Investigate the slowdown with VCM

[yuphin]

Compared to 2022 drivers, the current version of the VCM algorithm seems a lot slower (~%50% !). Investigate why.

[tigrazone]

VCM slowed down up to 50% or more after merging with bsdf branch.
Maybe material processing is slower or so.
Previous and current version of renderer I compile with VS 2019

[yuphin]

Yes, but there is an another source of slowdown I've yet to determine. Regarding the current BSDFs, I've plans to add callable shaders in the future which should hopefully improve the performance.

[tigrazone]

What you mean as "callable shaders"?

[tigrazone]

Maybe good ideas is

• use precompiled spv, not compile every time
• use ray_query because ray_query works faster

[tigrazone]

seems problem with sample_light_Le(), sample_light_Li().
sepearated version of this functions for every case was faster - w/o unneeded functionality.
I used older build and change only bsdf and VCM still works fast as it was before

[yuphin]

Thanks for looking into it. Yes, I think that part of the codebase needs to be revised. I'll keep that in mind going forward.