In this work, we present the RTX technology that enables hardware-accelerated rendering with real-time ray tracing. The RTX technology uses dedicated graphic card cores, advanced acceleration structures and a new rendering pipeline to accelerate ray tracing.
As part of this thesis, we implemented the path tracing algorithm, an advanced form of ray tracing, on central processing unit, graphics processing unit and using the RTX technology. With the former two implementations, we used Vulkan API. Since RTX uses acceleration structures to accelerate ray tracing, we also used acceleration structures, provided by the LogiSceneGraph library, in our other two implementations.
For analysis, we created two scenes that we used to compare the speed of the path tracing algorithm and the speed of building the acceleration structures in all three implementations. As expected, our GPU implementation is significantly faster than our CPU implementation, and our RTX implementation is even faster than the GPU implementation, approximately six times.
|
