Author: Vincent Diepeveen
Date: 03:33:43 01/24/04
Go up one level in this thread
On January 23, 2004 at 12:12:24, Anthony Cozzie wrote: >On January 23, 2004 at 00:06:34, Earl Klugh wrote: > >> >> >> >> How does the hardware of ICC monster Pure-Harcore compare to the used in the >>Kasparov vs X3DFritz Match? Here is it's notes >> >> >> >> >> Statistics for Pure-Hardcore(C) On for:31:01 Idle: 4 >> >> rating [need] win loss draw total best >>Bullet 2743 546 0 9 555 2743 (22-Jan-2004) >>Blitz 2988 986 11 43 1040 2990 (26-Dec-2003) >>Standard 2723 455 4 7 466 2723 (22-Jan-2004) >> >> 1: Dual Xeon 2.8ghz, overclocked to 3.15. FSB 375MHZ. Winfast A310TD 256MB >>nVidia GeForce FX5600. Using an ASUS PL-DL Duluxe motherboard. 2GIG 400MHz DDR >>Ram. 2 60GB Hard Drives in a RAID 0 configuration. >> >> >> >>This machine is a Monster, I wonder if it could draw Garry kasparov like >>x3dfritz. > > >Overclocking a dual is a _bad idea_. A _VERY BAD_ idea. > >I'd say that if its stable (unlikely) its about 65% as fast as the X3D machine. > >anthony It's a very fast machine, because overclocking the bus is kicking *butt* for parallel programs. The weakest chain for running 4 threads at a dual is *memory speed*. WCC machine is 4 x 2.8Ghz with 533Mhz bus or so and 133Mhz RAM, no HT. Now we have 4 x 3.15Ghz HT bus speed = 3.15 / 2.8 = 1.125 x 133mhz = 150mhz and that with cl2 memory most likely. That *kicks* butt. For parallel software it really kicks butt. Of course overclocking is something you need to know what you do. Ron Langeveld ran DIEP for me at a slightly overclocked P4 and it is *a lot* faster than a normal P4. Really *a lot* and HT works a lot better too. Just 10% busspeed with cl2 memory matters really a lot, the P4 chip has such tiny caches that this really is a weak chain.
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.