Return to BSD News archive
Xref: sserve comp.os.linux.misc:557 comp.os.386bsd.misc:741 Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!agate!howland.reston.ans.net!darwin.sura.net!haven.umd.edu!umd5.umd.edu!roissy.umd.edu!mark From: mark@roissy.umd.edu (Mark Sienkiewicz) Newsgroups: comp.os.linux.misc,comp.os.386bsd.misc Subject: Re: Why would I want LINUX? Date: 23 Aug 1993 16:11:45 GMT Organization: University of Maryland Lines: 18 Distribution: world Message-ID: <25aq81$abf@umd5.umd.edu> References: <MIKE.93Aug19115915@pdx800.jf.intel.com> <250m5t$dmk@europa.eng.gtefsd.com> <252n71$2d4@fnnews.fnal.gov> NNTP-Posting-Host: roissy.umd.edu In article <252n71$2d4@fnnews.fnal.gov> dejan@cdfsga.fnal.gov (Dejan Vucinic) writes: > > All this probably holds for Linux as well. It seems that DOS engineers >used some other mathematics in their time calculations. ;> > > Don't trust figures too much. Try and measure. You'll be surprized. When you look at DOS machines, you have to pay close attention to just what the benchmarks measure. Often, you will find the Unix-ish system had to support 32 bit integers and a megabyte of memory, while the DOS system got to get away with 16 bit integers and 64k. Your friend probably did a fair chunk of shuffling things around in large model. It probably also helped that you used GCC as a code generator. You might try running the same fortran through f2c and compiling it for DOS with a good compiler like Turbo C (or whatever they're calling it these days).