Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!bloom-beacon!apple!oliveb!amiga!cbmvax!daveh
From: daveh@cbmvax.UUCP (Dave Haynie)
Newsgroups: comp.sys.amiga.tech
Subject: Re: New stuff
Message-ID: <6440@cbmvax.UUCP>
Date: 29 Mar 89 18:04:57 GMT
References: <3700@amiga.UUCP>
Organization: Commodore Technology, West Chester, PA
Lines: 40

in article <3700@amiga.UUCP>, jimm@amiga.UUCP (Jim Mackraz) says:
> Keywords: 680x0's and computers in general

> )2) Math chips
> )	Two things ARE clear though: 1) it's FAST, and 2) we'd
> )	better start making noises to support this thing compiler
> )	/library wise so it's ready when the chip is available.

> The ieee libraries should be able to make use of any reasonable chip, even
> if it doesn't drop in as a coprocessor. 

This is true.  However, the ieee libraries may slow the thing way down, to
the point where you'd be better off running a 68881 or 68882.  For instance,
some of these quick math chips do adds in 3 clocks, multiplies in 5, at
any precision (register to register of course).  It's likely that a well
written IEEE library for such a beasty could do more complex math operations
perhaps faster than a Motorola FPU, but simple add and multiply instructions
could lose big going via library.  Two things could help this situation 

(1) Compilers that know about killer math chips.  Obviously, we aren't likely
    to see the likes of this at least until there's a standard killer math
    chip.  In 80386-system land, a version of the Weitek Abacus killer math
    chip is rapidly becoming a secondary standard; perhaps such a standard 
    in Amiga-system land would help this along.

(2) Math libraries with higher level functions.  If we had access to math 
    libraries that support real high level functions, like matrix math,
    the price of a call to such a library function might get so small as
    compared to the length of the function, that we'd really get to see
    a performance boost closely tracking the FPU in place.  This could even
    support vectorizing FPUs or multiple FPUs acting in concert.


> Jim Mackraz, I and I Computing	   	"Like you said when we crawled down
> {cbmvax,well,oliveb}!amiga!jimm          from the trees: We're in transition."
> 							- Gang of Four
-- 
Dave Haynie  "The 32 Bit Guy"     Commodore-Amiga  "The Crew That Never Rests"
   {uunet|pyramid|rutgers}!cbmvax!daveh      PLINK: D-DAVE H     BIX: hazy
              Amiga -- It's not just a job, it's an obsession