Skip to content

Monte Carlo based path tracer with both CPU(with openMP) and GPU support(with CUDA and Reinforcement Learning optimization). RL ideas from NVidia paper: Learning Light Transport the reinforced way

Notifications You must be signed in to change notification settings

repalash/CUDAPathTracerRL

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CUDA PathTracer with Reinforcement Learning

Monte Carlo based path tracer with both CPU(with openMP) and GPU support(with CUDA). This is done as the course project for the GPU course.

CUDA based renderer optionally uses Reinforcement Learning (based on ideas from NVIDIA paper: Learning Light Transport the reinforced way).

RL assumes that every ray has 8 actions available to choose from in each bounce(based on the 8 octants). So a QTable is built upon that and the most optimum path is chosen. With time, the rays learn the best path to take to reach the light source. Since this is online Reinforcement Learning, there is no training period and the values converge pretty fast.

The main code for this is in renderengine_gpu.cu The parameters for path tracing can be changed in pathtracer_params.h

For more information on implementation and the output, read the short poster or the full report for the project: Poster.pdf and Report.pdf

The project can be built using Clion(on Windows) or cmake. Check out template for Clion for use of CUDA: https://github.com/PalashBansal96/CUDAClionStarterProject/

Renders

Render 1 Render 1 Render 1 Render 1

CPU vs GPU comparision(no RL)

CPU vs GPU

Path tracing with Reinforcement learning on GPU

RL vs GPU RL vs GPU working

We can see here that in the same time, the rays were able to find a better path for indirect lighting, hence better illumination.

More comparisions

RL vs GPU RL vs GPU RL vs GPU

Issues with RL based lighting

RL issues

About

Monte Carlo based path tracer with both CPU(with openMP) and GPU support(with CUDA and Reinforcement Learning optimization). RL ideas from NVidia paper: Learning Light Transport the reinforced way

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published