Array Processors and UNIX

[Martin Taylor]

In article <dcdwest.170> phb at dcdwest.UUCP (Peter H. Berens) writes:


>  FPS and CSPI are the oldest of the array processor manufacturers.
>        FPS has always been very successful in marketing against CSPI and
>        thus has generally been considered the leader in the field of
>        array processors.  I too have heard the rumors about CSPI and
>        their support of UNIX with their mini-map array processor.  I would
>        only caution people to check it out throughly before buying it.
>        Insist on talking to people who already are running it under UNIX.
>....
> Currently I am not aware of any array processor manufacturer that
>        provides software that will run under UNIX.  To the best of my
>        knowledge my company (APUNIX) is the only source for UNIX array
>        processor software. 

It may be correct that no array processor MANUFACTURER provides UNIX
support, but APUNIX isn't the only place that could provide it.  Several
years ago (1978?) we commissioned HCR to develop UNIX support for the
CSPI MAP-300.  This included calls to their SNAP-II routines, plus
a pre-processor that allowed C-like calls to be written in your C programs
(it also preprocessed Ratfor and Pascal, using a compile-time switch).
The system originally ran under V6 on a PDP-11/34, and now runs on
a Perkin Elmer 3242 under Edition VII.  CSPI were at one time interested
in marketing the product, but decided that UNIX didn't represent a
big enough market. (Maybe that's why FPS outsells them; they don't
have much insight :-).  We didn't have full debugging support, or
C language access to the assembler code of the various internal processors
of the MAP-300, but it doesn't seem to have mattered.

I don't know whether HCR still maintains or sells this, but it was
a product at one time (perhaps it still is; I just don't know).

Incidentally, I can't speak to present-day array processors, but at that
time I felt CSPI had a better product for general purpose high-bandwidth
computing than did FPS.  The AP-100 was based on a rather long pipeline
that was a bit tricky to program, whereas the CSPI MAP-300 was based
on concurrent processors that were not pipelined, and could be programmed
as SIMD or MIMD, as well as totally independent I/O processors for
communicating with the host or with the outside world.
-- 

Martin Taylor
{allegra,linus,ihnp4,floyd,ubc-vision}!utzoo!dciem!mmt
{uw-beaver,qucis,watmath}!utcsrgv!dciem!mmt