On Tue, 20 Sep 2011 00:17:37 -0700 (PDT)
Michal Krol <mic...@vmware.com> wrote:
> ----- Original Message -----
> > On Sun, 11 Sep 2011 13:05:38 +0200, Tolga Dalman
> > <tolga.dal...@googlemail.com> wrote:
> > > Hi Kenneth,
> > >
> > > On Fri, 9 Sep 2011 14:41:45 -0700
> > > Kenneth Graunke <kenn...@whitecape.org> wrote:
> > >
> > > > - if (reg->type == BRW_REGISTER_TYPE_F) {
> > > > - return reg->imm.f == 0.0;
> > > > + if (type == BRW_REGISTER_TYPE_F) {
> > > > + return imm.f == 0.0;
> > > [...]
> > > > - if (reg->type == BRW_REGISTER_TYPE_F) {
> > > > - return reg->imm.f == 1.0;
> > > > + if (type == BRW_REGISTER_TYPE_F) {
> > > > + return imm.f == 1.0;
> > >
> > > Shouldn't this rather be something like
> > > fabs(imm.f - 1.0) < std::numeric_limits<float>::eps() ?
> > >
> > > Nevertheless, I like your patch.
> >
> > I don't think so: (some variable equal to 1.0) * (1.0 + epsilon)
> > should
> > return (1.0 + epsilon), while your change would optimize that to
> > return
> > the variable's value of 1.0, right? I wouldn't want to do that
> > without
> > some clear reason to.
> >
>
> And this is still the case since the optimisation applies for deltas that are
> *less* than epsilon, not *less-or-equal* as you imply.
I wouldn't exactly call it 'optimisation' :) However, I feel a bit uneasy about
using the == operator on float variables
as this could lead to unexpected errors.
Best regards
Tolga Dalman
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/mesa-dev
