Forgive me if this is a stupid question, as it certainly seems like you’ve focused on performance, but…
Is this not very similar to what PhysX ultimately does behind the scenes for overlap queries? The collision objects are stored in various acceleration structures to make it fast to do spatial queries, so I’m curious as to any specific numbers you have regarding the performance improvement of using your plugin to do a common sphere overlap test rather than just using the existing PhysX API.
I’m quite interested in using this plugin if it does prove to be significantly faster than a sphere overlap, so please don’t take this as an attack. I’d simply like to see more detailed profiling numbers before making a purchase. 