[RC5] stats per os/cpu - cpu id inside blocks

dan the person motion at es.co.nz
Tue Nov 16 11:29:59 EST 1999


Andre Schulze wrote:

> now it would be interesting if those x86 would be differentiated between
> amd k5/6, k7, cyrix, winchip, p1/2/3, celeron, xeon and so on.

Yes, that would be great. but i wouldn't bother with small changes in CPU
design, I see 3 options:
1) amd, cryrix, winchip, intel, and rise.
2) 386, 486, K5, K6, K7, 6x86,p5,p6,winchip.
3) am386, am486, 5x86, K5, K6, K6-2, K6-3, K6-2+,K6-3+, Athlon, AthlonUltra,
AthlonSelect, cyrix486, 6x86, 6x86L, 6x86MX,MediaGX, i386, i486, pentium,
pentiumMMX, pentiumpro, pentiumII, pentiumIII, Xeon, Celeron, Celeron(with
cache)...

So i'd so with option 2, option 3 is just overwhelming us with too much
data. And celeron xeon pII and pIII are all basically pentiumpros so it
makes sense to lump them in the p6 category.
If you went with option 3 then you'd also have to have sparc and ultrasparc,
arm and strongarm, ppc601, ppc603, ppc604, ppcG3, ppcG4... and then we might
be getting a bit overwhelmed with stats.

Of course, maybe would could go with option 3 but various stats scripts that
we run on our proxies could do a few calculations and also give us figures
for 1 and 2.

I strongly advocate moving to at least option 1 or 2.

--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list