Author: Anthony Cozzie
Date: 07:17:04 06/15/04
Go up one level in this thread
On June 15, 2004 at 00:06:12, K. Burcham wrote: > >lets see, if each core has 900+ pins and 2 cores per processor in a SMP machine, >that is 3600+ pins. 3600+ pins running a chess program could be a considerable >speed change in KNS. I will believe it when I read benchtest reports at top >sites. > >kburcham > > >http://zdnet.com.com/2100-1103_2-5233470.html > >The Opteron and Athlon 64 chips will differ in cache size, frequency and the >number of input/output links, but the chips' core will functionally be the same. >There are three iterations of the dual-core Opteron, code-named Egypt, Italy and >Denmark. A single dual-core desktop chip goes by the code name Toledo. > >Dual-core processors allow chip designers to ameliorate the growing problem of >power consumption, Crume said. A dual-core chip running at a lower frequency can >consume less power but provide more performance than a singe chip running at >multiple gigahertz. > >Contrary to conventional wisdom, software compatibility is not a significant >barrier to adoption of dual-core processors, he said. Windows XP and Windows >Server 2003 are already "threaded," which means that they can divide up tasks >across two processors. A substantial number of server applications are threaded >as well. > >The practice is far less common with desktop applications, but if the operating >system is threaded, a PC can increase its performance, because it can do two >things at once. > >AMD's first dual-core chips will be made on the 90-nanometer process, which AMD >has recently started using.
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.