Author: Robert Hyatt
Date: 06:20:47 03/25/02
Go up one level in this thread
On March 25, 2002 at 01:33:46, Tom Kerrigan wrote: >On March 24, 2002 at 21:36:17, Robert Hyatt wrote: > >>It seems to me that "absolute performance" is the key to look at. The percent >>gain crap can be manipulated to no end. Compare _any_ processor available to >>that 800K node per second value for the 600mhz alpha to see which processor >>comes _close_. Not a one can do so... The fastest single-cpu Intel box I have >>tried is a 750mhz PIII. It hits about 320K nodes per second using the same >>version Tim used. that is 800K/320K = 2.5X faster, yet the alpha was clocked >>at 600mhz or 80% of the speed of the PIII. Normalizing for that extra 150mhz >>makes the alpha 21264 3.1 times _faster_. That is significant. And the >>alpha didn't have any assembler "boosts" which when removed, slow Crafty a >>significant amount, 15% in fact. Adding this in brings this to 3.65 times >>faster than an equally-clocked Pentium... >> >>That is both significant, not not very "muddy". > >Yes, rah rah, go Alpha. But how are we supposed to use these NPS numbers to >guess how much faster the Hammer will be than the AXP? (Which, BTW, is the issue >at hand.) > If you believe Crafty is a good representative benchmark of whatever it is you want to run on either of those two processors, then simply run it on both and see which is faster. That is how _I_ benchmark anything... I just produce the raw data and compare the two numbers... Only works for Crafty as other applications might behave differently, of course... >-Tom
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.