Author: Jason Michael Rogers
Date: 15:52:44 08/16/03
Go up one level in this thread
On August 15, 2003 at 18:10:17, Aaron Gordon wrote: >On August 15, 2003 at 16:27:54, William Penn wrote: > >>I have had a 500MHz Celeron/256MB system for quite a few years. I just bought a >>new 2.6GHz Celeron/512MB computer and was expecting my chess software to run >>about 5X faster. It doesn't!? After considerable comparisons and benchmarking, >>it's clear that it runs about 2.2X faster. My tests & benchmarks are based on >>kN/s in infinite analysis mode using various practical chess positions and the >>Shredder 7.04 UCI engine in the CB GUI. >> >>What is the most important single other factor (besides processor speed) that >>would increase the speed of this engine in terms of kN/s? I'm sure that more RAM >>wouldn't do very much. Is it "BUS speed", or "memory speed", or "processor cache >>size", or "processor type", or what? I'm not interested in minor factors, just >>the most important one. >>WP > >This is what Intel counts on, people only looking at the MHz. Unfortunately you >got stiffed by Intel. Lots of benchmarks show the Celeron 2.2 overclocked to >well over 3GHz losing to a Duron 1.2GHz in most things, which should scale to >about what you have been seeing. > >The Pentium 4 is the same way (just not as bad). The best bet would be to get a >2800+ Athlon XP chip (if you don't want to overclock) as its the fastest MHz >wise (2250MHz) of any of the Athlons for chess, and faster than any of the P4s >out today for chess as well. If you'd prefer lots of speed and some >overclocking, try one of my 2.4-2.5GHz Athlon XP pretested chips.. They'll put >you up near P4-4GHz speed chess-wise and blow a P4-3.2 out of the water at >everything else. > >I've already done the tests vs a P4-2.53 @ 3.32GHz in multiple applications >(divx, mp3 encoding, compiling, encryption/decryption, sciencemark, chess, etc), >the P4-3.32 lost by a significant ammount in most of the tests. Plus, at $79 for >the 2.4GHz Athlon XP (3200-3400+ IMO) isn't bad at all.. > >Here are some chess benchmarks: >http://www.newageoc.com/crafty/bench.html > >Note: I don't have any freon-cooled Athlon systems tested yet (at 3.4GHz), so >that would beat the P4-4GHz w/ HT by quite a bit. The top Athlon on there is the >2.5GHz and that is AIR cooled. The top P4 is FREON (-50 celsius) cooled. > >You can see from here the 2.2GHz Athlon (Athlon XP 2800+ is 50MHz faster) beats >a P4-3.25GHz. Celeron 1GHz (p3 core) is faster than a Pentium 4 1.5GHz, and an >Athlon XP 1.86GHz (about like a 2200+) is almost equal to a P4-3GHz. Even the >Pentium 3 1.12GHz beats a P4-1.6GHz. Good job Intel. :) > >Intel has it pretty sweet, though. Only company I know that can produce a >product sigificantly worse than the last one and charge people 4-5 times more >and have people actually buy it. Do you care to qualify that statement about the P4 being "worse" than the P3? You make alot of very confident assertions about speed and performance, but you don't back it up with any techical specification comparisons or bench test results. For example, why is 2.4-2.5GHz Athlon "up near P4-4GHz speed chess-wise"?. > >My recommendation is take it back and build an Nforce2 based Athlon XP 2800+ >system. If you'd like to save a little money and get some extra speed, head over >to www.newageoc.com. If you don't know how to overclock I can manually 'lock' >the chip at 2.4GHz that way you don't need to do anything but drop in the chip >and put on the heatsink. :)
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.