I’m updating an OptiX prime model with RTP_MODEL_HINT_ASYNC, but when I profile I see it’s blocking and if I call rtpModelGetFinished immediately after it always returns finished.
I am using host buffer vertices and indices that I have locked.
It doesn’t run async on context type CUDA or CPU
Optix 5.1, Windows 10, GTX 1070, Driver: 416.34
I’m wondering if this is expected on my driver/card or a bug? Docs seem to suggest this should be async.