Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!jarthur!nntp-server.caltech.edu!toddpw
From: toddpw@nntp-server.caltech.edu (Todd P. Whitesel)
Newsgroups: comp.sys.apple2
Subject: Re: ACCELERATOR - ZIP OR TRANSWARP?
Message-ID: <1991May16.100046.6131@nntp-server.caltech.edu>
Date: 16 May 91 10:00:46 GMT
References: <730@generic.UUCP> <1991May16.074027.28758@ux1.cso.uiuc.edu>
Organization: California Institute of Technology, Pasadena
Lines: 23

dat33228@uxa.cso.uiuc.edu (Derek A. Taubert) writes:

>Ok, ok...  Sure, it runs cooler, but a bit faster?  Come on folks...  Until
>someone produces some HARD evidence that a 7 MHz Zip runs faster than a 7 MHz
>Transwarp GS, I would really appreciate it if we layed off of the AE bashing.

The Zip uses latch-on-write for slow bus writes. This means you can write
data to the slow bus at a full 1 mhz very easily, as long as you are running
entirely in the cache except for the write. I've clocked the transfer rate
at 1 mb/s and the performance for longer transfer loops was better than I
expected (i.e. you will still get .7 mb/s for loops that I thought would
only get .5)

This makes stack-romping graphics code (like the scrolling routines in LHG)
_scream_. Also, Apple's routine for clearing dialogs and windows to white
(the ONLY optimized rectangle primitive in QD it seems) uses a loop that is
tight enough to write to the screen at somewhere between .5 and 1 mb/s --
it's really neat to have large dialogs get cleared with a single screen tear.
I just wish they'd quit using super-slow genero-loops to draw things like icons
and the background picture!!

Todd Whitesel
toddpw @ tybalt.caltech.edu