On 12 Jun 99, at 7:49, Gary Diehl wrote:
> Maybe I'm just being silly but I wonder...How come I'm not on
> the (longer) top producer's list (albeit near the bottom of the list).
> Doesn't the list update automatically?
Should do, however sometimes (for a reason I don't know) it's
anything up to a day old.
>
> I had my system overclocked to P2-333 for 21 of the 27 days, and then I
> put the clock speed back due to heat and finished the test at P2-266 for
> the remaining 6 days.
>
> 123.49 days = 1 p90 year at 266mhz (365/(266/90))
> 98.64 days = 1 p90 year at 333mhz (365/(333/90))
> (these assume that MHZ is the deciding factor of speed, i.e. my 266 is
> 2.955556 times faster than a p90)
>
Actually it's irrelevant what speed your processor is working at so
far as credit is concerned. George Woltman (the author of the
software) benchmarks the program using a "standard" system, which
happens to be a P90 he has lying around somewhere (was probably state-
of-the-art when he got it). For exponent 7.87 million the transform
size is 448K, so the standard time per iteration is 1.498 P90 CPU
seconds (from http://www.mersenne.org/status.htm). So the credit is
(approximately) 7.87 * 1.498 * 10^6 P90 CPU sec (use the _actual_
exponent less 2) = 11.79 * 10^6 P90 CPU sec = 0.374 P90 CPU yr.
P6 architecture CPUs (PPro, PII, Xeon, Celeron, PIII) are much better
than you'd expect from raw clock speed, due to the improved
pipelining and branch prediction in the chip design. Conversely AMD
K6 chips aren't as good as you'd expect (no pipelining in the FPU at
all), and Cyrix chips are plain miserable.
> Did the change of MHZ on my config mess up the primenet server?
No. You might want to check the setting in the Prime95 setup
(Options/CPU) but this is cosmetic rather than meaningful to the
server.
Regards
Brian Beesley
________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
