was: Still no kmod for new nvidia- now: so Move On

Joonas Sarajärvi muepsj at gmail.com
Tue Aug 3 17:11:27 UTC 2010

2010/8/3, Roger <arelem at bigpond.com>:
> <snip>
> Given the reality, that users bought computers which Linux supported
> only a few years ago, and in some cases paid extra to get computers
> which ran Linux, it really sends a message to have that hardware become
> unsupported two years later. Thanks guys. Hand MSFT a big bag of FUD
> about "will Linux even run on your computer by the time it's depreciated
> or paid for?" Sadly, for once it's true.
> </snip>
> Why in the world would someone pay more for  computer to run Linux?
Because many people want to run something Linux based, or just
something that is free software.

> For the reasons expounded in
> http://lists.fedoraproject.org/pipermail/devel/2009-November/042327.html
> I personally have no reason to go past Fedora 11 and nvidia for Blender
> I could not get radeon to work well at all so swapped for an nvidia
> GT8600, now years old but works very well.
> However I'm at a loss to understand the reasoning that if Gnome and a
> few other 2D apps work then things are good. A strange notion which
> simply doesn't hold water.
2D-only support is better than no support.

> If 3D was not an absolute necessity  then I'm certain that nvidia and
> the other video card designers would not waste time and money developing
> it and trying to eliminate competition (linux).
If they can sell it, they will build it. There unquestionably is a big
market for graphics processors.

> My simplistic notion #1 --Get 3d working first and 2d apps shouldn't be
> a problem. They're basically a single layer in a 3d environment.
> The next step would be to have other 2D app GUIs  running in the layers
> above or below the visible layer. Then users can switch between layers
> to use a chosen app. Except when using Blender.

My understanding of graphics driver development is quite limited, but
I believe it is many orders of magnitude less work to have the basic
2D stuff work, when compared to writing a complete OpenGL

When there is a working OpenGL available, it often makes sense to use
it for 2D graphics, too. On the other hand, one would first need to
have such an implementation available.

> <snip>
> The current focus is on making graphics work for as many ppl as possible
> first, then 3D is always further down the list, this is just common
> sense.
> </ship>
> Open source will never be all things to all people so I cannot agree
> with the above contention.
> The priority to make graphics work on a very small variety of the most
> widely available fairly modern cards seems to be a way to move forward.
There mostly are only few different graphics processors available. At
least my impression is that all the modern ATI GPUs are very similar
from the perspective of a driver developer. I believe the situation is
quite similar on Nvidia side, too. At least the nouveau driver support
has been quite similar for a very wide range of different Nvidia GPUs.

> The problem is that no cards work well in Open source3D.
> Like with printers and scanners, many of which are crappy in Linux,
> couldn't the open source gurus focus on a very small range (say 2 or 3)
> of well known mid capability video card, get them firing on all 3D
> cylinders then promote those as the Linux / Open Source approved cards.
> It won't suit the 3D games people, high end power users but can help
> toward a solution for we who need to use Blender, and there are, I
> believe, thousands of us.
For the reasons I wrote above, I believe this is actually not a very
sustainable way forward. It seems the existing support we have is
easier to port to new devices than it is to write a complete high
performance OpenGL implementation. On the other hand, spending effort
on supporting a wide range of hardware where possible still allows
many more people take advantage of free software. Work on higher end
features seems also take place, but I guess the progress there is a
lot less visible due to the vastness of effort required.

> <snip>
> Current priorities are:
> 0) you aren't running a binary driver - if so no priority for you.
> a) Can you see stuff on the screen at install/boot?
> b) can you run GNOME desktop in reasonably useful manner? i.e. firefox
> runs okay, no glitches, major slowdowns etc.
> c) can you suspend/resume?
> d) can you run compiz/gnome-shell?
> e) can you run non-Gnome desktops at reasonable speed? (yes we have to
> prioritise gnome over KDE, it sucks but thats life)
> f) does misc 3D application run?
> </snip>
>   Pardon me but it's got to be a joke, can't be serious.
> (a) was priority 30 years ago
> (b,c,d) priority10-15 years ago
> (e)  6-10 years ago
> As for (0) I haven't a clue what that means or how it should affect 3D
> modelling and animation.
> My simplistic notion #2.
> Video card design is well known and pretty standard, so why can't an
> Open source electronics genius assemble a competent 3D video card for
> global Linux 3D power users?  Yep it's only a few hundred thousand
> cards, maybe a couple of million but has anyone researched the
> possibilities?
> If so what was the outcome?
> What are the barriers?
> Could it be used to raise funds for further development?
> Does anyone care?
There exists the Open Graphics Project:
http://wiki.opengraphics.org/tiki-index.php. They are developing a
completely open graphics hardware design. I have not followed them
lately, but at least the site seems still quite active.

Joonas Sarajärvi
muepsj at gmail.com

More information about the users mailing list