AMD's UVD2-based XvBA Finally Does Something On Linux

[gsacks]

This is good news. But as an end user, I really will only care when XvBA/VA-API finds its way into XBMC. Then maybe my 780G board can go back to being an HTPC board. Has 780G support been confirmed?

Edit: I did see Bridgman post that 780 support was there. Not sure if this is in the official announcement or not though.

[deanjo]

Only AMD/ATI would decide to support a stagnant api like VA-API. What a loss for ATI users.

[smitty3268]

VAAPI isn't necessarily a bad choice here - in fact it means we now have 1 api that covers all major hardware since there is also a VDPAU backend for it. Software support for it should come relatively quickly, i would think, assuming that it works of course. The comment about the 9.11 drivers possibly breaking this is a little worrying, however. Is that going to be a continuing problem with AMD driver updates, or is that just an issue with the early implementation? It's going to suck if everybody's drivers break each month after upgrading.

[Kano]

@evanjfraser

I did not test this beta driver as i refuse to get a login. I provided a hack to get rid of unsupported logo however.

[monraaf]

So I installed Catalyst 9.10, libva and xvba-video, build mplayer with the vaapi support, but when I try to run it:

$./mplayer -vo vaapi -va vaapi ~/Planet_Earth_From\ Pole_to_Pole_1080p_sample.mkv

[vo_vaapi] vaPutSurface(): the requested function is not implemented
[vo_vaapi] vaPutSurface(): the requested function is not implemented
...

MPlayer interrupted by signal 6 in module: decode video
- MPlayer crashed. This shouldn't happen.

This is on a RS780. Has anyone actually got it to work?

[Nille]

R600 doesn't have UVD. It was a pretty big chip already and it had enough shader power and bandwidth to do a fair amount of decode acceleration on the shaders.

But the HD3XXX Generation has UVD and as an big fan of ATI i'm feeling betrayed

[zoomblab]

He who controls the API, controls the univ.. market sorry!

AMD is stupid to go the VA-API (Intel) way. Or not! They want to pull the rub under VDPAU (Nvidia) but they know they can't do that alone. The enemy of my enemy is my friend?

[Kano]

@monraaf

You reminded me of updateing the infos at the end of the script. The way you tried works with nvidia and intel (there you need -2 script) but not with fglrx. fglrx only works with -vo vaapi:gl[:reflect]. Don't know when this will change.

2nd hint: if it still does not play check mediainfo if it is h264 L5.1.

[deanjo]

VAAPI isn't necessarily a bad choice here - in fact it means we now have 1 api that covers all major hardware since there is also a VDPAU backend for it. Software support for it should come relatively quickly, i would think, assuming that it works of course. The comment about the 9.11 drivers possibly breaking this is a little worrying, however. Is that going to be a continuing problem with AMD driver updates, or is that just an issue with the early implementation? It's going to suck if everybody's drivers break each month after upgrading.

We all know how well front end and backend efforts like this have played out in the past. *cough* pulseaudio *cough*

Using a bandaid solution is just bad project planning and is probably linux's greatest flaw.

[n0nsense]

I just wonder, how long it will take to get this news in usable form.
I mean my Sabayon MC doing a great job on Nvidia 9400 IGP.
On the other hand main machine with HD4850 is unusable for watching HD video under Linux (Gentoo). So the question is when i'm going to see it in portage
Why portage ? Since ~amd64 has all the good staff unlike Ubuntu's official repos