Multiple OpenGL 4.x Adjustments Hits Mainline Mesa

[phoronix]

Phoronix: Multiple OpenGL 4.x Adjustments Hits Mainline Mesa

A slew of patches hit mainline Mesa over the night that take care of various OpenGL 4.x related work items...

http://www.phoronix.com/scan.php?page=news_item&px=OpenGL4-24-May-Mesa

[Azpegath]

It really looks like 4.4 might be within reach since only ONE thing is missing for it. And that is in progress.

[SpyroRyder]

Since we reached a new Opengl version isn't it mesa 12? Or were they waiting for 4.4?

[dragorth]

Is it just me, or since Intel has support for OpenGL ES 3.1, shouldn't that be enabled? Also, according to Mesamatrix, 90% of the work required for 4.5 robustness is implemented in the 4.3 robustness extension. It looks like Intel could seriously go for 4.5 in an extremely short time.

(Also, are the extensions released last year for 4.5 on this list?)

[jhenke]

Since we reached a new Opengl version isn't it mesa 12? Or were they waiting for 4.4?

I guess that will happen at the time of the branching. So the release manager will branch the next release with 12.0 and update the version in both the new release branch and master afterwards.

[boffo]

Is it just me, or since Intel has support for OpenGL ES 3.1, shouldn't that be enabled? Also, according to Mesamatrix, 90% of the work required for 4.5 robustness is implemented in the 4.3 robustness extension. It looks like Intel could seriously go for 4.5 in an extremely short time.

(Also, are the extensions released last year for 4.5 on this list?)

Indeed only two extensions + testing & testing.
What are the devs going to do after finishing 4.5?

[Fegr]

Is it just me, or since Intel has support for OpenGL ES 3.1, shouldn't that be enabled? Also, according to Mesamatrix, 90% of the work required for 4.5 robustness is implemented in the 4.3 robustness extension. It looks like Intel could seriously go for 4.5 in an extremely short time.

It should be enabled!
And they deleted KHR_robust_buffer_access_behavior from the GL3.txt that only Intel had. It was never a core extension.

(Also, are the extensions released last year for 4.5 on this list?)

Those extensions are not part of OpenGL 4.5. It called OpenGL 2015, but Khronos will surely soon release OpenGL 4.6 that contain those parts.

[FireBurn]

Indeed only two extensions + testing & testing.
What are the devs going to do after finishing 4.5?

Hopefully look at improving performance and making sure that new hardware is supported well in advance

[Azpegath]

Indeed only two extensions + testing & testing.
What are the devs going to do after finishing 4.5?

I think they have truckloads of performance improvements to take care of! "On disk shader cache" is also one of the things they need to fix.

[V10lator]

Is it just me, or since Intel has support for OpenGL ES 3.1, shouldn't that be enabled?

You're right, Intel has everything in place, too, but the code which finally enables the extension sits in a subfolder called "state_tracker". While I'm not sure I guess that means the gallium state tracker, so Intel can't benefit from it.
Anyway, enabling the same extension for the Intel driver should be a simple task, so let's have a look at r600g instead (according to mesamatrix GL_ARB_compute_shader, GL_ARB_shader_atomic_counters, GL_ARB_shader_image_load_store, GL_ARB_shader_image_size and GL_ARB_shdaer_storage_buffer_object are missing - does somebody know if these are worked on?).