Date   : Thu, 30 Jul 1992 08:47:05 GMT
From   : sun-barr!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!ira.uka.de!math.fu-berlin.de!informatik.tu-muenchen.de!regent!mch@ames.arc.nasa.gov (Michael Hermann)
Subject: Re: z800?

ajv@xhost92.csd.mot.com writes:

>Did either the z800 ever see the light of day?  I stumbled across
>some old BBS messages concerning the z800, and it sounded like
>quite a neat little chip.  Did Zilog ever manufacture it?

>It strikes me that a marriage of a z800 with the multiprocessor
>support of S100 would make for a *very* interesting little
>parallel computer....

>                           Andy Valencia

The Z800 didn't ever show up (at least to customers). However
the Z280 really close to the Z800 tech. manual. I didn't do 
a real thorough comparison, but from a first look the Z280 is
only missing a timer. It also comes in a single pinout, as
compared to 4 different versions of the Z800. It inputs to
select 8/16-bit bus and Z-BUS mode.

However my Z280 didn't work as described in the manual. I 
could quite easily lock the chip (for example by enabling
DATA/INST-mixed cache).
Don't know about the current revision, mine was an eng. sample.

There is no real multiprocessor-support (at least I'm not aware
of it). That is, you're on your own to maintain cache coherence,
no bus-snooping).

The chip seemed to be quite slow (but this may have been fixed, too).
Even the data-sheet claimed a cache HIT to take about 5 cycles !!

A nice feature: It has builtin EPU-support. I think, this was meant
for devices like the Z8070-FPU (which didn't show up either), but
the protocol is quite close to the 32000-protocol and you can
attach 32081-FPU w/ some HW-support. This is a cheap yet reasonable
fast way to get floating point.

Michael


End of INFO-CPM Digest V92 Issue #78
************************************