[comp.realtime] 68340 Flt Pt ?

jandres@sag4.ssl.berkeley.edu (Jeff Andres) (06/12/91)

This message is being posted in behalf of my wife! :)

She is working on an embedded realtime 68340 project which requires the
use of floating point calculations. Their current design is not to add
a floating point processor into the design due to power and size
limitations. The question is "How can they perform floating point
operations or emulate them?"

I'll take the answers through email (address below). Thanks in advance.

^     _____________	                                                ^ 
^    |             |                  Jeff Andres                       ^
^    |             |              jandres@ssl.berkeley.edu              ^
^    |             |             Center For EUV Astrophysics            ^
^    |_____________|        2150 Kittredge Blvd, Berkeley CA 94720      ^

glenn@ready.eng.ready.com (Glenn Kasten) (06/13/91)

From glenn Wed Jun 12 10:09:25 1991
To: jandres@ssl.berkeley.edu
Subject: Re: 68340 Flt Pt ?
Cc: +683xx
Status: R

According to Tom Starnes, Motorola product marketing manager
for the 68330 and 68340, the first batch of 683xx processors
has no floating-point support.  He believes only 10% of
intended market requires floating-point, and this can be handled by:
a. Motorola's "fast floating-point" software package
(non-IEEE format, 32-bit precision, very fast, no C support
b. C compiler vendor's full IEEE software floating point
c. 68881/68882 as a peripheral chip using F-line traps or macros
These are Tom Starnes opinions, not mine.
---
Glenn Kasten
Ready Systems 470 Potrero Ave. Sunnyvale CA 94086 
glenn@ready.com (408) 522-7357 

-- 
Glenn Kasten
Ready Systems 470 Potrero Ave. Sunnyvale CA 94086 
glenn@ready.com (408) 522-7357 

bill@mwca.UUCP (Bill Sheppard) (06/15/91)

In article <1991Jun12.171040.12991@ready.eng.ready.com> glenn@ready.eng.ready.com (Glenn Kasten) writes:
>According to Tom Starnes, Motorola product marketing manager
>for the 68330 and 68340, the first batch of 683xx processors
>has no floating-point support...

>...this can be handled by:

>...b. C compiler vendor's full IEEE software floating point

Similarly, a real-time OS such as OS-9 can be used with the 68332/340 (we
have a drop-in version tailored specifically for these chips), which will
afford the capability of using our floating point emulation with our
compilers either via libraries or trap handlers (C, PASCAL, FORTRAN) or
with our BASIC interpreter.

We have a free evaluation software kit for use with the 68332/340 EVS and
a PC - please contact microware (800-475-9000) for more information.


-- 
 ##############################################################################
 # Bill Sheppard  --  bills@microware.com  --  {uunet,sun}!mcrware!mwca!bill  #
 # Microware Systems Corporation -- OS-9 / OS-9000 / CD-RTOS -- (408)980-0201 #
 ##### Opinions expressed are my own, though you'd be wise to adopt them! #####

chaplin@keinstr.uucp (chaplin) (06/20/91)

In article <1991Jun12.145027.12204@agate.berkeley.edu> jandres@sag4.ssl.berkeley.edu (Jeff Andres) writes:
>This message is being posted in behalf of my wife! :)
>
>She is working on an embedded realtime 68340 project which requires the
>use of floating point calculations. Their current design is not to add
>a floating point processor into the design due to power and size
>limitations. The question is "How can they perform floating point
>operations or emulate them?"
>
>I'll take the answers through email (address below). Thanks in advance.
>
>^     _____________	                                                ^ 
>^    |             |                  Jeff Andres                       ^
>^    |             |              jandres@ssl.berkeley.edu              ^
>^    |             |             Center For EUV Astrophysics            ^
>^    |_____________|        2150 Kittredge Blvd, Berkeley CA 94720      ^

Try the Motorola Freeware BBS.  (512)891-3733 1200 baud.  There is a file
with this description:

68343ffp.arc Source code for the M68343 Fast Floating Point routines.
.            (Unsupported!) 

I suspect these will work for the 68340 as well, perhaps with minor
modification.  I have not downloaded this file, so I can't comment further.
I just hoped to point you in the general direction.

There are probably many other sources for M68000 floating point libraries.
Chances are that you may find out where they are by asking on the Freeware BBS.
-- 
Roger Chaplin / Instruments Division Engineering / "Though I'd been lost now I
chaplin@keinstr.uucp / CI$: 76307,3506          / felt I was found when He
#include <disclaimer.h>                        / looked at me with His
#include "disclaimer.h" /* cover all bases */ / forgiving eyes." - Michael Card