Re: A89: sine and cosine

To: assembly-89@lists.ticalc.org
Subject: Re: A89: sine and cosine
From: Zoltan Kocsi <zoltan@bendor.com.au>
Date: Mon, 29 Nov 1999 11:30:59 +1100 (EST)
In-Reply-To: <3841AB5F.641E52A6@ocf.berkeley.edu>
References: <0.5aed99ee.2572ff13@aol.com> <3841AB5F.641E52A6@ocf.berkeley.edu>
Reply-To: assembly-89@lists.ticalc.org


If you want to calculate sine or cosine on a fixedpoint range
(for example, 0-1023 representing 0-90 degrees) then the quickest
method is the table. 

If space is at permium, you should use the CORDIC algorithm, which 
takes as many iteration as many bits you need and each iteration 
consist of a shift and a few adds. That is, the CORDIC gives you 
a sine/cosine pair in the time of a SW multiplication which IMHO 
is *damned* quick. CORDIC is also suitable for many other 
transcendental functions. 

If anyone is interested, I have a PDF file which explains the whole 
thing. The paper is written from an FPGA point of view, however, the 
first half of the paper gives an extremely clean explanation of the 
CORDIC algorithm. 
I have no idea where I downloaded it from, so no URL :-( but I'm happy 
to send the file to whoever is interested in it (it's about 120K).

Regards,

Zoltan

References:

A89: sine and cosine

From: BLoWh0Le@aol.com

Re: A89: sine and cosine

From: Patrick Davidson <pad@OCF.Berkeley.EDU>

Prev: Re: A89: AMS 2.00 Released & DoorsOS Beta 2
Next: Re: A89: AMS 2.00 Released & DoorsOS Beta 2
Index(es):
- Main
- Thread