts and the association are not additive, some other projects needs important or specifics ressource. And the real usage of the same computer at the same time by normal application (not calc grid) interact, and not time stable. With one computer it is not so difficult, but with many not the same and not the same usage, how to choice really between projects for each.

How many computed time for how many credit for a computer by project, is probably better to understand, interaction between project in a same computer and if performance (processor, memory, OS) are enough / optimized for the project.