Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Introducing "No-Moore's Law"

Author: Matt Taylor

Date: 06:08:25 03/01/03

Go up one level in this thread


On March 01, 2003 at 03:11:43, Jeremiah Penery wrote:

>On March 01, 2003 at 02:44:08, Matt Taylor wrote:
>
>>The increase of computing power per processor is tied directly to the increase
>>of transistors/mm^2.
>
>It's not an exclusive relationship.

Do you wish to argue that processor power is NOT primarily influenced by
transistors/mm^2? I also have pointed out that other factors are influential.
Now, which drives more performance? Processors have extracted performance from
pipelining, superscalar execution, etc. How often do designs change to yield
more performance? ...

The Pentium Pro launched somewhere around 180 MHz (upper end) and was scaled all
the way up to 1.4 GHz. That is nearly a 10x increase in speed. The only
architectural change through this time that I can think of is SIMD. MMX and SSE
are great buzzwords, I suppose, but they do very little in the way of creating
real performance for general applications, and they are littered with problems
(such as lack of support, lack of compiler support, competing extensions, etc.).

Enter Pentium 4. In 0.13 micron form, the Pentium 4 scales to 3.06 GHz. This is
roughly equivalent to a 2 GHz Pentium 3. Is there a speed increase over the 1.4
GHz Pentium 3? Yes. Is it 10x? No.

-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.