Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Athlon @800 Super Bypass results

Author: Ed Schröder

Date: 22:17:26 02/22/00

Go up one level in this thread


On February 22, 2000 at 20:21:01, Eelco de Groot wrote:

>On February 21, 2000 at 19:47:01, Jay Wilson wrote:
>
>>Hello All--
>>
>>I recently purchased an Athlon-700/K7M combination and was fortunate to have the
>>chipset support the "Super Bypass" feature which reduces memory latency between
>>the processor and main memory and can result in an estimated 5 - 10% overall
>>system performace boost.  In general, a 100mhz increase gives you an 5 or 6%
>>increase, so as you can see I was really excited with my new system!  I'm able
>>to overclock my system to 800mhz/SB stable with no problem and was anxious to
>>see the benchtest result of the programs that I have.
>>
>>The following are the benchtest/ratings from various programs on my computer:
>>
>>1. CM6000 @700 -2757 @800(super bypass)2775
>>2. CM7000 @800(sb) 2711
>>3. Fritz5.32 @800(sb)- 540 fritzmarks!
>>4. RebelCent @800(sb) - 2703 rating
>>5. Rebel10 @800(sb)- Benchtest 2579!
>>
>>I thought it was interesting that my Rebel benchmark result was only 3 rating
>>points behind Shroeder's #1 Super G(100mhz) and only 3 points higher than the
>>AMD 600mhz, is this some sort of pattern for AMD or is there some bottle neck
>>with the Rebel program?  Nevertheless Rebel is screaming and I've seen NPS peak
>>to as much as 380,000 Nps in the middle/end game!
>>
>>I've seen my Fritz hit over a 1.1 million nodes plus along with Junior 6 close
>>by...Shedder4 has reached close to 300,000nps in the endgame also.
>>
>>If anyone would like to see the results of any simple problem tests...forward
>>them to me and I'll share the results.
>>
>>Jaybird
>
>
>Hello Jay,
>
>Seems like pretty fast system! About the Rebel benchmark, there are only 7
>rating points between an Athlon 650 and a Kryotech Athlon 1000 MHz but that is
>maybe an artifact of the used formula at these high speeds? Just a guess. The
>solution times and NPS probably are more significant, you can check that for
>yourself, the rating result seems to be about what you would expect. Did you
>enter your result for the benchpage? http://www.rebel.nl/bench.htm Your machine
>is listed among the possible entries.
>
>Regards, Eelco.

The term "bench rating" is a bit out-dated since its introduction 1.5 year
ago. The bench rating goes to a maximum of 2600, the formula should be
changed. You better can compare the solution times than the bench rating
these days.

Ed



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.