Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!usc!jarthur!nntp-server.caltech.edu!toddpw
From: toddpw@nntp-server.caltech.edu (Todd P. Whitesel)
Newsgroups: comp.sys.apple2
Subject: Re: SuperConvert/SHRCONVERT
Message-ID: <1991Jan26.144248.8599@nntp-server.caltech.edu>
Date: 26 Jan 91 14:42:48 GMT
References: <1991Jan25.145637.13116@infonode.ingr.com>
Distribution: comp
Organization: California Institute of Technology, Pasadena
Lines: 30

unknown@ucscb.UCSC.EDU (The Unknown User) writes:

>Not likely, as it seems people were DOUBLING SHRConvert (2.1?)'s speed
>with purportedly "simple" basic/machine language hacks..

That was my claim-- I wrote a sleazy GIF parser/decoder in Basic/Assembly,
which was 2-3 X faster than SHR convert with the 320x200x16 test picture
I was using. I just hauled it out to do some sample benchmarks (turned off
the Zip so these are all unaccelerated tests):

program		320x200x16	320x200x256	800x600x256

Basic/ML	13 sec		13 sec		No thanks
GIF3200v0.2	11 sec		12 sec		1:05
LHG 0.whatever	2 sec (!!)	4 sec		25 sec

I tried to time the actual picture decompression, and not the interface stuff
(which varies, since GIF3200 is a desktop program and LHG is still a raw EXE
or S16 which only responds to finder messages).

Those of you who are bugging me for beta's of LHG, sit tight, the color
crunching is not even GIF3200 Pre Scale quality yet. (Jonah, don't send me
anything; I want to figure this out myself.) After I get some sleep I will
be rewriting a much better color cruncher. At that point it will be really
tempting to release LHG betas (the real version will be shareware or something
because this is finally something worth money and I am really broke this term).
However, I need to figure out what to do about copyrighting it first.

Todd Whitesel
toddpw @ tybalt.caltech.edu