> Secondly, the net effect of this policy is to penalize those who have more
> efficient computers. I have two computers processing WU's, one of which is
> very efficient and one which isn't very efficient. In every case, my
> efficient computer's claimed credit is greater than the credit granted. To
> get credit based on someone else's less efficient processor does not seem to
> make sense. My efficient computer generates a high claimed credit, is usually
> the first to report a WU, and in every case the credit granted is less than
> that claimed. Yes I usually get more credit granted than claimed on the less
> efficient computer, but since the more efficient computer processes about 3
> times more units than the other, I will never make up for the difference
> between claimed and granted. Why not give full credit to the first computer
> to report a WU, and then average the other two?
I think you've got things backwards. I'm assuming by "effecient", you mean "faster"?
A faster computer will ask for less credit, not more. If yours is, then you're experiencing some sort of OS/benchmarking issues that aren't the norm.
It actually equals out. A faster computer will be able to process more WUs in the same amount of time as some slower systems. More lesser credits amounts to much the same as one bigger credit.
On the other side, the quicker systems often chop the much higher credits requested by a slower machine.
So really, it balances out fairly well in the terms of total granted credits across systems of various speeds.
As I understand it, the goal is to actually get all machines regardless of speed to request roughly the same amount of credit per the complexity of the WU processed. That's obviously not quite there yet, but the credits still balance fairly well. :)