Re: alpha/12623: strtod(3) FPE on alphaev56

[ Available lists | Index of freebsd-alpha | Month of Jul 1999 | Week of 16 Jul 1999 | Raw email | View thread | Wrap long lines | Reply | Tag ]
From
Doug Rabson <dfr@nlsystems.com>
Date
16 Jul 1999 01:44:04
Subject
Re: alpha/12623: strtod(3) FPE on alphaev56
Message-ID
Pine.BSF.4.10.9907160944230.58023-100000@salmon.nlsystems.com

In reply to
Referenced by

[ Hide this part ]
On Thu, 15 Jul 1999, John Polstra wrote:

> Doug Rabson wrote:
> > On Tue, 13 Jul 1999, John Polstra wrote:
> >
> >> Yes, or the equivalent "-mieee". In my opinion, all of the libraries
> >> (if not the whole world) should be compiled that way. In fact, this
> >> option should be the default. Users who wanted a little extra speed
> >> and who knew what they are doing could turn it off.
> >
> > I agree (at least as far as the libraries go). Setting it as default would
> > be easy but there are performance implications. On the other hand, the
> > only applications whose performance is affected are ones using floating
> > point..
>
> Yes, I think "correct by default" would be more in keeping with the
> FreeBSD tradition than "fast but possibly wrong by default". So I
> still think -mieee should be used for everything.
>
> My only real concern about making it the compiler default is that it
> causes us to deviate from standard egcs a little bit more. But the
> deviation would be confined to "egcs/config/alpha/freebsd.h", which is
> probably OK. I think we should add a "-mno-ieee" option to disable
> it, too. That's not hard.

I agree. As long as there is a -mno-ieee option for those programs which
need to squeeze the last drop of performance then this is the right thing
to do.

--
Doug Rabson Mail: dfr@nlsystems.com
Nonlinear Systems Ltd. Phone: +44 181 442 9037




To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-alpha" in the body of the message



Elapsed time: 0.134 seconds