Re: SIGFPE and -mieee
On Tuesday 17 June 2003 01:04 pm, Helge Kreutzmann wrote:
> Hello,
>
> On Tue, Jun 17, 2003 at 11:10:01AM -0400, George France wrote:
> > Everything built for alpha should use -mieee. It is not need on most
> > other architrectures, but it is almost a requirement for the alpha
> > architecture.
>
> Sure, since all other (popular) architectures implement the IEEE
> standard.
>
> > This has been a subject of lengthy discussion on several mailing list.
>
> Well, which ones? I can't remember such discussions with the outcome
> you mention on either debian alpha nor redhat-axp.
>
Try the archives of the gcc@gcc.gnu.org mailing list. About May of 2002 by
Richard Henderson containing mieee. Richard is the defacto maintainer for
the toolchain for the Alpha Architecture. IIRC, there is even more if you go
further back.
> One of the reasons for the speed of alpha is the ability to calculate
> fast, eliminating often unnecessary checks. So why should we do away
> this advantage by re-inserting those checks on a broad basis again
> when compiling software?
>
> If a program divides by zero, it is a bug. If it can't handle an
> overflow, it is a bug. If the author *knows* there could be a division
> by zero or simmilar, he/she has to provide checks.
>
> -mieee is for software where speed is absolutly irrelevant and where
> bugs like those mentions are no problems (e.g. an office suite). Also
> it can be a first step to get an application running on alpha at all.
> But IMHO this should be decided on a case by case basis and not be
> stuck into every compilation; especially since several programs are
> fixed now and can work fine without -mieee (or never needed one in the
> first place).
>
> > I hope this helps.
>
> Nope, sorry.
This is not a speed issue but a library issue.
Best Regards,
--George
Reply to: