On Tue, 9 Jul 2002 @ 5:17pm (-0500), Richard Rauch wrote:

RR>   http://mail-index.netbsd.org/current-users/2002/06/17/0000.html
RR>
RR> (This is a post to the current-users mailing list from back in June.)

...how MI is this code?  I.e., will it work on netbsd/sparc64 w/ the
onboard ATI rage pro?

all the XFree86 world seems to be stuck on linux/i386

-lava

RR>
RR> DRI requires XFee86 4.x, which is now going to be the default in
RR> NetBSD/i386 1.6, I gather.  (I'm a bit mixed on the change to XFree86 4.x.
RR> I have a nice, working 3.x config, and was all set to postpone switching
RR> to 4.x until DRI support was there.  4.x offers nothing new that I care
RR> about, and will require me to change configs.  I may just keep the 3.3.6 X
RR> server that ships with NetBSD 1.5.2.  (^&)
RR>
RR>
RR> At this time, accelerated 3D under NetBSD is only possible via the Utah
RR> package (a Mesa derivative that works with XFree86 3.x, but *not* XFree86
RR> 4.x).  I never got anything useful out of the Utah package.  (My video
RR> card wasn't directly supported by the version we had in pkgsrc.  I didn't
RR> have time to seperately try to port a more current version.  The "genric"
RR> support (via SVGA) was supposed to offer some improvements even so (by
RR> imposing fewer software layers and buffer-copies, I guess), as I recall;
RR> but it offered little or no improvement for me and resulted in *terrible*
RR> degradation of quality.  (When using the SVGA X server, I had to cut my
RR> display resolution and the shading was done with coarse
RR> dithering/stippling rather than smooth color selection as I recall. There
RR> was something else (the depth buffer?) that was vastly degraded.)
RR>
RR> Some people reported that the Utah package worked well for them, though.
RR> It just didn't work well for my (S3Virge) card.
RR>
RR> I think that the Utah project is dying (or has died) slowly.  It was a
RR> stop-gap measure before XFree86 4.0 was released.  When XFree86 4.x was
RR> released and GNU/LINUX kernels had DRI support, the raison d'etre for Utah
RR> largely evaporated.
RR>
RR>
RR> The unfortunate reality for as far as I can tell is: If you want hardware
RR> acceleration and if Utah doesn't cut it for you, you need to use a
RR> different OS.
RR>
RR>
RR>   ``I probably don't know what I'm talking about.'' --rauch@math.rice.edu
RR>

later -           | _BonaFide[] = { coder, author, | /~\ The ASCII Ribbon
Brian A. Seklecki | problem solver, scholar, BOFH, | \ / Campaign Against
                  | vegetarian, runner, NetBSD     |  X  Exchange, Outlook
                  | advocate, spiritual machine }; | / \ & HTML Email

PGP: 0111 4618 0111 4618 4AB2 24EB C853 6611 75C4 A91E A7A7 4973

"GNU/Linux: About as stable as the elements at the bottom of the periodic table"