Reply With QuoteThis ought to make a heck of a long term release in about a year.
Phoronix News Bot
Mesa 7.9 Makes It To The Release Candidate Stage
Phoronix: Mesa 7.9 Makes It To The Release Candidate Stage
Mesa 7.9, one of the biggest updates to this open-source graphics driver stack in quite a while, is almost here. Ian Romanick has made available a release candidate for Mesa 7.9 and is saying the final release will be here one week from now (4 October) barring any major issues cropping up...
http://www.phoronix.com/vr.php?view=ODYzNA
Senior Member
This ought to make a heck of a long term release in about a year.
Senior Member
Will Galium3d be the default or mesa classic?
Senior Member
Depends on driver..Originally Posted by Beiruty
AMD Linux
I think the idea is 300g for 3xx-5xx, 600c for 6xx-EG...
... so "yes and no"
X.Org Developer
There is no "default", it's up to distros what they package. All we can do is to build the driver for them.
If you build both classic and gallium and type "sudo make install", gallium gets chosen basically (it first installs classic and then gallium, overwriting classic).
AMD Linux
There were some useful new fixes for r600c (HD2xxx-HD5xxx) checked into 7.9 over the last day or so, including stencil support for Evergreen (HD5xxx). This would be a great time to try out the latest code before 7.9 locks and ships.
If you are running HD5xxx I believe you need the latest DDX from the evergreen-accel branch as well.
Apparently OpenArena runs properly on Evergreen now
Senior Member
ioquake3 has been working on evergreen for a while, it's been playable since the mipmap fixes. (it was playable before, just ugly). Didn't test openarena, but it's based on the same engine.
I noticed that even simple 3d apps are affected by resolution, even though fill rate shouldn't be a problem there. I guess it's because colortiling is still missing for evergreen?
At one point all gfx froze. I could still move the mouse pointer, but there were no more framebuffer updates, no VT switches, couldn't kill X, nothing. This happened before, but due to a lack of messages in dmesg or xorg.0.log or any reliable reproducability, I wouldn't know how to properly report it.
