In the demo 'modern rendering in metal', I saw that you split the bistro 3d model into mesh chunks. Can you tell me which algorithm is used here? kmeans or others? Thanks.

The new Xcode 14 Metal visualization and debugger features are awesome. However, one major pain point for me with indirect compute command buffers and indirect argument buffers, is that the buffer viewer is not typed when viewing buffers bound to an argument buffer. It defaults to displaying everything as float types. This is for compute encoding, I haven't tried it for render encoders. For non-trivial types (structs), it's painful to figure out which piece of memory is what, therefore I convert everything back to regular MTLComputeCommandEncoders (non bindless) when debugging especially complicated compute pipelines, which is far from ideal. Are there any workarounds or is this a current limitation?