Mesa 21.0 Merges Direct3D 12 Gallium3D Driver

[phoronix]

Phoronix: Mesa 21.0 Merges Direct3D 12 Gallium3D Driver

Mesa 21.0 will allow running OpenCL and OpenGL on top of Gallium3D for any hardware on Windows 10 supporting Direct3D 12 acceleration...

http://www.phoronix.com/scan.php?page=news_item&px=Mesa-21.0-Direct3D-12-Gallium3D

[wswartzendruber]

They really are working hard to keep their advertising platform (Windows 10) relevant...

[karolherbst]

shitty comments incomming in 3.. 2.. 1..

[xxmitsu]

So this is like d3d12 gallium backend. NOT gallium driver, right ?

[schmidtbag]

I don't think I even know a Gallium D3D12 driver was in development. This could be a great alternative to those with CPUs that can't spare enough threads for DXVK.

[jrch2k8]

I don't think I even know a Gallium D3D12 driver was in development. This could be a great alternative to those with CPUs that can't spare enough threads for DXVK.

this is not an DX12 for linux driver, is a OpenGL/CL driver for Linux over DX12 for WSL2

[JPFSanders]

Microsoft is trying to embrace Linux by absorbing it fully.

[Danny3]

Can't Microsoft keep their patches for their vendor-locked crap in their own branch and stop polluting open source projects ?
I don't see the point of open source developers being slowed down because of this bloatware that only Microsoft needs for their spyware OS.

[tildearrow]

shitty comments incomming in 3.. 2.. 1..

0!

Embrace Extend Extinguish Embrace Extend Extinguish Embrace Ex-

In a more serious way, I thought this was a typo, but I am surprised...
Is this even a real Direct3D 12 driver (like Gallium-Nine) or just a shim for WSL?

[mannerov]

Can't Microsoft keep their patches for their vendor-locked crap in their own branch and stop polluting open source projects ?
I don't see the point of open source developers being slowed down because of this bloatware that only Microsoft needs for their spyware OS.

If you look only in the interest of the Linux user, this merge is positive.

Indeed it means Microsoft was ready to invest money into getting this merged.
Thus I would assume it sees some value into having this backend.
It would make sense for Microsoft to maintain and improve the backend after the merge. And these improvements are likely to be affecting code that benefits everyone.
For example if they decide to work on improving the OpenCL backend, that's a significant win.

You could argue the new backend adds maintenance burden, but I think this should be minimal. The gallium architecture enables to have many drivers, backends, etc live together, and you do not usually need to modify anything in other drivers/backends/etc when working on your driver/backend/etc.