GPU acceleration for RigidStatic <> RigidKinematic collisions

[SumantBagri]

Hi,

I was wondering if the current version of the PhysX SDK (5.2.1) provides GPU acceleration for RigidKinematic <> RigidStatic collisions. Specifically, if I create RigidStatic and RigidKinematic actors with triangle mesh geometries and set the PxSceneFlag::eENABLE_GPU_DYNAMICS flag in the sceneDesc would the contact generation be performed on the GPU ? Or is the GPU accelerated contact generation supported onlly for RigidDynamic objects ?

Also, PxSceneFlag::eENABLE_PCM is enabled by default now right ? (https://nvidia-omniverse.github.io/PhysX/physx/5.2.1/_build/physx/latest/class_px_scene_desc.html#_CPPv4N11PxSceneDesc5flagsE). So we dont need to explicitly enable it in the sceneDesc for GPU narrowphase?

Thank you in advance!

[SumantBagri]

Contact generation from GPU accelerated simulate() for RigidStatic and RigidKinematic is possible with triangle mesh and convex mesh collisions and with a ContactReportCallback. (Also see #180 (comment))

This works for me for now