Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!usc!polyslo!rcfische
From: rcfische@polyslo.CalPoly.EDU (Raymond C. Fischer)
Newsgroups: comp.sys.mac
Subject: Re: 68882 vs. '881
Message-ID: <25da0cdd.7103@polyslo.CalPoly.EDU>
Date: 15 Feb 90 01:58:53 GMT
References: <651@runxtsa.runx.oz.au> <1514.25d95a24@csc.anu.oz> <1990Feb14.050221.16897@wam.umd.edu>
Reply-To: rcfische@polyslo.CalPoly.EDU (Raymond C. Fischer)
Distribution: usa
Organization: Cal Poly State University -- San Luis Obispo
Lines: 25


In article <1990Feb14.050221.16897@wam.umd.edu> ddev@wam.umd.edu (Don DeVoe) writes:
>    Can anyone tell me if the 68882 and '881 are pin compatible?  I recently
>obtained an '882, and I hope to replace my mac II's '881 with it.  What I
>need to know is A) will this switch work?  B) Will I see any real performance
>improvement?  I kinda-sorta recall hearing that the '882 will work in a
>mac II, but I don't really feel like risking my motherboard on it.  

The 68881 IS pin compatible with the 68882.  Although I haven't tried it,
I referred to the "MC68881/882 Floating-Point Coprocessor User's Manual"
to make sure.

Will you see any performance inprovement?  Yes.  How much?  It depends.
The '882 is faster, but it's speedup comes from increased efficiency rather
than a complete redesign.  The '882 improves the coprocessor interface
to increase throughput and allows multiple '882 instructions to execute
concurrently.  With code optimized for the '882, the speedup can be 50%.
No compilers that I know of do this optimization however, so your actual
speedup will probably be more along the lines of 10%.

Ray Fischer
rcfische@polyslo.calpoly.edu
>    
>    If anyone out there has successfuly swaped these chips in a macII, please
>let me know!