🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: comp.sys.next.misc
Article #4248

Re: 100 Mips Intel NeXT.

#4248
From: Christopher Songer
Date: Fri, 11 Dec 1992 18:01
38 lines
2869 bytes
In article <1g7sq8INNggf@matt.ksu.ksu.edu> ovrskeek@matt.ksu.ksu.edu (I am the OVER-skeek!) writes:
>lschultz@pdx255 (Leonard Schultz) writes:
>
>
>The P5 (I hope this name sticks, pentium sucks!) may be superscalar,
>but it is not available yet, now is it? How many more transistors does
>it need to match the performance of a MIPS r4000? Which is already available!
>IF the P5 makes its (claimed) performance figures, it wouldn't be in the
>running for fastest chip if it was released tommorow!
>

Still, Intel drives volumes high enough that while their chips may not be
the fastest per transistor, they are often the fastest per dollar. Economy
of Scale is very important here. (Hence the reason NeXT didn't go with the 
88k.) Next is not in the super fast workstation market -- they are in the 
desktop market. They need a processor base that has a reasonable tradeoff 
between speed and price. Motorola cannot not even keep up with Intel in terms 
of chip speeds and Intel lags the Riscs by about a year.

Next could adopt a "low volume" high performance Risc architecture, but they
would have to abandon the low end of the market. The Pentium will sell well and
run far faster than the current 68040 based machines. It will be out probably
a year before the 68060 and Next has already done a significant amount of
porting work to the family. The 80486 is already out and running at higher 
speeds than the 68040. The silicon is probably cheaper.

The 80x86 dominates the applications markets. It will have a huge market for 
the next decade at least. There is too much development invested in it for it
to go away overnight. If Next stays with Moto's 68K they will perform like a dog 
for the next year and a half. If they go to a Risc like the R4000 or the 88K 
they are commiting to paying premium prices for their processors and doing a port to
an architecture that does not have an installed base to insure its survival. 
Their choices seem quite limited and the Intel line seems the best. It is ready 
soon, the work for a "complete" port has been going on for some time, it is
going to be cheap and there will be a "next chip" in the family.

Anyway,
-Chris
Technical Details
Message-ID: <1992Dec11.180118.105908@lexmark.com>
Path: arc1-ca.pugleaf.net!sparky!uunet!ornl!utkcs2!emory!swrinde!zaphod.mps.ohio-state.edu!cs.utexas.edu!wupost!spool.mu.edu!agate!netsys!ukma!lexmark!songer
References: <1992Dec8.205422.746@qb.rhein-main.de> <1992Dec10.090411.12142@ichips.intel.com> <1g7sq8INNggf@matt.ksu.ksu.edu>

Show Full Headers
Complete Article Headers
Newsgroups: comp.sys.next.misc
Path: arc1-ca.pugleaf.net!sparky!uunet!ornl!utkcs2!emory!swrinde!zaphod.mps.ohio-state.edu!cs.utexas.edu!wupost!spool.mu.edu!agate!netsys!ukma!lexmark!songer
From: songer@lexmark.com (Christopher Songer)
Subject: Re: 100 Mips Intel NeXT.
Message-ID: <1992Dec11.180118.105908@lexmark.com>
X-Disclaimer: These views are the poster's and not necessarily those of Lexmark
Sender: usenet@lexmark.com (News Dude)
Nntp-Posting-Host: 9.51.7.57
Organization: Lexington, KY
References: <1992Dec8.205422.746@qb.rhein-main.de> <1992Dec10.090411.12142@ichips.intel.com> <1g7sq8INNggf@matt.ksu.ksu.edu>
Date: Fri, 11 Dec 1992 18:01:18 GMT
Lines: 41