lkml.org 
[lkml]   [1997]   [May]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: Pentium II Math Bug
Date
From
Dan Merillat wrote:

> So yes, comparing a P100 with a 486/100 is valid. It shows the P100
> is faster.

NO! The way the '486 is built it can be pumped up to about 120MHz,
The way a Pentium is built, it can be pumped up to about 200MHz.
It is valid to compare them at their max speed. In the example
you site, (P100 vs 486/100) the Pentium is both faster in
work/clock and clocks/second.

The case at hand, PPro versus Pentium-II, there is about a 5%
drop in work/clock and a 15% increase in clocks/second.

> Comparing the MMX 166 with a P100 to show the "new" pentium is
> faster is NOT valid.

If they would both be "state-of-the-art", it would.

If you are benchmarking two chips at the same clock you are comparing
the "work per clock" figure. What interests people is what they can
expect for real-life performance from a chip. For real-life, you're
going to run your new Pentium-II at 233MHz, and not at 200. Simple.

Possibly even a more interesting measure is the performance per
dollar, but as the prices continually change, your're likely to see
that type of numbers change a lot over time, so whenever you are going
to make a decision based on the benchmarks you might have to reexamine
the prices and recalcualte the new performance/dollar figure.

Roger.




\
 
 \ /
  Last update: 2005-03-22 13:39    [W:0.058 / U:0.020 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site