Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!ut-emx!ccwf.cc.utexas.edu From: greg@ccwf.cc.utexas.edu (Greg Harp) Newsgroups: comp.sys.amiga.advocacy Subject: Re: Computer Architecture question -- Daye Haynie Message-ID: <48815@ut-emx.uucp> Date: 11 May 91 06:25:34 GMT References: <1991May10.211958.25387@sbcs.sunysb.edu> Sender: news@ut-emx.uucp Reply-To: greg@ccwf.cc.utexas.edu (Greg Harp) Distribution: comp Organization: The University of Texas at Austin Lines: 30 In article melling@cs.psu.edu (Michael D Mellinger) writes: > >In article <1991May10.211958.25387@sbcs.sunysb.edu> dtiberio@eeserv1.ic.sunysb.edu (David Tiberio) writes: > > Does the 68040 have a math coprocessor? I remember hearing that there were > problems with the coprocessor and it was removed. I assume that a 68882 would > be slower than a 68040, but could two math processors also be used together? > >The math coprocessor is built into the 68040. Moto. claims that for >most things it is 5-10 times faster than the 68882. Well, the 040 is missing the trancendental functions that the 882 had. These, of course, could be emulated with traps. I've heard that it was anywhere from 2X to 10X faster than the 882. I'd tend to believe more of the lower speeds under normal use. Maybe if you used the "one function that was super-optimised over the 882" you'd get 10X the speed. I'd be willing to bet that under normal use you get a 3 to 4 times speedup. That is, of course, when using the I and D caches. I have, of course, done no benchmarking of the FPU functions of the 040. This is pure speculation on my part, but hey, I'm allowed that once in a while. ;-) Greg -- Greg Harp |"I was there to match my intellect on national TV, | against a plumber and an architect, both with a PhD." greg@ccwf.cc.utexas.edu| -- "I Lost on Jeopardy," Weird Al Yankovic