OT: nVidia driver [was: Wish list] -- understanding the GPU market

[Bryan J. Smith <b.j.smith@ieee.org>]

From: Dave Jones <davej at redhat.com>
> Except the GeForce doesn't have any possibility of working 3d
> any time soon, whereas the X300 does have a good chance.

You _do_ know that there is some DRI/UtahGLX for nVidia, correct?

I'm glad to see that some people are reverse engineering ATI R300+
since ATI started withholding the specs, but that just means that
they aren't doing anything more than nVidia.

> It's somewhat different in reality. The nvidia involvement in nv
> is one engineer in his spare time when he feels like it.
> NVIDIA don't actually employ him to do nv development, and theres
> no commitment whatsoever from Nvidia to enhance it further. (I don't
> think he actually does work on their binary 3d driver any more either
> coincidentally).

I'm sure he doesn't.  But then again, ATI isn't exactly helping either.

> As I mentioned in an earlier mail, at best we're going to be one
> generation behind.

I disagree.  The NV40 (GF6800) was supported rather quickly after
release.  I've typically seen nVidia support for specific core/cards
in the 2D MIT within 2-3 months.  It's nice to boot a card that is
not very old on a new distro and have it come up, working, and
the docs say the specific revision is supported, so there are no
quirks.

> Which for modern Xorg based servers is next to useless.
> That code has festered in UtahGLX for years, and no-one seems
> interested in doing anything with it.

I'm sure one of the reasons why people aren't sponsoring DRI
development for nVidia is because they view it as redundant
compared to nVidia's Standardware drivers.

All I'm saying is that there's a "double standard" that is applied
to nVidia versus ATI, and that's what most of us are complaining
about.  It's like watching the same "double standard" applied
to Red Hat versus IBM, which makes me curious.  ;->


--
Bryan J. Smith   mailto:b.j.smith at ieee.org