Catalyst 10.7 Released

[Pfanne]

Since apparently I can't edit the previous post, I'll just put my edit here.

It looks like this:

http://i27.tinypic.com/of8p5w.jpg

Except only on some of the screen.
Apparently this is described as "tearing" and there is a solution to it.
So... Im'ma try to use that. Reporting back shortly.

i dont think this qualifies as tearing.
tearing is way more subtle and most of the time you dont notice it.
this is something else.
have you tried running the application in a window (not full resolution)?
10.7 reintroduced a bug that causes weird behaviour if opengl applications are running with full resolution.

[Ishayu]

i dont think this qualifies as tearing.
tearing is way more subtle and most of the time you dont notice it.
this is something else.
have you tried running the application in a window (not full resolution)?
10.7 reintroduced a bug that causes weird behaviour if opengl applications are running with full resolution.

Well sure it works - but it's a strategy game so I'd like to edge detection scroll - that's disabled if running windowed.

However, all the text is garbled up. It's like 90% of the letters are just gone.
I don't know if that is a Wine bug but it's not anywhere to be found on winehq.org and several NVIDIA users have shown screenshots that look just fine.

[tmpdir]

sure? maybe my english is just to bad..

to buy nvidia 'as well' do not turn you into an nvidia hater and only nvidia haters are cool you know ?

Every man has his vices... can't be cool 24/7

[mguiddy]

Well sure it works - but it's a strategy game so I'd like to edge detection scroll - that's disabled if running windowed.

However, all the text is garbled up. It's like 90% of the letters are just gone.
I don't know if that is a Wine bug but it's not anywhere to be found on winehq.org and several NVIDIA users have shown screenshots that look just fine.

Text is garbled... 90% of the letters are gone...
I had (have) that issue with Alien Swarm.
And from the bug entry on winehq.org, people with nvidia have the issue as well as other games (GalCiv2).

[XabiX]

Also, VAAPI is not working with radeon hd5000 series video cards:

for example, take the following screen-shot using xbmc with vaapi support:




Related ubuntu packages:
xvba-video v0.7.2-1
libva1 v0.31.1+sds4
fglrx v8.753 (Catalyst 10.7 )
xbmc svn ( revision 32181 )

Hi Dandel,

I am not that skilled with linux so I am struggling compiling vlc/mplayer with VA API without affecting the my Lucid 64 installation.
Am i correct to say that there is nothing to expect for evergreen (HD5xxx) users from VaAPI and XvBA?

What will be your recommendation in order to be able to watch DVD, blue-rays and H264 movies without lags etc....?

FYI I am using the fglrx 10.7 drivers but don't feel really satisfied with the quality of videos.

Thanks in advance for some insights and directions so I know where to go with my HTPC.

BTW do you know when XvBA will be supported on evergreens or when should we have a good vaapi implementation??

Merci
XabiX

[gbeauche]

What will be your recommendation in order to be able to watch DVD, blue-rays and H264 movies without lags etc....?

Use Intel or NVIDIA HW.

BTW do you know when XvBA will be supported on evergreens or when should we have a good vaapi implementation??

I am sorry to tell you, but no fix is planned. BTW, the VA-API driver already exceeds what XvBA can do, since it implements some things with OpenGL that are missing in the XvBA/GL backend, or are terribly bugged or partially implemented. This is far from perfect as everything can't be workarounded and I actually don't care... I am really sorry but I am not a masochist and I am terribly fed up.

ATI should just release the current mess if they don't want to fix anything and actually don't have anybody working on that for, uh, a very long time. The fact that Evergreen support is still broken shows that the original people that worked on XvBA @ ATI are no longer there (former trainees?) and very few people, if any, actually know what to do when new chips are released. I am still amazed by how trivial bugs are still there for a very long time... and people prefer waste time discussing whether the fix is really needed (sure it is!) vs. actually fixing the bug. Typically, a 48H round-trip spent for Q/A + convincing was more than enough to actually fix the bugs in the first place.

[maciejus]

"X-window no longer fails with some ATI card plugged in NV's C51 IGP board" - This is what I've been waiting for since the catalyst 9.9.

Many thanks to ATI/AMD and all forum members who were involved in the discussion about the bug for finally solving the issue.

[barbarbaron]

Use Intel or NVIDIA HW.

hehehe one of the best comments on this forum ever.. It seems amd/ati did its thing again serving a canned horseshit full of bugs pretending to be a driver.. Well ati user folks.. The best thing to do is to change your gfx card (if it is a card) with an nvidia one till the gallium3d matures.. or switch to window$ which your gfx manufacturer urges you to do

[klikklak]

OK, ATI you've got me sold. Next few months I'll do the switch to one of your cards. Now, just release your open stereo-3d simultaneously on linux and windows.

[XabiX]

Use Intel or NVIDIA HW.



I am sorry to tell you, but no fix is planned. BTW, the VA-API driver already exceeds what XvBA can do, since it implements some things with OpenGL that are missing in the XvBA/GL backend, or are terribly bugged or partially implemented. This is far from perfect as everything can't be workarounded and I actually don't care... I am really sorry but I am not a masochist and I am terribly fed up.

ATI should just release the current mess if they don't want to fix anything and actually don't have anybody working on that for, uh, a very long time. The fact that Evergreen support is still broken shows that the original people that worked on XvBA @ ATI are no longer there (former trainees?) and very few people, if any, actually know what to do when new chips are released. I am still amazed by how trivial bugs are still there for a very long time... and people prefer waste time discussing whether the fix is really needed (sure it is!) vs. actually fixing the bug. Typically, a 48H round-trip spent for Q/A + convincing was more than enough to actually fix the bugs in the first place.

Thanks Gbeauche for the info. That's sad to hear but if this is the case then we should either stop buying their stuff for linux or they should start investing in an OS getting more and more attention.

If using intel is better then ATI from a linux support standpoint then I may revert back to my Intel Graphics HD embedded video chip then the extra evergreen video card bought recently.

XabiX