Author: Francesco Di Tolla
Date: 00:41:59 06/11/99
Go up one level in this thread
On June 10, 1999 at 10:07:25, HungryGoldfish wrote: >There is a concept which states that every 18 months there is a doubling of >processor speed, and it has held to be true for a very long time. this is called Moore law and it is an empirical observation >However, due >to quantum effects, we will soon be approaching a upper limit. As a >microbiologist, I have no idea what that limit might be, but would some of you >be willing to hazard a guess to the max. processing speed we can expect from our >computers (single processors, not multi's)? at the Empirial college they are working on devices activated by single electrons! The only limits we have from nature in principle are Heisenberg uncertainity and the speed of light. And we are far from having devices that approach that limits, very far. >And so many chips! AMD's, and those fast Celeron chips, and Pentium III's. >Assuming equal MHz ratings, are there any significant differences between these >chips? Why is Celeron so much cheaper? Marketing only marketing: the celeron was a Pentium II with the cache removed to be sold as cheap as the K6. At the beginning it was a little flop because of the total removal of the secondary cache reduced to much the speed. The new version (codname Mendocino) reintroduced a little secondary cache: 128 kb respct to 512 of the Pentium II, but (and it is a big but) it is at the same speed of the chip, while on Pentium II it was at half speed! So the celeron at 400 MHz is extremely similar to the PII: some apps are faster some slower, but the better price makes it a big seller (I've just ordered one). Why the price is so low? To disturb AMD. After all the K6-3, with a secondary cache on chip too, is a big challanger for the Pentium II and Pentium III, but at that price I see no good reason for getting anything different from a celeron: best bang for bucks by far. The few % extra performance of the other chips do not justify the extra 100's $. And with the money you same you can easely double your ram, and you know that helps the chess programs.... >Is there a chip which seems to be better with chess programs? it depends on the kind of program you use: a program specifically written for a given CPU can perform poorly on other. My guess is that if you use more programs the celeron gives highest nps/$ ratio. bye Franz
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.