Another OpenGL 4.5 Extension Marked As Done In Mesa

[newwen]

Also Dave Airlie updated it for recent codebase changes two days ago:
http://lists.freedesktop.org/archive...ly/089280.html

Very nice! It looks like it's ready. I hope they make it into mainline so we get full OGL 4.0 in the next mesa release.

[tomtomme]

hopefully in the next week, we can merge tess and shader subroutine to master which will give nvc0 4.1, and then radeonsi tess code along with a couple of bits of ARB_gpu_shader5 which I've written but need some llvm changes to work properly. Intel users will have to wait until someone figures out tess and an fp64 strategy. It looks like for fp64 Intel like most r600 gpus will need a softfloat implementation of some sort.

that would be: boooombastic (radeonSI user here)
thanks david for yourupdate here!
thanks to all the foss devs!

[Khudsa]

hopefully in the next week, we can merge tess and shader subroutine to master which will give nvc0 4.1, and then radeonsi tess code along with a couple of bits of ARB_gpu_shader5 which I've written but need some llvm changes to work properly. Intel users will have to wait until someone figures out tess and an fp64 strategy. It looks like for fp64 Intel like most r600 gpus will need a softfloat implementation of some sort.

That's great!!! It's for radeonsi users, I'm a r600 user (6970 card), it will take long time for r600 to reach ogl4 compliance? Thanks!

[bokal]

Thanks for the links. Let's hope it all gets merged before mesa 10.7.

[drSeehas]

... which will give nvc0 4.1, ...

Will this run also on NVE0 (Kepler)?

[My8th]

So only 4 extensions left til 4.2 compliance and they're all in progress. Awesome!!

[Ancurio]

hopefully in the next week, we can merge tess and shader subroutine to master which will give nvc0 4.1, and then radeonsi tess code along with a couple of bits of ARB_gpu_shader5 which I've written but need some llvm changes to work properly. Intel users will have to wait until someone figures out tess and an fp64 strategy. It looks like for fp64 Intel like most r600 gpus will need a softfloat implementation of some sort.

Dave, GL3.txt says "ARB_shader_precision" is still in development, yet the infrastructure in mesa looks already to be done, and since nvc0's pipe reports a GLSL feature level of 410, it seems to already expose that extension. What is up with that?