Re: A86: ASM Programming Test - div 3

To: assembly-86@lists.ticalc.org
Subject: Re: A86: ASM Programming Test - div 3
From: ZeromusMog@aol.com
Date: Thu, 10 Dec 1998 00:50:56 EST
Delivered-To: assembly-86-outgoing@towerguard.unix.edu.sollentuna.se
Reply-To: assembly-86@lists.ticalc.org

In a message dated 12/9/98 9:41:20 PM Pacific Standard Time, electrum@tfs.net
writes:

> Hmm, a look-up table.  My favorite.  Problem is, I don't see an easy way to
>  do division with a table.  Multiplication is easy, division doesn't seem to
>  be.  I see two possible ways.  One way would be to store all numbers from 0
>  to the max.  The number at that position would be the quotient.  This would
>  take up alot of memory but would be fast, and you could also use it to
store
>  remainders (that would get a bit tricky, but possible).  The other way
(that
>  I can see) would be to have all the dividends in the table (assuming they
>  would always be multiples of the divisor).  You would search for the
divsor,
>  and the position in the table woudl be the quotient.  This would take less
>  space, but would be slower and more limited.
>  
>  If anyone knows a better way, please let me know.
>  

What I was thinking (once again, disregard the fact that I'm talking to you in
an IM right now) is that a program that repeatedly had to divide by 3 for
numbers that were, say, below 30, it would be fast and easy to have a table
that had the values in them. Depending on the purpose of the program, the
table could be pre-rounded as well. Some simple code from a newbie like me
could probably do the trick:

Code:
 ...
 ld a,27
 call Div_A_By_Three
 ...

Div_A_By_Three:
 cp 0
 ld b,0
 cp 1
 ld b,0
 cp 2
 ld b,0
 cp 3
 ld b,1
 cp 4
 ld b,1
 ...
 cp 27
 ld b,9
 ...
 ret

I'm sure there's a faster/more efficient way, but you get the idea... (perhaps
you can show me something cool by optimizing that? :)