Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!mnetor!seismo!rutgers!mit-eddie!genrad!decvax!ucbvax!ANDREW.CMU.EDU!rs4u# From: rs4u#@ANDREW.CMU.EDU (Richard Siegel) Newsgroups: comp.sys.mac Subject: Re: Floating point Fortran Message-ID: Date: Mon, 8-Dec-86 20:51:08 EST Article-I.D.: andrew.MS.V3.18.rs4u.80021112.easton.ibm032.5266.7 Posted: Mon Dec 8 20:51:08 1986 Date-Received: Tue, 9-Dec-86 05:02:46 EST Sender: daemon@ucbvax.BERKELEY.EDU Organization: University of California at Berkeley Lines: 30 [Line-Eater? What Line-Eater? *Chomp* 8-) ] Organization: Carnegie-Mellon University, Pittsburgh, Pa. Position: Confused Undergraduate To get MS/Absoft Fortran to use the 68881, you need to purchase from {Absoft, Microsoft} a version of the compiler that includes a library that does the 68881 calls instead of the Absoft floating-point package. I believe you can purchase that library separately, or as part of MacFortran/020, which is a 68020 Fortran compiler. Sorry, I can't help as to the price... By the way, MacFortran's floating-point performance on a regular Macintosh is simply INCREDIBLE; it's on the order of 10 times faster than SANE at {x, /, +, -}; I'm not sure of the improvement for transcendentals. It's a shame that those routines can't be used by other programs, perhaps in place of the regular SANE package... Good luck! --Rich Richard M. Siegel Arpanet: rs4u@andrew.cmu.edu (the only way to get to me!) Disclaimer --> Disclaimers are bogus.