Author: Slater Wold
Date: 02:19:47 06/01/02
Go up one level in this thread
On June 01, 2002 at 02:12:12, David Dory wrote: >On May 31, 2002 at 23:54:24, ERNIE COLLADO wrote: > >> >>Would someone please tell me which bench mark tests are most relevent in regards >>to running chess software. I'm refering to the tests done between the Pentium >>and AMD chips. Thank you for your help. > >It depends on several factors, among them: > >1) What chess program are use going to want to use? TSCP, for instance, runs >slightly faster on Pentium CPU's than on the "about equal", AMD Athlon. Crafty, >however, can be compiled to run faster on "equal" Athlons. If Crafty is NOT >re-compiled, however, Athlons are still faster for Crafty, but the margin is >much slimmer. > >2) If you want to overclock - well, talk to Aaron Gordon (on CCC). He's is the >over-clocker from HELL (or is that HEAVEN?) In general, o/c'ing works better on >Athlons than on Pentium's, but it's somewhat risky. Not everyone's cup of tea. >and every mobo, memory, BIOS system is different! > >3) If you don't have a good quality board, memory, BIOS, fans, heatsinks, blah, >blah, blah, you shouldn't o/clock. One guy fried his Athlon in 5 seconds. His >own error, but that's sobering, IMO. You might even start a fire! > >In my own test - which I'll send you via email, I have a little Tic-Tac-Toe >program with full mini-max, on an somewhat larger TTT board. It fits nicely on a >floppy, and I test a lot of computers. The tests correspond well with a full >chess program test. > >In my tests, a 1700+ Athlon with 1Gb of SDRAM, is just slightly better than a >Pentium4 at 2GigHrz with any memory type or size. (This program is not effected >by nor require a HD, or a large amount of memory.) This appears pretty >consistent with other tests I've read on the net, excluding multimedia-intensive >stuff (like sound or video editing). > >Athlon's with DDR ram memory, and using 266 or higher FSB (front side bus), will >outdistance a regular P4 with any memory, even more. If the P4 has the newest >533 FSB and memory that best suits that FSB speed, that's another question. I >don't have one like that around to test, (yet!) You can read about them a little >at: > >http://developer.intel.com/design/motherbd/845GBoards/index.htm > >I'm sure the Pentium would do better with this FSB speed and the memory to use >it the best. Also, while the Athlons do better with integer speed, the Pentium's >are undoubtedly faster with multi-media work. I would guess by 5-15 percent, >depending on the exact application. > >Doing a good benchmark is a REAL can of worms. Unbelievably complex to be >"fair". To a Crafty user it's only "fair" to recompile Crafty code for each >machine. To a Fritz7 user, it's impossible to recompile the code, so that's not >"fair". There are so many driver, BIOS, video and mobo issues, you can't shake a >stick at all of them. Even the noise of the fans - especially for the hotter >running Athlons. > >And of course, everyone believes "their" way of testing is the ONLY "fair" way >to test, period!! Be ready for "hardware wars", as Slater stated. Some folk are >not interested in anything but SPEED, while others want SPEED AND VALUE. >Frankly, I like the way my Athlon's case LOOKS (it's all a light grey - even the >floppy drive and cdrom drive face). > >With the current power and stability of today's computers, I'm sure which ever >one you choose, you should be well-served. Be sure to do some performance test >on your prospective choices before buying. The differences can be large! > >David Amen. ;)
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.