Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!utgpu!water!watmath!clyde!rutgers!gatech!mcnc!ecsvax!wmcb From: wmcb@ecsvax.UUCP Newsgroups: comp.sys.mac Subject: Re: MS Fortran Math from Lightspeed Pascal Message-ID: <3046@ecsvax.UUCP> Date: Wed, 6-May-87 11:39:38 EDT Article-I.D.: ecsvax.3046 Posted: Wed May 6 11:39:38 1987 Date-Received: Fri, 8-May-87 01:25:15 EDT References: Reply-To: wmcb@ecsvax.UUCP (William C. Bauldry) Organization: UNC Educational Computing Service Lines: 24 In article rs4u#@ANDREW.CMU.EDU (Richard Siegel) writes: > >I am about to start writing code to make it possible to call >MS Fortran's floating-point math routines from Lightspeed Pascal >(or any other, for thatmatter). I'm doing this because MS Fortran's >math is about a factor of 10 faster than SANE is (on a regular Mac). > true, but what about the accuracy of the routines - does anyone know what tradeoff's were made to get the speed ? I'd want to know that first - especially in the transcendetal routines, etc. Also, can the environment be altered and monitored as in SANE ? I'd really be interested in the answers to these (there are several LSP programs I have the would love to be an order of magnitude faster!!) Any kind soul that would post and/or email the answers would have my eternal (or at least until Friday) gratitude... Bill --------------------------- B. Bauldry wmcb@ecsvax.UUCP or whatever Dept of Math Sci Appalachian State U Boone, NC 28608 ---------------------------