3 Nov
2022
3 Nov
'22
6:52 p.m.
On Thu Nov 3 18:50:18 2022 +0000, Zebediah Figura wrote:
In theory. In practice that's not going to work very well with vkd3d; there are a lot of missing pieces in our vkd3d -> sm1 path. We should probably adjust this later.
One related point that's also worth noting is that ideally "minimum shader model" wouldn't be interpreted as "only shader model". I think that ideally we'd at least be doing an sm1 and an sm4 shader model for all applicable tests. But that needs more work in the shader_runner infrastructure, and probably more planning with respect to what tests we actually do care about running.