Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!mips!wyse!vsi1!hsv3!jls
From: jls@hsv3.UUCP (James Seidman)
Newsgroups: comp.windows.ms
Subject: Re: Which graphics-adapter is best?`
Message-ID: <5783@hsv3.UUCP>
Date: 14 Nov 90 18:51:41 GMT
References: <90315.094056MUHRTH@DB0TUI11.BITNET> <1511@frapper.nsc.com> <1990Nov14.030313.16366@watserv1.waterloo.edu> <1519@berlioz.nsc.com>
Reply-To: jls@hsv3.UUCP (James Seidman)
Distribution: na
Organization: Video Seven / Headland Technology
Lines: 20

In article <1519@berlioz.nsc.com> my@berlioz.UUCP (Michael Yip) writes:
>But from what I understand, Window 3.0 still does almost
>all the low/high level graphics primitives using the
>host CPU (80386) on the PC.  Although there is a TI30410(sp?)
>on the Hercule and the NEC board, I am not sure how much 
>is the onboard processor helping Window 3.0.  Well, it is
>faster  somehow.

How much of the work the host CPU does depends on how the display driver
is written.  There is a great deal of opportunity in the driver to
offload work to a TMS34010 (although I thought the boards mentioned
used TMS34020s... I could well be wrong, though).  One of the best
examples is block transfers, or BLTs, which happen every time you
move a window.  Another is pattern fills, which happen all the time.
The list goes on and on...
-- 
Jim Seidman (Drax), the accidental engineer.
"It doesn't have to work... they'll be paralyzed just from laughing at me."
							- Dr. Who, _Shada_
UUCP: ames!vsi1!hsv3!jls	         INTERNET: hsv3.UUCP!jls@apple.com