Path: utzoo!attcan!uunet!husc6!mailrus!ames!decwrl!nsc!taux01!yuval
From: yuval@taux01.UUCP (Gideon Yuval)
Newsgroups: comp.sys.nsc.32k
Subject: Re: 1: floating math library; 2: 32CG016.
Message-ID: <798@taux01.UUCP>
Date: 2 Jul 88 18:14:26 GMT
References: <3202@imag.imag.fr>
Reply-To: yuval@taux01.UUCP (Gideon Yuval)
Organization: National Semiconductor (Israel) Ltd.
Lines: 25


 
>   I just realized today that none of the two C compilers we have (standard V.2
> and CTP, both for 32k), feature the "-f" option (compile and link for no FPU).
> A quick look in the libraries showed no floating math functions.
> 
> We are considering to use the 32CG016 in an industrial design without FPU (not a
> CMOS chip ).
> 
> The "undefined" trap could help but need to write the code interpreter/
> emulator.

Two ways out: 
(1) NSC's FPS (Floating Point Simulator) is now a product. It emulates the
32081/381/580 bit-for-bit, & is therefore slower than ...

(2) U.S. Software (Portland, Oregon) have a S/W floating-point library which,
on the 386, beats 287 HARDWARE speed. If they have a 32K version (or if
you convert their 386 code to 532 - their source-code sells for $1250),
you should get a pretty fast S/W FP package. NOT bit-compatible with
081/381/580 results -- you pays your money & you takes your choice.
-- 
Gideon Yuval, yuval@taux01.nsc.com, +972-2-690992 (home) ,-52-522255(work)
 Paper-mail: National Semiconductor, 6 Maskit St., Herzliyah, Israel
             (alternative E-mail address: decwrl!nsc!taux01!yuval@uunet.uu.net)