Including that for the instance's meshlet_slice_start using the other buffer we uploaded gives us the global meshlet index in the general meshlet asset buffer. The thread can then produce out the two calculated IDs for your cluster. The ways taken are just about just copied from Nanite (besides Nanite does https://bookmarkblast.com/story17413852/how-github-io-games-can-save-you-time-stress-and-money