Author: Slater Wold
Date: 13:25:17 10/08/02
Go up one level in this thread
On October 08, 2002 at 15:00:52, Yen Art Tham wrote: >On October 08, 2002 at 14:05:59, Robert Hyatt wrote: > >>On October 08, 2002 at 13:54:16, Rex wrote: >> >>>Can someone show me just 1 thats ONE example where this 8 CPU FRITZ program >>>gained an advantage over lets say a single cpu in this GM ver Comp match. >>> >>>I believe the dual/8 cpu is hog wash and just makes the program not play >>>stronger but rather widens the TREE search wider/faster which causes the >>>efficiency to decrease. >> >> >>We've all posted plenty of results showing that two processors are faster than >>one by a >>significant amount. >> >>I can post some results if you want, but there is little doubt that two >>processors are better >>than one, by a significant amount, if all run at the same clock speed... > > >Just curious, if both a single and a dual gets the same nps, >would they be equally strong or if one has an advantage over >the other? >Are there any games to prove one way or the other? > >yat Interesting question. For example: Say you have a P3 733 getting 600k nps. And a Dual Celeron 400 also getting 600k nps. (Not actual figures, just for example.) Which would get the solution first? My opinion: The single CPU system. Simply because it almost ALWAYS (99% of the time) takes more nodes to solve a problem on a dual CPU machine. However, there are weird instances where it takes a LOT less nodes (sometimes 3-4 times less) to solve the problem. This mostly based on random branching. Hence the speedup arguement.
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.