COMP 5704: Parallel Algorithms and Applications in Data Science
School of Computer Science Carleton University, Ottawa, Canada
Project Title: CUDA Accelerated Raytracing
Name: Sivakumaran Gowthaman
E-Mail: sivakumarangowthaman@cmail.carleton.ca
Project Outline: In CUDA accelerated raytracing implementation, when threads are called recursively to carry ray-surface intersection and shading functions, stack overflow occurs due to the limited shared memory of the GPU. This results in low quality rendered images. This paper addresses this issue by converting the recursive functions into iterative functions with explicitly defined stack depth. Redundant ray-surface intersection restricts the raytracing implementation from achieving real time rendering. This paper handles this issue by deploying Binary Volume Hierarchy (BVH) acceleration structure.