The State Of Mesa OpenGL GL3/GL4 Updated

[89c51]

Do we know how long it will take (approximately) for radeon drivers to reach feature parity with intel??? 3.x stuff.

[Nille_kungen]

Do we know how long it will take (approximately) for radeon drivers to reach feature parity with intel??? 3.x stuff.

That depends on when intel catch up
This is not about feature parity but rather the current state of OpenGL support.

[Ancurio]

... did Michael just write an entire article about someone reformatting a text file??

[rudregues]

Those f*** geometry shaders and glsl must very hard to implement!

Anyway, nice to see the docs updated and get a general view of the things with Michael's article.

[rudregues]

... did Michael just write an entire article about someone reformatting a text file??

Michael, in the end of that file, there is a link http://dri.freedesktop.org/wiki/MissingFunctionality/
This link is very outdated, you should point this in your article.

[smitty3268]

Do we know how long it will take (approximately) for radeon drivers to reach feature parity with intel??? 3.x stuff.

The only thing missing right now is the hardware driver geometry shader support.

AFAIK, no one is currently working on it.

It probably wouldn't take more than a week or two for somebody familiar with the codebase and hardware to finish it off. If it's a volunteer who doesn't have much experience, well, who knows? It's not an easy task for somebody to just jump into, unfortunately.

[Wilfred]

The only thing missing right now is the hardware driver geometry shader support.

AFAIK, no one is currently working on it.

It probably wouldn't take more than a week or two for somebody familiar with the codebase and hardware to finish it off. If it's a volunteer who doesn't have much experience, well, who knows? It's not an easy task for somebody to just jump into, unfortunately.

But why is nobody working on it, if it's quite doable and needed for OpenGL 3.3 compliance? I do not presume to know about AMD's priorities, but I assume that it's not as easy as you make it out to be or that other stuff has higher priority, wouldn't it?

[smitty3268]

But why is nobody working on it, if it's quite doable and needed for OpenGL 3.3 compliance? I do not presume to know about AMD's priorities, but I assume that it's not as easy as you make it out to be or that other stuff has higher priority, wouldn't it?

AMD seems to be entirely focused on Radeon SI support, and enabling their new hardware. And finishing off DPM. I assume they're hoping someone from the community will finish off geometry shaders for them.

[tarceri]

GL_ARB_arrays_of_arrays (GL 4.3) Indiegogo campaign has been been quite successful so far. Basic target (reached) is to write piglit tests and stretch goal ($657 away) is to actually write the extension. Considering that basic goal was funded in about 24h I'm positive this extension will be ready soon too!

I just wanted to clarify that I have not made any promises as to when I will be able to get the extension finished only that it will get finished, the stretch goal will help a lot but just want to make it clear that it may take a bit longer longer. Hopefully it will be ready for a Mesa release early next year. Just don't want to mislead anyone.

[przemoli]

With Marek working in AMD... no hope for quick 3.3 for r600g.

Good news is that AMD is really, really, really trying hard to make radeonSI best thing ever.


And maybe that if/when Nouveau try to tackle geometry shaders in their code base, r600g will get lots of things for free.
Beauty of cooperation in open source.