www.delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp/1994/11/13/20:35:44

Date: Mon, 14 Nov 1994 10:37:05 +1100
From: Bill Metzenthen <BILLM AT vaxc DOT cc DOT monash DOT edu DOT au>
Subject: Re: DJGPP & MATH
To: UCKO AT VAX1 DOT ROCKHURST DOT EDU
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu

Aaron Ucko (ucko AT vax1 DOT rockhurst DOT edu) wrote:

> I'd be about the last person to criticize GCC.  What I am criticizing is
> DJGPP's (mis-)initialization of the FPU (which was mentioned in different
> guise a little while ago)

I must have missed that discussion (I skip most djgpp messages unless
the subject line indicates something which might interest me).  Since
you seem to be happy with gcc I guess that you are unhappy with the
default action on the detection of an FPU error.  This is probably a
"religious" issue.  I prefer to have the default of having the FPU
cause an interrupt when an error is detected.


[stuff deleted]
> >have seen is `paranoia' (available on netlib).  It doesn't crash, but
> um...where's netlib?

netlib is a major source of numeric code.  It is mirrored at a number
of sites.  The main ftp site is (or was) "netlib.att.com", directory
"/netlib".  You should find a pointer there to other, perhaps more
convenient, ftp sites.


[stuff deleted]
> >(*) You may find that it reports more flaws or errors if you use a
> >floating point emulator other than wmemu1xx.
>
> We aren't talking about _emulators_ here (although if they're truly good

Perhaps not, but if the emulator isn't good enough it just compounds
the "problems".

Cheers,

--Bill

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019