OpenGL Overload: Implementing OpenGL Over Vulkan

[sarmad]

Doesn't sound useful to me. OpenGL is already cross platform, drivers are already close to be 4.5 compliant, and OpenGL development seems to be slowing down so we don't expect much future improvement to it anyway. The amount of work to re-implement it seems more than that amount of work needed to fix the shortcoming of the drivers and add whatever new features that might be coming.
A better project to work on is to implement DX11 or DX12 using Vulkan.

[droste]

I'm slightly confused about what this actually does. Is this basically just a compatibility layer so you can, for example, run a Vulkan application using a GPU that only supports OpenGL?

The other way around. Vulkan driver running OpenGL applications.

[FLHerne]

It's obviously intended to be used by more than linux on AMD/intel

So is Mesa, LLVMpipe runs on Windows and is used by VMWare among other things.

This doesn't seem useful for hardware with existing GL drivers, but it'll save work in the future.

[schmidtbag]

The other way around. Vulkan driver running OpenGL applications.

Weird... unless it allows for the performance enhancements of Vulkan, this sounds very unnecessary. No wonder I found it confusing.

I agree with sarmad - an implementation of DX11 or DX12 would be great.

[ruthan]

Well, so great Vulkan feature would be, that it could be downgraded back to OpenGL?

[Nille]

Doesn't sound useful to me. OpenGL is already cross platform,

It is very useful. Yes, OpenGL is Crossplattform but not Crossvendor friendly. You see it every Day with the games on Steam or WINE.

The OpenGL Spec is not clean enough and none of the current OpenGL Implementation run all piglit tests. With OpenGL 4.4 and 4.5 Khronos makes a bit better with the conformance test.

[totoz]

Doesn't sound useful to me. OpenGL is already cross platform, drivers are already close to be 4.5 compliant, and OpenGL development seems to be slowing down so we don't expect much future improvement to it anyway. The amount of work to re-implement it seems more than that amount of work needed to fix the shortcoming of the drivers and add whatever new features that might be coming.
A better project to work on is to implement DX11 or DX12 using Vulkan.

Well, I guess that it could become more relevant in the future, for newer GPUs having only vulkan driver. And OpenGL being considered legacy, it wouldn't be worth developing an OpenGL driver for it. Because having to support OpenGL + Vulkan (+ DirectX + Metal + ...) will start to be quite expensive for driver vendors.

[schmidtbag]

Well, I guess that it could become more relevant in the future, for newer GPUs having only vulkan driver. And OpenGL being considered legacy, it wouldn't be worth developing an OpenGL driver for it. Because having to support OpenGL + Vulkan (+ DirectX + Metal + ...) will start to be quite expensive for driver vendors.

Considering how long it took (or really, is still taking...) DX9 to die, I don't think GPU vendors are going to be letting go of OGL for a while. I don't think OGL support will be dropped for at least 10 years.

[airlied]

someone should tell him to open the issues on his repo, since he says in the README to file issues. I can't find an obvious email address for him, and I don't twit.

[illwieckz]

Sounds nice, but its not easier to port Mesa3D? A Vulkan Backend like Gallium3D .

It's obviously intended to be used by more than linux on AMD/intel

You mean on AMD or nVidia, because as far as I know Intel driver does not use Gallium3D, so any Vulkan-based Gallium3D replacement would not work with Intel drivers.