[geeks] Jesus this guy could be me

[Frank Van Damme]

On Sunday 19 January 2003 08:05, Joshua D Boyd wrote:
> On Sat, Jan 18, 2003 at 11:01:31PM +0100, Frank Van Damme wrote:
> > On Saturday 18 January 2003 22:46, Joshua D Boyd wrote:
> > > Open hardware?  Who the hell makes anything reasonably resembling an
> > > open 3D accelerator?  Certainly not ATI, Nvidia, or Matrox.  Or the
> > > people who make that Kyro card.  Who makes an open video IO card
> > > (well, Linux Media Labs makes a rather mediocre one that they price
> > > gouge you on). About the only open hardware is SCSI and ethernet, and
> > > stuff where the chip maker will sell the chips to anyone so it wasn't
> > > too hard to reverse engineer the rest.
> >
> > Nothing is perfect in this world, but I'm still getting acceleration
> > with fairly recent ati cards (8500 radeon) while no one will be able to
> > the same for an nvidia card. Well not with opensource drivers at least.
> > And Matrox is notorious among linux users for its good opensource
> > support.
>
> I haven't seen much good support for a Matrox card for 3D acceleration
> since the G200.  Perhaps I'm not up to date in this area.
> I actually would love to get a few G200s in PCI configurations, just
> because they are so well documented.

The whole Gxxx series should be equally supported under linux as under 
windows. I have no idea about the parhelia though.

> I haven't heard particularly good things about the open source radeon
> drivers, but I keep hoping that that will change.  At the moment the
> Weather Channel seems to be sponsering development, but I haven't heard
> anything about T&L support.

There's no card which has so many drivers for the same platform as this one 
;-)

There will normally be accelerated drivers in XFree86 4.3. I installed a lot 
of binary snapshots over the last half year or so, the last thing I did was 
th compile the dri sources from cvs, which worked pretty well with quake3. 
Except for the mouse (only the buttons worked, but no movement), but that 
was a quake bug (which is now solved). Also, I'm talking about r200 based 
cards. radeon 9000 and so on are r250 or r300 based cards, who have very 
buggy support. dri just lacks developping time unfortunately... 

> The old RivaTNT drivers seemed fairly decent to me.  But, just
> supporting a rasterizer is a lot easier than supporting a T&L system I
> guess.

Indeed. Actyally I owned a tnt2, which made my system crash line windows 9x. 
And there's nothing you can do about it. And if there's a bug, I think 
nvidia would rather like you to buy a geforce3 then to fix it. That's why I 
find open source drivers important. And availability of hardware specs. 
That said, you have to be an active xfree86 developper and sign an nda to 
develop radeon drivers.

> > Who are Linux Media Labs? :-)
>
> http://www.linuxmedialabs.com/
>
> They make the LML33 which is an MJPEG board that is to my understanding
> somewhat related to the Iomega Buz or the Pinnacle DC10+.  Except they
> charge a lot more than the other two go for.  They now have a LML33R10
> board.  I'm not sure what the difference is other than it is less than
> half the cost, but still not particularly cost competitive compared to
> the others.

Ahha. I thought it was a joke of yours :-)

-- 
Frank Van Damme