VK9, the open source project to implement d3d9 over Vulkan continues to improve

Hmh, I don't really get the point in the project except if it is for academic purpose (which is perfectly fine, don't get me wrong there!). Providing a d3d9->vulkan mapping for systems which ship with d3d9? In the end, 2020 we'll rather see Vulkan/DX12 things popping out than having D3D9 or even 11 being used, since most engines will be there by then.

But yet, for Wine which already does all this (basically, providing a WINAPI, HRESULT, IDirect3D9* stuff and windows event handler / STDMETHODCALLTYPE etc.), it may be a good starting point if they wanted to do a d3d9 to vulkan mapping.

And it will be a nice starting point to port your games from Windows/DX9 to Linux/Vulkan.

Quoting: STiATHmh, I don't really get the point in the project except if it is for academic purpose (which is perfectly fine, don't get me wrong there!). Providing a D3D9->Vulkan mapping for systems which ship with D3D9? In the end, 2020 we'll rather see Vulkan/DX12 things popping out than having D3D9 or even 11 being used, since most engines will be there by then.

But yet, for Wine which already does all this (basically, providing a WINAPI, HRESULT, IDirect3D9* stuff and windows event handler / STDMETHODCALLTYPE etc.), it may be a good starting point if they wanted to do a D3D9 to Vulkan mapping.

In any case, it's good all around. We know that Vulkan is a standard already so that makes old applications (on DX9) still available in the future. This potentially can increase performance for the old DX9 with continued support thanks to Vulkan.

So by your definition, just because such a platform already exist this implantation of DX9 over Vulkan is a waste?
Nah, we all know that alternatives can be adapted, and since it's open source it really can be adapted to Wine if they deem it worthy.


Last edited by Sethox on 28 November 2016 at 1:11 pm UTC

Quoting: STiATHmh, I don't really get the point in the project except if it is for academic purpose (which is perfectly fine, don't get me wrong there!). Providing a d3d9->vulkan mapping for systems which ship with d3d9? In the end, 2020 we'll rather see Vulkan/DX12 things popping out than having D3D9 or even 11 being used, since most engines will be there by then.

But yet, for Wine which already does all this (basically, providing a WINAPI, HRESULT, IDirect3D9* stuff and windows event handler / STDMETHODCALLTYPE etc.), it may be a good starting point if they wanted to do a d3d9 to vulkan mapping.

If this matures, it means that GPU vendors can stop worrying about D3D9 support in their future drivers, putting all their focus on Vulkan support, which is good news for us.
As others have stated, it would also help Wine eventually.

Quoting: STiATHmh, I don't really get the point in the project except if it is for academic purpose (which is perfectly fine, don't get me wrong there!). Providing a d3d9->vulkan mapping for systems which ship with d3d9? In the end, 2020 we'll rather see Vulkan/DX12 things popping out than having D3D9 or even 11 being used, since most engines will be there by then.

But yet, for Wine which already does all this (basically, providing a WINAPI, HRESULT, IDirect3D9* stuff and windows event handler / STDMETHODCALLTYPE etc.), it may be a good starting point if they wanted to do a d3d9 to vulkan mapping.

Let me disagree with your last point. To concentrate on this translation layer, I believe it is just easier to work on... Windows, which already supports both APIs, and as such, is just one DLL remplacement away, not involving wine.

That's interesting, but I wonder if there isn't reusable things in gallium nine, which does the same, albeit a bit higher level.
It also has its use cases where G9 isn't available, too.

Is that under a MIT or BSD styled license? I'm confused as to why they wouldn't use a FSF approved one. If so, they should be more clear as it would otherwise be dead on arrival and confuse people.


Last edited by t3g on 28 November 2016 at 4:08 pm UTC

This could help port older Dx9 games over to Linux. And possibly with better performance than before.
Also im happy to see the project coming along nicely, we need something other than directx12... Long Live Vulkan!

Will probably be good for porting older games to Linux or running them in Wine.

I like how he approaches the problem, with realistic measure of success and roadmap. His reasoning is laid out and makes sense as well.

Additional kudos if he can keep up with his plans for the next three years!

Its possible that DX9-VK will give better performance then DX9-OGL, in saying that it would have been more useful if it was DX11, but I guess you gotta start somewhere.