Author: Robert Hyatt
Date: 16:33:36 11/15/05
Go up one level in this thread
On November 15, 2005 at 18:05:56, Vincent Lejeune wrote: >On November 15, 2005 at 17:40:54, Robert Hyatt wrote: > >>On November 14, 2005 at 14:31:48, Theo van der Storm wrote: >> >>>On November 13, 2005 at 23:24:31, enrico carrisco wrote: >>> >>>>On November 13, 2005 at 20:49:56, Jorge Pichard wrote: >>>> >>>>>On November 13, 2005 at 13:17:38, Eduard Nemeth wrote: >>>>> >>>>>>Much stronger than all other programs! >>>> >>>>According to the official website, DIEP used stronger hardware than Zappa. >>>> >>>>-elc. >>> >>>There is no statement that DIEP used stronger hardware than Zappa. >>>The official website says: >>> >>>Zappa 2xAMD dual-core Opteron, 2200 MHz >>>Diep Quad Opteron 2400 MHz >>> >>>As a first very crude step in the comparison: >>>2 x 2 x 2200 Mcycles = 8800 Mcycles >>>is faster then >>>4 x 2400 Mcycles = 9600 Mcycles. >> >>That is wrong. I'd take 4 x 2400mhz any day. >> >>9600mhz > 8800mhz in the math I use. :) >> >>Plus I'd rather have two single-core cpus rather than one dual-core. Or in this >>case, four single-cores rather than 2 dual-cores. > >Surprisingly the benchmark appears better for 1 dual-core than 2 single core in >this test : >http://www.tomshardware.com/cpu/20051107/single_core_cpus_aint_dead_yet-10.html >except for the momery transfert where the single is 2 times faster >http://www.tomshardware.com/cpu/20051107/images/image029.gif > >It's a bit puzzling :-s Depends on the test. a single dual-core can beat two single-cores, if there is not a lot of "shared data" to deal with in cache. Most benchmarks are written like that... > > >> >> >>> >>>furthermore: >>>If from >>>http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_8796_8800~97051,00.html >>>we may conclude that a dual AMD Opteron processor gives about 86% additional >>>power in comparison with single processors at the same speed, >>>Diep has an even bigger advantage. >>> >>>Still, a thourough analysis using chess programs as benchmark >>>on the actual machines is required to prove your statement. >>>Evidently the tournament organisation does not have access >>>to these machines, so it is beyond their(our) scope. >>> >>>Can anyone present real measurements? >>> >>>Theo
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.