Author: Matt Taylor
Date: 21:34:23 12/16/02
Go up one level in this thread
On December 16, 2002 at 23:03:00, Robert Hyatt wrote: >On December 16, 2002 at 20:01:44, Matt Taylor wrote: > >>On December 16, 2002 at 11:29:22, Robert Hyatt wrote: >> >>>On December 16, 2002 at 10:46:55, Jorge Pichard wrote: >>> >>>>On December 16, 2002 at 09:33:55, Dan Andersson wrote: >>>> >>>>>It's not that simple. In every case there is complex reasons for any decision. >>>>>While it is true that the bottom line is money. In this case the P4 design is >>>>>having trouble scaling in frequency. >>>>> >>>>> >>>>>MvH Dan Andersson >>>> >>>>You are right Intel is reaching a dead end road with the P4 it took them 2 years >>>>to increase the speed from P4 1.5 Ghz to P4 3Ghz. But the upcoming technology >>>>will be based on the true Pentium V (Remember what Pentium stand for) :) >>> >>> >>>That is a normal rate of improvement. Doubling every two years. >> >>Murphy's Law which has held for a couple decades... > >I think you mean "Moore's Law". > >:) > >Murphy's law would be a whole different proposition. :) Yes. Argh. Moore is Intel's founder. Chips "working" according to Murphy's Law would be a bit of a paradox. Maybe I was thinking of software. Software is more in accordance with Murphy's Law. >>The statement that AMD or Intel is reaching a brick wall is also baseless. A >>while ago, several hardware sites said that AthlonXP wasn't scaling when AMD >>released AthlonXP 2200 processors. They've got AthlonXP 2800 processors >>available now. Only proof that the monkeys who run hardware sites lack the >>capacity to think. > >They are all running into known problems. At 10 ghz electrical properties >are far different than at 2ghz. IBM has been studying metal migration on >silicon chips at such extreme clock rates and the problems look interesting >to solve... The designs for desktop x86 chips have been iterated, but it hasn't taken as much work as a brick wall would entail. Their plans likely take into account the fact that smaller processes are necessary to run the chip efficiently at higher speeds. As for 10 GHz...well...yeah. Those're some fast gates. -Matt
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.