This 'problem' I have I had on the BOINC Beta as well, and have it here on the Seti Live site as well as on Predictor@Home. I am claiming huge credits, getting granted whatever the next machine gives for a work unit.

Maybe it's time someone looked into the claims a P4 Celeron makes, as that is the CPU I am using. It is a P4 Celeron 2.3 that I overclocked to 2.76GHz (120x FSB).

--- - 2004-07-06 03:27:22 - Running CPU benchmarks
--- - 2004-07-06 03:27:23 - Suspending computation and network activity - running CPU benchmarks
--- - 2004-07-06 03:28:24 - Benchmark results:
--- - 2004-07-06 03:28:24 - Number of CPUs: 1
--- - 2004-07-06 03:28:24 - 1782 double precision MIPS (Whetstone) per CPU
--- - 2004-07-06 03:28:24 - 3757 integer MIPS (Dhrystone) per CPU
--- - 2004-07-06 03:28:24 - Finished CPU benchmarks

As you can see, the Whetstone and Dhrystone calculations are a bit off for a CPU with only 128K L2 cache. I am slower than an equal P4 at crunching (about 6.5 to 8.5 hours for a Seti WU. 4.5 to 6 hours per PAH unit, about 50 minutes per charmm unit), claiming around 65 and above per unit.

Oh, before you ask me to normal clock my CPU, I did that this weekend. My Dhrystones stayed the same, while my Whetstones went down by about 200 MIPS.

But could someone look into this? Could it be the P4 Celeron structure that is throwing off the benchmarks? These Celerons are based on the P4 Northwoods.

P.S: yes, I have asked the same question on the PAH Q&P forums. Maybe that both Dev teams want to look into this.
----------------------
Jordâ„¢