Author: Charles Worthington
Date: 13:23:56 02/18/03
Go up one level in this thread
On February 18, 2003 at 16:16:56, Yen Art Tham wrote: >On February 18, 2003 at 13:47:58, Dann Corbit wrote: > >>On February 18, 2003 at 13:11:24, Charles Worthington wrote: >> >>>I have noticed a great deal of bias toward Intel products on this site so I have >>>decided to conduct an experiment using the fastest Intel Xeon technology and >>>AMDs fastest system which I assume is the 2600. If not someone please correct me >>>before I order it. I propose to test both systems using Deep fritz 7 with hash >>>settings of 32,64,128,256, 512, and 1024MB. Both systems will be equipped with >>>the best technology offered by current manufacturers. There will be no >>>overclocking. They will be tested "as is" out of the box using the Deep Fritz 7 >>>Program and Deepfritzmark test. I will post the pictures here for all hash >>>settings then we can lay the speculation to rest. Someone please leave me a post >>>telling me which AMD Motherboard would be best to conduct this test with. As for >>>Xeon testing I will conduct one test with hyperthreading enabled and one with it >>>disabled. I know everyone likes to clock the AMD processors up but I will not >>>overclock the Xeons (even if I could). They are far too expensive to be >>>tinkering with and I doubt that the Intel board would support it anyhow. So we >>>put them side to side and let them shoot it out and then it ends the speculation >>>once and for all. >> >>I think a test is a very good idea. I know Slater Wold usually has some very >>fast hardware, so he might have something to benchmark against. Perhaps a >>uniform series of tests can be created, and then run on many different systems >>so that we can create a performance table. >> >>One suggestion would be to set crafty to 100 million nodes on a dozen positions >>or so, and we could see who is fastest to target. Various commercial benchmarks >>will also be interesting (Fritzmark, etc.) >> >>Nature of the positions and available tablebase files will be a consideration, >>of course. > > >I don't think Crafty is the engine one should use to test AMD vs Intel. >It is well known that Crafty doesn't like dual AMD systems. I will conduct the benchmark test using a variety of different engines. Some Engines may be friendlier to certain type processors than others. Some may be more or less friendly to hyper-threading also. The only way to get a good comparison between processors is to use as many engines as possible then average the results into one figure.
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.