[Mesa-dev] [PATCH] r600g: Add support for ROUND
cand at gmx.com
Mon Aug 8 00:30:57 PDT 2011
> > On 8 August 2011 03:58, Jose Fonseca <jfonseca at vmware.com> wrote:
> > > It's subjective. It depends on the expected input distribution,
> > > which is effectively impossible to characterize in general. One
> > > can easily find datasets where one method gives biased results and
> > > the other not, and vice versa. And if one takes all possible
> > > numbers, they are equally good.
> > >
> > This is probably largely irrelevant to the patch in question, but
> > just
> > for arguments sake, I don't think that's true. The function floor(x +
> > .5) will introduce positive bias regardless of input distribution,
> I thought the patch implemented round to +/- infinity. Indeed round to +infinitity indeed has a bias.
IMHO, positive numbers are more common in shaders than negative, and for those, the floor(x+0.5) approach would create correct results (= human rounding, away from zero) while rndne would not.
Will rewrite to use rndne though if that's the consensus.
More information about the mesa-dev