Path: utzoo!utgpu!news-server.csri.toronto.edu!torsqnt!hybrid!torag!utdoe!generic!pnet91!ericmcg
From: ericmcg@pnet91.cts.com (Eric Mcgillicuddy)
Newsgroups: comp.sys.apple2
Subject: Re: Animation (self-modifying code)
Message-ID: <661@generic.UUCP>
Date: 26 Apr 91 23:30:13 GMT
Sender: root@generic.UUCP
Organization: People-Net [pnet91], Etobicoke, ON
Lines: 26

>slowing you down.  That is, if you're using indirect indexing to
>load and store bytes to and from the screen, you may wish to change
>this to absolute indexing and self-modify the base address for each
>new row.  The same applies if you're using indirect indexing for
>accessing the shape tables.  Self-modifying the code is always faster
>when dealing with large amounts of data.
>
><\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/><\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/
><  Philip J. Stephens                ><   "Many views yield the truth."      

You do not need to use indirect indexing, absolute indexing from a base of
0000
is more than adequate. setup time is about as long and the data bank 
register must be the same (althoug I suppose absolute long addressing could be
used for the shape table) for the shape table. Using the X register you can
access any byte in a given bank and this is currently adequate to support the
GS SHR (but throws out my earlier comment about the maintenance in the
future). The cost of adjusting the X register for the next line may be as
simple as INX or as complex as TXA, CLC,ADC #onerow-size_of_sprite, TAX.
Regardless it is less than or equal to self modifying code and more
"standard".

Eric

UUCP: bkj386!pnet91!ericmcg
INET: ericmcg@pnet91.cts.com