home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.os.os2.programmer:7349 comp.graphics:13438 comp.os.msdos.programmer:11823 comp.sys.ibm.pc.programmer:744
- Path: sparky!uunet!pipex!bnr.co.uk!uknet!canon.co.uk!ads
- From: ads@canon.co.uk (Adam Billyard)
- Newsgroups: comp.os.os2.programmer,comp.graphics,comp.os.msdos.programmer,comp.sys.ibm.pc.programmer
- Subject: Re: Fixed point sqrt anyone?
- Message-ID: <1993Jan6.093937.7848@canon.co.uk>
- Date: 6 Jan 93 09:39:37 GMT
- References: <1idsq6INNced@flop.ENGR.ORST.EDU>
- Reply-To: ads@canon.co.uk
- Organization: Canon Research Centre Europe Ltd
- Lines: 14
-
- murrayk@prism.CS.ORST.EDU (the Dodger) writes:
-
-
- > Well? Does anyone know of a quick and dirty way to get a reasonably accurate
- > square root of a fixed point integer number? My numbers are stored in a 16.16
-
- It depends on what "reasonably accurate" means. I used a Newton-Raphson
- approximation in a 3D pool game I did a while back, and I got enough accuracy
- and quickly enough for my needs..
-
- Adam
-
- NB If you're looking to do vector normalization then you probably need quite a
- lot of accuracy.
-