[Nouveau] Nouveau support for GeForce GT 730 or GTX 750 Ti?

[Ilia Mirkin]

On Sun, Jan 3, 2016 at 1:52 PM, Csányi Pál <csanyipal at gmail.com> wrote:
> 2016-01-03 19:32 GMT+01:00 Ilia Mirkin <imirkin at alum.mit.edu>:
> I can't know without buying it whether is it Fermi or Kepler, right?
> I get only these informations from that card:
> VGA GIGABYTE NVIDIA GEFORCE GT730, GV-N730D5-2GI, 2GB DDR5,
> 902/5000MHz, HDMI, DVI-D, D-sub
>
> Is this a Fermi or a Kepler card?

Moderately sure this is a GK208 (Kepler2), you can tell because it's
x8 PCIe, not x16, and 28nm process, while Fermi's (and other Keplers)
were higher, and it supports 4k screens, which I don't think Fermi
did. It also matches the last column of

http://www.geforce.com/hardware/desktop-gpus/geforce-gt-730/specifications

which, while not authoritative, does tend to indicate it's a Kepler.
(While if it matched the first column, it'd be a Fermi - note the
higher wattage and lower core count.)

> > For the best open-source support, I recommend sticking with an
> > on-board Intel GPU, or if that's insufficient (esp recent stuff has
> > been getting pretty speedy), getting something from AMD.
>
> I don't have on-board Intel GPU on this motherboard.
>
> So there is an AMD card too:
> VGA ASUS AMD Radeon R7 240, R7240-2GD3-L, 2GB DDR3, 128bit,
> 730/1800MHz, HDMI, DVI-D, D-sub
>
> What would be the best choice here:
> GEFORCE GT 730
> or
> AMD Radeon R7 240 ?

I don't want to recommend one or the other, esp as I'm unaware of the
specifics around exact AMD GPU support. But in general, if things
don't Just Work (tm), AMD has full-time employees (plural) working on
various aspects of the driver, and those employees are armed with
documentation and access to hardware engineers. NVIDIA provides next
to no support for the nouveau team outside of what's applicable to
Tegra (embedded) devices, there's currently one employee at RedHat
who's the maintainer, and another who is currently looking at making
OpenCL a reality on nouveau. And I contribute what I can in my spare
time, along with a few other volunteers. [NVIDIA does, on occasion,
provide extremely helpful bits of info, and we're thankful for them,
but it doesn't compare to having full access to documentation.] If I
were betting on a team, I'd definitely bet on the AMD one :)

OTOH the NVIDIA proprietary driver is what all linux game ports are
developed against, so by using it you're more or less guaranteed to
have something that works. (Not because it conforms to standards -- it
doesn't -- but because the applications expect precisely its level of
non-compliance.) Your call :)

  -ilia