Re: [PATCH] 32-bit ppc rounding functions (v2) (again)

[Geoffrey Keating <geoffk at geoffk dot org>]

Steve Munroe <sjmunroe@us.ibm.com> writes:

> libc-alpha-owner@sources.redhat.com wrote on 06/29/2004 05:29:20 PM:
> 
> > Dwayne Grant McConnell <dgm69@us.ibm.com> writes:
> >
> > > On Fri, 11 Jun 2004, Dwayne Grant McConnell wrote:
> > >
> > > > The attached patch implements ceil, floor, rint, round, trunc
> > for double and
> > > > floot in ppc assembly.
> > > >
> > > > Resent with ChangeLog this time.
> > > >
> > > > Resent (again) with patch this time.
> > >
> > > Is there anything else I need to do regarding this patch?
> >
> > Well, I'm no Uli, but as a starting point I would suggest implementing
> > these in C using the fesetround() macro.
> 
> The fesetround() is expensive by comparison to this implementation. This
> code represents the minimal use of the expensive mffs/mtfsf instructions.

In that case, you may wish to use the macros in fenv_libc.h, or add
some of your own if they don't do quite what you want.