On Thu, 25 Mar 1999, Brian Paul wrote:
> > Since there are rumours of a couple of the IHV's getting geared up
> > to putting geometry hardware onto low end cards (in the next year or
> > so), it may be time to anticipate what that would do to the
> > Mesa core. It's kindof a similar problem to the 3DNow!/PIII issue.
>
> It's definately something to keep in mind. But until the hardware
> is available we've got plenty of other things to accomplish so I'm
> not going to worry too much about it right now.
Oh - sure. I just think that this is something we should all bear in
mind when considering what to change in the geometry processing side
of things.
> Ideally when the hardware does transformation most of the API entry
> points will practically map right onto the hardware.
Yes - we could hope for that.
> What could make
> it a mess is if the hardware only implements _some_ of the OpenGL
> transformation features. Then a mix of hardware and software paths
> will be needed. Yuck.
I fear that's likely. I could easily imagine them leaving off
the ability to read back vertices (for example) - so things like
feedback mode might have to be a software fallback...but that's
just speculation.
Certainly you could imagine that if anything in the RENDERING
path needed a software fallback then that would imply a need for
the corresponding GEOMETRY operations to fall back on software
too....'Yuck' is probably an understatement if that is the case.
Steve Baker (817)619-2657 (Vox/Vox-Mail)
Raytheon Systems Inc. (817)619-2466 (Fax)
Work: [EMAIL PROTECTED] http://www.hti.com
Home: [EMAIL PROTECTED] http://web2.airmail.net/sjbaker1
