AMD's Open-Source Radeon Driver After Four Years

[smitty3268]

Yea, I'm still reading up on MLAA, but from what I've read, MLAA is far more taxing on graphics hardware than MSAA is.. So even if they did implement it on older hardware, I wonder if it would even be usable for anything.

MLAA is visually far superior than MSAA, but there's a lot more processing involved in it. All I really want is just 2XMSAA with some decent framerates.. MLAA will give amazing graphics but will have bad framerates on older hardware, no doubt.

I hope they do bring MSAA to the Gallium3D R300 drivers eventually..

MLAA does require at least r500 hardware, so MSAA would be nice to have for r300/r400.

Once it's integrated into Mesa and the hardware drivers are working, I hope Phoronix can do some benchmarks of MLAA. It's not entirely clear to me right now exactly how taxing it is.

[drag]

Of course AMD is not going to abandon their Windows drivers.

I didn't say they should stop developing Windows driver. I said they should drop development for the proprietary Linux driver.

Catalyst for Linux = 95% cross-platform code shared with the Windows driver + 5% of Linux-specific glue code.

I think if you pay close attention to what you written it may make it a bit more obvious why this approach may be very suboptimal from a Linux standpoint.

Anyways the evidence speaks for itself. The catalyst driver is much older, has a much higher budget and workforce behind it and yet it is still one of the worst graphics drivers available for Linux.

If it wasn't for the OSS driver I wouldn't touch it with a ten foot pole and modern Linux desktops would be horribly broken by default for anybody wanting to use ATI's hardware for any purpose (including professional workstations)

It's not like that have 500 Linux hackers working on the Linux blob. It's mostly guys doing cross-platform code, which ends up in the Windows driver.

Don't care. Doesn't matter.

I don't think that there would be any big difference if they dropped Catalyst development, other than losing 90% of their workstation customers who need OpenGL 3 and blob-only features.

yeah right. How competitive do you think their proprietary driver is against Nvidia's?

(hint: They are not doing OSS driver support just to make Linux nerds happy.)

[curaga]

Jimenez' MLAA is actually very good. It's ~12x faster than MSAA 8x (upstream measurement w/ DX; may not be equivalent to the Mesa implementation), better quality than Sony's PS3 one, and better quality & faster than AMD's.

I don't know how it compares in speed with MSAA 2x though.


The minimum hw is r500 / gf6, and it runs quite well on my E-350.

[BlackStar]

Yea, I'm still reading up on MLAA, but from what I've read, MLAA is far more taxing on graphics hardware than MSAA is.. So even if they did implement it on older hardware, I wonder if it would even be usable for anything.

MLAA is visually far superior than MSAA, but there's a lot more processing involved in it. All I really want is just 2XMSAA with some decent framerates.. MLAA will give amazing graphics but will have bad framerates on older hardware, no doubt.

Not necessarily and not necessarily: MLAA is a post-processing effect that doesn't have access to subpixel information, which leads to worse image quality on thin lines. On the other hand, it filters the whole scene, not just polygon edges, which can lead to better quality on scenes with very detailed textures.

Regarding performance, GPUs have dedicated hardware for MSAA which can be quite fast, indeed. However, this hardware has two drawbacks: (a) it's very limited and doesn't cooperate well with modern 3d engines; (b) performance is dependent on the number of polygons in the scene (the more polygons the lower the performance).

MLAA, on the other hand, can be used in every scene and has constant overhead that depends only on the resolution used. Most Xbox360 and PS3 games use variants of MLAA, so we know for a fact that this filter can run on modest hardware at 720p. Intel has also published papers that show the effect running on the CPU in real-time (i.e. render the scene on the GPU, download to the CPU for filtering and re-upload for display).

Without looking at the code or any benchmarks, my guess is that MLAA should be usable on most mid/high-range SM3.0 card and maybe a few high-end SM2.0b cards, provided enough memory bandwidth.

[Sidicas]

The open-source drivers recently gained support for MLAA, an antialiasing post-processing filter. This is similar to techniques utilised in modern games which do not support MSAA (for various technical reasons that go beyond this discussion).

Yea, I'm still reading up on MLAA, but from what I've read, MLAA is far more taxing on graphics hardware than MSAA is.. So even if they did implement it on older hardware, I wonder if it would even be usable for anything.

MLAA is visually far superior than MSAA, but there's a lot more processing involved in it. All I really want is just 2XMSAA with some decent framerates.. MLAA will give amazing graphics but will have bad framerates on older hardware, no doubt.

I hope they do bring MSAA to the Gallium3D R300 drivers eventually..

[BlackStar]

So everywhere I look, it's all about trading off features (losing MSAA) to get other features (open source drivers don't have screen refresh problems)....

The open-source drivers recently gained support for MLAA, an antialiasing post-processing filter. This is similar to techniques utilised in modern games which do not support MSAA (for various technical reasons that go beyond this discussion).

In short, the open-source drivers have now reached feature parity with 9.3, are more stable, offer faster 2d, slightly slower but more featureful 3d (closer to GL 3.0 vs GL 2.1 in fglrx 9.3). Work is also underway for video acceleration. Soon, there'll be little reason to not use the open-source stack on older hardware.

[Sidicas]

WOW!! Great article!!!

Yea, your results are pretty much exactly in line with mine.. I'm pretty much forced to use the older Catalyst 9.3 drivers if I want to run 2xMSAA on games smoothly..

I really wish AMD didn't discontinue the 9.3 drivers because it locks me into older versions of Xorg.... nvidia still offers proprietary driver support for their 7000 series GeForce cards in their latest Linux proprietary driver, so people can run new versions of Xorg with old hardware and still get all their graphic chip's features.

I certainly DON'T think that everybody should move over to Gallium3D drivers on older hardware as for many people, that'd be a loss of features (no MSAA, no Hyper-Z, less 3D performance, etc.)... User's shouldn't have to sacrifice features...

I've only been running Catalyst 9.3 for a week and I've found some annoying bugs in it where it doesn't refresh certain parts of the screen properly in some 2D applications like Firefox 5&6, but manually calling xrefresh takes care of it.. Strangely, starting up Compiz-Fusion, the refresh problem goes away... but then the desktop just isn't as smooth (especially scrolling in Firefox 6 not as smooth with Compiz-Fusion enabled). And also videos are choppy with compiz-fusion enabled.. More weird than that, 3D games run just as smooth with Compiz-Fusion enabled as it does with it disabled... So it's only 2D apps that suffer badly from Compiz-Fusion with the Catalyst 9.3 drivers.

So everywhere I look, it's all about trading off features (losing MSAA) to get other features (open source drivers don't have screen refresh problems)....

At least I have choices which is a good thing, but right now I've got xrefresh bound to a hotkey on the laptop and I'm still sticking to the Catalyst 9.3 drivers for my MSAA and the 3D performance..

[disi]

The cards tested came out like 2007? I got an HD4770 ~2009 after I was pissed off with Nvidia binary for 7950 GT. Then I got Nvidia (460M) again and was again not happy after testing and got myself a new NI AMD laptop card. That would mean my new card is ~2015 fully supported?

With the kernel firmware blobs the AMD 6970M runs fine, power management is fine, desktop is not sluggish, videos play perfect. I am not sure what other people play all the time :/ I played cogs a lot and it looks perfect. OilRush does only 7-10 fps@1920x1080 (on low power profile) and there are parts of the objects not rendered (black), but this is to be expected if there is only OpenGL 2.1 support.

As soon as they put something like Gentoo has ACCEPT_LICENSE="*" into the Mesa Makefile, I will be really happy

[Kano]

Gnome Shell does not matter, only Unity does as U uses it. Didnt you notice that only U releases are tested? Maybe ask U if they want to switch

[pingufunkybeat]

If they dropped LINUX Catalyst development.

Editing broken.