Fixed point matrix representation considered harmful

Joerg Sonnenberger joerg at britannica.bec.de
Wed Mar 19 08:59:29 PDT 2008


On Wed, Mar 19, 2008 at 08:25:39AM -0700, Keith Packard wrote:
> > What guaranties do you want to put on precision for the
> > *implementation*? Is the server free to use any floating point data type
> > it wants to use?
> 
> I think we would let it even continue to use the current fixed point
> representation internally. Render doesn't offer precise pixelization
> guarantees anymore...

OK, so you basically want to specify a 32bit (or whatever) IEEE format
without all the nasty parts? E.g. subnormals and exception values?

Joerg



More information about the xorg mailing list