Mesa Developers Eye Removing Clover Once Rusticl OpenCL Code Hits Parity

[Veerappan]

As one of the guys who has worked some on Clover in the past, and did quite a bit of the coding of functions in the libclc library, all I can say is: If we can get Rusticl on par with Clover for R600/SI generations, I'm all for deleting Clover. There's enough additional features and momentum in Rusticl, and Clover ended up being a dead-end due to lack of consistent contributors.

I've done a few custom builds of Rusticl to attempt to get my RX580 running, but haven't had the time to get too far into it yet. Maybe once the weather stops being nice out and I start spending all of my time indoors, if someone hasn't beaten me to it.

[karolherbst]

As one of the guys who has worked some on Clover in the past, and did quite a bit of the coding of functions in the libclc library, all I can say is: If we can get Rusticl on par with Clover for R600/SI generations, I'm all for deleting Clover. There's enough additional features and momentum in Rusticl, and Clover ended up being a dead-end due to lack of consistent contributors.

I've done a few custom builds of Rusticl to attempt to get my RX580 running, but haven't had the time to get too far into it yet. Maybe once the weather stops being nice out and I start spending all of my time indoors, if someone hasn't beaten me to it.

Might want to use this MR https://gitlab.freedesktop.org/mesa/...requests/19232

[Veerappan]

Might want to use this MR https://gitlab.freedesktop.org/mesa/...requests/19232

Well, that's too easy

I'll give it a build and a test and leave any feedback either in the MR or somewhere you can find it.