CS338 Parallel Processing final project. A path tracer implemented sequentially and parallelized on the GPU with the CUDA computing platform.
Implementation based on Peter Shirley's Ray Tracing in One Weekend.
- Construct large input scene to use for all testing
- Collect benchmark performance data from sequential program
- Collect performance data from 4 different block size configurations
- Assign more computational work per thread
- Reduced branching
- Metropolis light transport
- Importance sampling