Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!apple!agate!ucbvax!ucsd!ucrmath!yoshio
From: yoshio@ucrmath.ucr.edu (yoshio nakamura)
Newsgroups: comp.arch
Subject: Re: sin(pi) (Was: Re: bizarre instructions)
Summary: References needed for numerical computation
Keywords: reference numerical computation
Message-ID: <12339@ucrmath.ucr.edu>
Date: 27 Feb 91 19:26:38 GMT
References: <15485@lanl.gov> <3028@charon.cwi.nl> <3030@charon.cwi.nl>
Sender: news@ucrmath.ucr.edu
Followup-To: comp.arch
Distribution: usa
Organization: University of California, Riverside
Lines: 18

In article <3030@charon.cwi.nl> jurjen@cwi.nl (Jurjen NE Bos) writes:
>
>Oops.  I was asleep.
>The correct value of sin(3.14159265359) in radians (not degrees!) is of course
>-2.06761537357E-13.  The calculator uses 32-bit argument reduction, so that in
>can compute 12-digit sine values up to 1E20.
>I am embarrassed.

As long as we are on the subject of numeric computation and accuracy of
math operations, can someone point me to a reference on 32-bit argument
reduction and how to implement routines like cosine to an accuracy of 
arbitrary n-bits?  ( In English, not Dutch  :-)  )

-yoshio

-------------------------------------------------------------------------------
Yoshio Nakamura                      Internet: yoshio@ucrmath.ucr.edu
University of Calif., Riverside      uucp: {ucsd, uci, ucdavis}!ucrmath!yoshio