mes@aplvax.UUCP (12/20/83)
My query about the death of the 432 newsgroup was met with a sparse
number of responses. Most of them were somewhat cynical comments
about the speed (or lack thereof) of the 432. (Nothing wrong with that.)
But there did seem to be enough genuine interest shown that I figured
a few words about the 432 'Users Group' meeting at Purdue would be in
order.
There were probably about 30 attendees, mostly from universities and Intel.
The meetings turned out to be a forum for describing work being done
with 432's. Intel also spent a good deal of time relating the in's and out's
of there new version chip set, operating system, and compiler.
The compiler is much closer
to full Ada, but still far from validation. Intel seemed to indicate that
while they were definitiely mounting an effort to validate the compiler,
they might stop short of validation if the going got too tough. There new chip
set has some notable improvements, such as higher clock rates and microcode
changes to make dealing with complex data types (i.e. having both pointers
and data ) more efficient. They also put in a context preallocation capability
to help reduce the effect of a very time-inefficient procedure call.
I can't describe all of the efforts with the 432 that were presented, but a
few stuck in my memory. U. Michigan's effort in robotics applications of
the 432 stood out as the most advanced in terms of the amount of application
code that had been produced. They had encountered innumerable hurdles
and successfully surmounted them, and had a partially working
system based on a 432 system that had at least two of what Intel calls
attached processor systems (e.g. 8086's). At U.C. Santa Barbara, a Unix
for the 432 was under development with some success. A relatively common
feature that emerged from many of the other projects was their use of the
432 as a controller, with the 'real processing' going on somewhere else.
Some were as network controllers, and at least one was as a controller of
a large system of 32 bit micro's.
Well I suppose I've used enough net time and will end here. If any that
attended the Purdue meeting can comment on new results, or anyone else for that
matter, I would welcome hearing from you, either by mail or news.
Mark Schmid
...seismo!umcp-cs!aplvax!mes