Author: Zdroj
Date: 20:19:37 03/26/98
Go up one level in this thread
On March 14, 1998 at 21:05:03, Rajen Gupta wrote: >I have recently read a lot of rave reviews about the new Cyrix 266 >processor (some reports have put it as fast as a P11 333!)I have also >heard that in general the Cyrix is an absolute donkey at chess because >of its poor floating point performance while on the other hand the AMD >K6 is a scorcher at chess) Is this true? > >Rajen gupta I have wondered about this myself. I have a Cyrix 6X86MX PR233 chip running Fritz 5.00 under Windows 95 OSR2 with a total of 64 MB SDRAM. I have experimented with hash tables ranging from approximately 16 MB to 45MB, as any hash larger than that would result in constant hard drive access and in turn a considerable reduction in performance. It appears that Fritz performs best on my machine with 40 MB hash tables, which in itself seems a bit odd unless you consider that I might be pushing the upper bounds of my systems resources with a hash between 40-45MB . However, what I don't understand is why I am only able to achieve a maximum Fritzmark of 121 +/- 0 with 184 kN/s running the short test, and then again only rarely! Usually I can land it at 118+/-0 with 179kN/s. I have tried reloading the engine, reloading the entire program, disabling all other programs on startup rather than shutting them down before starting Fritz, but nothing serves to eek out a higher Fritzmark. What's more, the variance of these tests within a session is astonishing. On one occasion, using 40 MB hash tables, I obtained an inital value of 101+/- 10, and after 8 successive engine reloads I was able to max out at 117+/- 1. I checked the Chessbase website and discovered something interesting there on a related subject. Apparently it has been reported that Fritz may lock up on systems running a 200MHZ Cyrix chip. In closing, I can't say for sure if the Cyrix is a donkey at chess. If Chessbase states that a Pentium 200MMX turns in a Fritzmark of 119+/- 0, and I KNOW I have configured my system right down to the BIOS settings for optimal performance and I can only muster (rarely!) 121 +/- 0 on my Cyrix 233MMX, I would have to conclude at the very least that possibly some chess programs, if not all chess programs, may run faster on a particular chip...but Cyrix is not that chip.
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.