I have noticed that, on one of my computers, CPDN seems to award ~50% more credit per second of CPU time than Predictor does. How can this be fixed?

Some people complain that CPDN awards too few credits on their machines :)

There are several sides to this. One is that the software under which each project operates (not BOINC) is more efficient on some machines than others. For example, CPDN seems to be most efficient on Intel processors running Linux, and does badly on Athlon 64s under Windows. Other projects will have very different results.

Secondly, CPDN has a fixed price system for rewarding work done because all WUs are roughly equal. Others do not, and again this tends to benefit certain machines.

Lastly, the benchmark for credit tends to be SETI, and it is against this that the CPDN credits were calculated. I am not aware that CPDN is generally regarded as too generous, neither am I aware that dedicated credit chasers have been flooding here. If your experience with Predictor is typical, it may be that it is Predictor's credit that is wrong.

---