Author: Matt Taylor
Date: 13:36:34 02/19/03
Go up one level in this thread
On February 19, 2003 at 14:52:41, Aaron Gordon wrote: >On February 19, 2003 at 08:09:12, Charles Worthington wrote: > >>On February 19, 2003 at 00:35:21, Aaron Gordon wrote: >> >>> >>>> >>>>Interesting....the 2 fastest cpus on your list of overclocks are still >>>>intels...not amds. Not bad considering most like to say that intels don't >>>>overclock well. >>> >>>As I said before, that was a special cpu running with Freon. I know someone >>>running 3.1GHz with freon on an AthlonXP, if an 2.52GHz XP gets 1.57 million >>>nodes/second then you can roughly figure a little over 1.93 million >>>nodes/second. Still faster than a P4-4GHz with HT. :) I am in the process of >>>getting in touch with him, I'll get him to run a Crafty benchmark whenever I get >>>an email back. Also, when you compare overclocks it's good to compare >>>air-vs-air, liquid-vs-liquid, freon-vs-freon, etc. >> >> >>Actually though due to the differences in processors its hard to conduct a fair >>test where overclocking is concerned. That was why I proposed the test under >>factory conditions to begin with. Neither AMD nor Intel recommend overclocking >>their products and both will boot your warranty straight out the window for >>doing so. I understand that many users like to tinker with their cpu clock speed >>but most consumers won't do that. That's why I consider the out-of-the-box test >>to be more accurate. When we buy cars we buy them on the basis of performance >>off the showroom floor not on the basis of how fast our mechanics can make >>them. Same with cpu's. > >L2 & Memory really don't affect Crafty performance at all. A Tbird 1.33GHz with >SDRAM gets almost identical (Within margin of error it would appear) NPS to an >AthlonXP 1500+ (1.33ghz) with DDR. With this in mind a clock for clock >comparison should be alright considering the other things aren't much of a >factor. To me that implies dependency on memory latency. Since DDR isn't really any faster than SDRAM (just more bandwidth), the results would be the same. If Crafty blows the cache often, L2 won't make much difference either. -Matt
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.