[NTLUG:Discuss] FC4 and Dell Inspiron 8600

[Bryan J. Smith]

On Mon, 2005-08-22 at 13:02 -0500, . Daniel wrote:
> I have an ATI Radeon Mobility 9200 with 128MB RAM in my machine... I 
> selected that option over the stock NVidia because I foolishly assumed ATI 
> was better supported under Linux and X.org.

Yep, this is a common _farce_.  Just looking at video drivers ...

ATI didn't develop the "clean-room" DRI drivers, but Precision Insight
was contracted by The Weather Channel (I believe, or NOAA?).  ATI only
released _some_ 3D interfaces for the R1xx series (Radeon 7500-8000),
and some features of the R2xx series (Radeon 8500-9000) have been
adopted.  A few of the "value" RV2xx series (Radeon 9100-9200) have
limited support.

ATI has withheld all 3D specifications as of the R3xx series (Radeon
9500+), and they have always had a lot of proprietary 2D as well.

nVidia actually funds development of the MIT 2D drivers, and is open
with their codecs/video-in/video-out (largely because they are industry-
standard licensed designs).  But yes, they horde their 3D drivers ever
since they almost got sued by Intel and Microsoft in their old NV0x
series (TNT through original GeForce[1]) drivers for XFree86 3.3.x.
There is some DRI/UtahGLX work based on those, but they are pretty
useless beyond the NV10/11 series (GeForce/2/MX).

Now in reality, both ATI and nVidia's "mobile" chipsets don't work well
with the stock MIT 2D drivers in XFree86/Xorg.  That's not so much their
fault, but the tier-1 OEMs who get customized designs.  The general
drivers just don't work.

Heck, I have problems with the lack of signal control on my Intel i855GM
chipset integrated video.  It works great, aged GLX support out-of-the-
box, but I can't change what frequencies the DB15 output sends.  So it
matches the (often non-VESA) standard modes of the 1024x768 panel of my
Toshiba M35X notebook.

> I was mostly wrong as it turns out as I have read more about success with
> S3/S4 modes with NVidia users lately.  Sad.

Now the nVidia _nForce_chipset_ is _very_, _very_ GPL and open.  Don't
confuse the nVidia GeForce graphical processor units (GPUs) and the
nVidia nForce chipsets.  In fact, if you have kernel 2.4.23+ or 2.6.5+,
you basically have 100% nForce chipset support!

The only time nVidia had closed source drivers on the nForce was before
the newer ALSA drivers for audio (old nvaudio driver, now replaced by
GPL i810/ALSA work/codecs), and before nVidia secured the IP rights to
its NIC (old nvnet, now they can legally work directly on the GPL
forcedeth).

Fedora Core 3 x86-64 on my nForce4 worked out-of-the-box, 0 drivers.
>From the AMD PowerNow! support down to the full ATA/SATA support.

Also, with the advent of PCI-Express (PCIe), Intel no longer considers
AGP a "trade secret."  So nVidia has opened up the nForce AGPgart
support.

I know this seems like contradictory because Intel is so open with their
i75x/85x/95x GPUs, but those GPUs are _trailing_ edge and do _not_ use
Intel more advanced IP/designs (let alone Microsoft key ownerships thanx
to SGI and other "we swindled you" deals), including nVidia's kernel
driver for GPU-CPU memory mappings.  Even Intel doesn't share a lot of
its registry/timing of its integrated chipsets, and that's largely
because of yet other IP issues.

The video interconnect and GPU space is an IP landfield, and Intel and
Microsoft are holding a lot of the cards -- especially Microsoft.  I
believe it will be the first place Microsoft strikes, and they were
smart to wrestle control over a lot of OpenGL patents in the late '90s.


-- 
Bryan J. Smith     b.j.smith at ieee.org     http://thebs413.blogspot.com
----------------------------------------------------------------------
The best things in life are NOT free - which is why life is easiest if
you save all the bills until you can share them with the perfect woman