CUDA-memory based canvas. More...
#include <iindex_canvas.h>
Inherits mi::base::Interface_declare< 0x631ca66c, ... >.
Public Member Functions | |
virtual mi::neuraylib::ICanvas_cuda * | get_cuda_canvas () const =0 |
Returns the CUDA memory based canvas for external use, e.g., by a video encoder as input. More... | |
virtual mi::Sint32 | get_device_id () const =0 |
The CUDA memory canvas is hosted on a CUDA device. More... | |
CUDA-memory based canvas.
Using a CUDA-memory based enables NVIDIA IndeX to run rendering and image compositing on the GPU only without ever transferring data from GPUs to main memory. NVIDIA IndeX support the creation of a built-in CUDA canvas that derives from the following interface class.
|
pure virtual |
Returns the CUDA memory based canvas for external use, e.g., by a video encoder as input.
|
pure virtual |
The CUDA memory canvas is hosted on a CUDA device.
The CUDA device ID tells the user by which CUDA device the canvas is managed.