From: salnick@dejavu.spk.wa.us (Bob Salnick)
Subject: Re:  VGA from hell!
Date: 31 Oct 91 05:44:03 GMT



In article <1991Oct28.001450.4768@milton.u.washington.edu> rcd@raven.eklektix.
com (Dick Dunn) writes:

>dstamp@watserv1.waterloo.edu (Dave Stampe-Psy+Eng) writes:
>>OK.  I ran some timing tests on my somewhat-super VGA Paradise card.
>>Talk about wait states!...
>
>This is common on VGAs.  It's one of those things you get out of hardware
>design driven almost entirely by cost...and it's quite an unfortunate
>thing, because it usually presents a "brick wall" for video performance
>that happens before you get enough to come close to animation.  It's a
>disappointment, and also a caution to folks who'd like to build low-end VR
>systems.  I've done a little dabbling with this stuff, and talked to folks
>who've done a lot.  10-12 bus cycles seems fairly common.  An AT bus cycle
>is 125 ns; the normal min is 3 cycles for a transfer.  Your numbers sound
>right (where "right" means "what other people have found", certainly not
>"what they should be":-).  Yeah, there really IS twice as much delay as
>time-spent-doing-useful-work-on-the-bus.
>
>The sticky part is that these delays don't seem to show up in the specs for
>the VGA cards--"everybody does it and nobody talks about it."  I don't know
>what you'd do to find faster VGA cards short of beg/borrow/buying a bunch
>and testing them.
>
>The one hope is that video cards might eventually be built which keep up
>with the bus, and which aren't too outrageously priced.  But VGA has got so
>much goo in it (with more accreting on each "super/extended/hyper/turbo"
>new product) that we'd be much better off with a "dumb frame buffer."
>That's outside VGA-land.


I wonder what the comparison would be on an Amiga?  Has anyone ever done 
one?

bob
--
Bob Salnick, Spokane,WA |  USENET:     oliveb!isc-br!tau-ceti!DejaVu!salnick
  Amiga 1000, WB 1.3    |  INTERNET:   salnick@DejaVu.spk.wa.us
      WA9BVE            |         (FireStorm '91 survivor)



[MODERATOR'S NOTE:  For those engaged in the VGA debate, I would appreciate
email sent to my private account regarding the propriety of this discussion
on sci.virtual-worlds, and why it should not be shunted over to comp.
graphics or comp.graphics.visualization, where the readership might be
better.  Thanks. -- Bob Jacobson]