ti-raytrace

  • The code use Taichi programming language

Current implement acceleration


How to run

  • First config your anaconda workspace, and open the anaconda prompt
  • Second you need install taichi language, pip install taichi
  • Last you type ti Render.py, that’s all

Some image produced by this project

  • metal

  • glass

  • non metal


Performance

  • 100000 triangle mesh get 30fps with 2070super

Other algorithm:

  • radix sort

  • morton 3d code generation

  • compact & downweep


Reference

  • Nvidia 2012: Maximizing Parallelism in the Construction of BVHs,Octrees, and k-d Trees

  • Lauterbach 2009:Fast BVH Construction on GPUs

GitHub

https://github.com/lyd405121/ti-raytrace