Path: utzoo!attcan!uunet!super!udel!gatech!cwjcc!mailrus!ames!oliveb!amiga!boing!dale
From: dale@boing.UUCP (Dale Luck)
Newsgroups: comp.sys.amiga
Subject: Re: Math Chip Interface
Message-ID: <596@boing.UUCP>
Date: 18 Jan 89 13:27:48 GMT
References: <6736@louie.udel.EDU>
Reply-To: dale@boing.UUCP (Dale Luck)
Organization: Boing, Milpitas, Ca.
Lines: 23

In article  iphwk%MTSUNIX1.BITNET@cunyvm.cuny.edu (Bill Kinnersley) writes:
>[In "Re: Math Chip Interface", Dale Luck said:]
>It apparently talks to the 68881 by using the "CPU address space".
>It loads 7 into the Function Code registers (SFC and DFC) then sends
>stuff to the 68881 registers using the MOVES instruction.

>Probably means I am out of luck for 1.3 support, doesn't it?

Yup, using the cpu space stuff means the 68881 software would not work
on a 68000 equiped machine. Since there were many more 68000 machines
than 68010 machines we chose to only include support for what we felt
would be the 99% of the cases.

Also the extra time required to use the cpu space by changing the SFC and
DFC would have only added additional overhead to an already overhead
bound feature.

>


-- 
Dale Luck     GfxBase/Boing, Inc.
{uunet!cbmvax|pyramid}!amiga!boing!dale