Sparse binding for RTX acceleration structures

Hello NVIDIA,

i’ve been playing around with your new raytracing API and noticed that there is no sparse binding feature for the acceleration structures. While the underlying memory layout is opaque and there is no meaningful way to support sparse residency, support for sparse binding would make memory reuse for acceleration structures easier. From what i have seen so far this should just be an API/driver change, not a hardware change.

I would really like to see this feature in either a future NV or KHR raytracing extension.

Regards