Xref: utzoo comp.sys.dec:2222 comp.unix.ultrix:2242
Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!purdue!decwrl!granite!vajra!rwood
From: rwood@vajra.dec.com (Richard Wood)
Newsgroups: comp.sys.dec,comp.unix.ultrix
Subject: Re: 3100 mix and match
Message-ID: <703@granite.dec.com>
Date: 29 Nov 89 20:49:34 GMT
References: <4305@helios.ee.lbl.gov> <1989Nov28.191803.18970@sun.soe.clarkson.edu> <8732@cbmvax.UUCP>
Sender: news@granite.dec.com
Reply-To: rwood@vajra.dec.com (Richard Wood)
Organization: Digital Equipment Corporation
Lines: 23

In article <4305@helios.ee.lbl.gov>, envbvs@epb2.lbl.gov (Brian V.
Smith) writes:
> What were the reasons for not using the GPX processor in the DS3100?
> To go to a dumb frame buffer driven by the CPU seems like a giant step
> backward.

The graphics performance of the DS3100 is substantially faster than the
GPX subsystem would allow, even without an accelerator.  While the GPX
subsystem would have reduced the CPU load somewhat, it also would have
forced an upper limit on the graphics performance of the machine.

Considering that adding in the additional subsystem would take
additional time, increase cost and complexity - all without adding
substantial performance gains - it became obvious that the solution was
a dumb frame buffer.  The philosophy was "we got this extremely high
powered chip in the middle of this design, let's make it work for a
living."  Thus the elimination of DMA and graphics accelerators made the
system much simpler to design, less expensive, and much more reliable
when complete.

-- ---------------------------------------------------------------------
Richard Wood     Corporate Worksystems Team      Digital Equipment Corp.
========================================================================