Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site omsvax.UUCP
Path: utzoo!watmath!clyde!akgua!sdcsvax!sdcrdcf!hplabs!intelca!omsvax!plb
From: plb@omsvax.UUCP (Phil Barrett)
Newsgroups: net.wanted
Subject: math routines
Message-ID: <903@omsvax.UUCP>
Date: Fri, 11-May-84 11:03:38 EDT
Article-I.D.: omsvax.903
Posted: Fri May 11 11:03:38 1984
Date-Received: Sun, 13-May-84 08:28:14 EDT
Organization: Intel Corp, Hillsboro, OR
Lines: 23

<insert witty bug-deriding comment here>

Sorry if this is a well worn saw.

I am looking for a set of math routines (sin, cos, ...) that
use the 8087 transendental support to its fullest advantage.
In particular, I'm interested in 80 bit (temp real) computations.
The current libm routines just do a standard approximation
which isn't particularly accurate when compared to what can
be achieved. I need to be able to call these routines from
C (C source + a little assembler would be perfect, sorry,
just dreaming).  I'm particularly intrested in libm.a compatable
code but will take anything that even remotly resembles it
(er, them?).  Public domain or (*gasp*) commercial are equally
acceptable.

Please reply by mail and (assuming responses) I will post the results

					Phil Barrett
					Intel Corp
					tektronix!ogcvax!omsvax!plb
					hplabs!intelca!omsvax!plb