🚀 go-pugleaf

In article <2AF18671.6642@deneva.sdd.trw.com> thomsen@spf.trw.com (Mark Thomsen) writes:
>andrew@cubetech.com writes
>> All you have to do is look at the number of vendors that are using the
>> processors and the volume of the markets in which these vendors
>> computers are going to be sold:
>> 
>> 88k:
>>      DataGeneral - pretty low volume workstations and servers.
>>                    Good machines though...
>>      Motorola    - low volume workstations
>
>And lots of embedded applications in industry. I know full computer systems on  
>desktops are fun, but there seems to be a continuing arrogance thinking that  
>the volume of processor chips is dominated by desktop computers. A lot of  
>processors - the majority three years ago - are embedded inside non-computer  
>products and equipment. From three years ago I saw data that indicated that  
>Motorola does very, very well in the embedded processor industry.

Motorola makes a wide range of microprocessors explicitly for embedded
applications.  The ones I am aware of are based on the 6800 and 68000
families.  Intel has the i960 series, AMD has a slew of 29000 varients,
and the R3000 has a wide range of varients for embedded applications.
What varients exist of the 88100, and how do they sell against these other
chips?

>Putting the software questions aside for this posting, the low quantity of  
>processors in the desktop computer system market does not directly imply that  
>the vendor will not continue to strongly support the processor evolution.
>
>Mark R. Thomsen

Has the 88K family evolved beyond just the 88100/88200 pair?  Is the
only planned evolution the 88110?  The fact that almost every maker
of 68000 workstations (Sun, Apollo, HP, SGI, Sony, etc...) aside of
Tektronix didnt go with the 88K line was a bad omen.  The R3000 and
the 88K were neck and neck at the same clock rate; the improved R3000A,
R4000 (and soon the R4000A) are out, and Im still not aware of when the
88110 will ship in quantity.  It would appear the vendor currently does
not strongly support the processor evolution.

The reason any of this is important in this forum has to do with NeXT's
choice of processors, and both the perception and reality of NeXT's
ability to remain competative.  Over the course of a year, NeXT will
need a 40 SPECmark box to compete with Apples R601 offering, and it
has to be less than twice the price.  It will need to evolve at the
rate of the R60x series (ie 4 chips promised over a two year period.)
While Apple promises $3,500 R601 systems, Intel and Microsoft are
moving with the Pentium CPU and a multiprocessing operating system.
NeXT also needs to have strong low cost offerings to compete against
current PCs.  486DX2s at 66MHz, with 8M RAM, 200M disk, and accelerated
local bus 8-24 bit graphics, and an array of expansion slots are selling
for around $2,500.  NeXT needs to have their 8/200 Mono slabs selling
for $2,500 on the street.  No expansion slots, no color, no graphics
acceleration, a small power supply, and a fixed frequency greyscale
monitor which should be no more expensive than the color multiscanning
SVGA displays should make up for at least some of the cost of the
licenses behind NeXTstep (AT&T UNIX, Adobe Postscript, etc...)

I realize that NeXTstep has a lot of added value, but I also realize
that there are a LOT of business folk who dont even know of NeXTs existence.
NeXT sold 80K machines in 3 years, and Microsoft sold 20K copies of Windows NT
in three months, and it isnt even a product!  Perception guides a lot
of buying decisions, and NeXT needs to choose a CPU architecture that
is perceived as advancing fast, to have a broader product offering, and
to generate volume across the board to be perceived as a viable solution
for mainstream problems.  Lowering the cost of NeXTstep 486 could advance
these goals as well.  -Steve
-- 
~~~~~~~~~~~~~~~~~~~~
Stephen E. Halpin                    halpin@vsg1.dsg.dec.com

"You might just have to waste your life just to live."  - Soul Asylum