OpenGL 3.1 Core Support Lands In X.Org Server's GLAMOR

[phoronix]

Phoronix: OpenGL 3.1 Core Support Lands In X.Org Server's GLAMOR

A number of GLAMOR commits landed today within the X.Org Server Git repository...

http://www.phoronix.com/scan.php?page=news_item&px=GLAMOR-GL3.1-Lands

[Azpegath]

So as I understand it, GLAMOR will now use OpenGL3.1 for rendering everything instead of using the old fixed function pipeline? This should possible speed it up quite a bit, right? Improve GPU utilization and decrease CPU overhead?

[andrei_me]

What were the previously required OpenGL version? 2.1?

[agd5f]

What were the previously required OpenGL version? 2.1?

It didn't require any specific version before, but the features and extensions it used generally required OpenGL 3.x capable hardware anyway.

[smitty3268]

So as I understand it, GLAMOR will now use OpenGL3.1 for rendering everything instead of using the old fixed function pipeline? This should possible speed it up quite a bit, right? Improve GPU utilization and decrease CPU overhead?

I don't think it's really changing much, just initializing a core context instead of a compat one. Looks like the main changes were to stop using quads, use vao and vbos in a couple new places, and to use GL_RED instead of GL_ALPHA.

[Azpegath]

I don't think it's really changing much, just initializing a core context instead of a compat one. Looks like the main changes were to stop using quads, use vao and vbos in a couple new places, and to use GL_RED instead of GL_ALPHA.

Well, using VAO/VBOs could really speed things up if batched correctly, using way less draw calls and using subbuffer-updates, etc. AZDO glamor! =)

[nanonyme]

Well, using VAO/VBOs could really speed things up if batched correctly, using way less draw calls and using subbuffer-updates, etc. AZDO glamor! =)

Then you look away for a moment and someone builds prototype vamor just because they want rid of GL completely and move to Vulkan (which winds up a dead project as negligible benefits)

[Azpegath]

Then you look away for a moment and someone builds prototype vamor just because they want rid of GL completely and move to Vulkan (which winds up a dead project as negligible benefits)

"vamor"? And do you mean that Vulkan is a dead project? You sir, I dub Sir Troll of Trolls.

[smitty3268]

Well, using VAO/VBOs could really speed things up if batched correctly, using way less draw calls and using subbuffer-updates, etc. AZDO glamor! =)

I think most of the core bits already were, and this just cleaned up some leftover bits that weren't real performance sensitive anyway. But I'm no expert on glamor or GL, so maybe it'll make a difference.