[IP] more on risc etc Why the Intel switch is bad
Begin forwarded message:
From: Mike O'Dell <mo@xxxxxxx>
Date: June 7, 2005 5:07:13 PM EDT
To: dave@xxxxxxxxxx
Subject: Re: [IP] Why the Intel switch is bad
this riff against RISC machines is nonsense
the MIPS processors were and still are monsters at what they do well,
far exceeding the performance of x86 lamers. just ask cisco how
many of their products replaced funny hardware with MIPS parts
10 years ago. and the MIPS cores are still embedded in many
very high performance communications processing chips. turns out
you can process entire packet headers in the luxurious register
set in a MIPS, and with a 20K, you can do it bloody fast.
the latest round of "deep inspection" packet crackers are all
based on MIPS cores.
and the ARM (esp StrongArm) is the undisputed champion of computes
per transistor and
computes per milliwatt dissipated. there are more ARMs in the world
than x86 machines by a large margin
and the dearly departed DEC Alpha was so far ahead of its
"competitors" that is was deemed by many to be beyond what
anyone could need.
in fact, *many* engine control units in automobiles just happen
to have a PowerPC inside them, some cars have several.
and in the embedded avionics and signal processing world,
the G5 PPC with the Altavec vector unit routinely delivers
performance seen only in supercomputers not very long ago.
the count of PowerPCs in the FA-22 Raptor and the Joint
Strike Fighter is non-trivial.
so if you wanna define "success" as "running on desktops",
yup, nothing but scorched earth as far as the eye can see.
if you wanna define success in terms of taking money and
doing jobs well that x86 machines cannot touch, then
that's another matter entirely.
-mo
-------------------------------------
You are subscribed as roessler@xxxxxxxxxxxxxxxxxx
To manage your subscription, go to
http://v2.listbox.com/member/?listname=ip
Archives at: http://www.interesting-people.org/archives/interesting-people/