clt@pur-phy (Carrick Talmadge) (03/08/88)
Last week I sent out a query regarding the A/UX Fortran compiler, which didn't appear to work. Apple has since confirmed that they distributed in their initial release of A/UX a broken f77 compiler. The f77 compiler which we received does not support the 68881 chip. The f77 libraries, however, were compiled assuming the '881 chip. Consequently, any calls to the standard f77 floating point libraries give bogus results. The f77 compiler also does not work with the "-g" flag (you can't even successfully link an executable in this case). Thought you might like to be warned... Carrick Talmadge Disclaimer: I don't use f77 clt@newton.physics.purdue.edu personally, so no hassles, ok?
verber@fruit.cis.ohio-state.edu (Mark Verber) (03/10/88)
If some is doing serious Fortran-77 work with number crunching I would strongly recommend the Greenhill's Fortran compiler. One of our professors was interested in the speed of the MacII for his huge number crunching program. The MacII using the Greenhill's compiler was as just about as fast as his Sun 3/260 with FPA using Sun f77. Note: the Mac was on a prototype MacII (2x the number of memory wait states, running a beta version of A/UX). We all know that the Sun should have creamed the MacII. This either says something about f77 or about Greenhill Fortran (or maybe both). Cheers, ----------------------------------------------------------------------- Computer Science Department Mark A. Verber The Ohio State University verber@tut.cis.ohio-state.edu +1 (614) 292-7344 cbosgd!osu-cis!verber