Date: Tue, 27 Jun 2000 12:35:08 -0400 (EDT) From: Daniel Eischen <eischen@vigrid.com> To: Steve Kargl <sgk@troutmask.apl.washington.edu> Cc: freebsd-hackers@FreeBSD.ORG Subject: Re: Default (x86) floating point precision Message-ID: <Pine.SUN.3.91.1000627120043.14641A-100000@pcnet1.pcnet.com> In-Reply-To: <200006271525.IAA59630@troutmask.apl.washington.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 27 Jun 2000, Steve Kargl wrote: > Daniel Eischen wrote: > > > > Oddly, this causes problems with GNAT (Ada is a high level language) > > because it wants/expects 64-bit extended precision. It seems as if > > GNAT for linux-i386 also uses 64-bit extended precision. The only > > other GNAT i386 platform that doesn't use 64-bit precision is NT. > > > > So is the above comment still valid? > > > > Does GNAT use the math library in /usr/lib? I've been testing > our math library against UCBTEST, and there appear to be some > pecularities. I need to dig deeper to understand all the info > produced by UCBTEST. The point of this note is that turning on > 64-bit extended precision in GNAT might be compromised by libm.a. Yeah, I kinda wondered how this would affect libm.a (after I changed GNAT to use extended precision :( ). GNAT does provide an Ada interface to some of the math functions in libm. You can configure GNAT to use either a 64-bit or 53-bit precision interface to the math functions. Actually, you can configure it to use 53-bit precision floating point numbers for everything else, but I based this part of the run-time on the Linux run-time which uses 64-bit precision. The more I think about it, I think I should just teach GNAT to expect 53-bit precision floating points and not change the precision of the FP processor. -- Dan Eischen To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.SUN.3.91.1000627120043.14641A-100000>