kririae / fledge

A personal renderer project aiming to have fun and simultaneously practice my coding skills.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Fledge Renderer

A simple scene with some artifacts. This geometry is from Ariel simulator.

Description

Originally A small physically based render to implement volumetric algorithms. Currently, a platform to implement seemingly non-trivial or non-conventional but (probably) useless render architectures and algorithms.

The ultimate and not yet achieved goal for this render is to integrate as diverse techniques within a consistent, fast, aggressive, yet elegant structure as possible ;) In a word, going a step further from conventional PBRT practice!

Anyway, this is solely a personal project aiming to have fun and simultaneously practice my coding skills.

Usage

Unix Makefile is recommended. Use make help to get more information.

Code Structure

  • src/:
    • common/: Some header files that could be used in both CPU and GPU backends.
    • experimental/: Some files that perform some unstable optimizations.
    • external/: Those libraries that could optimize some portions of the renderer.
    • gpu/: Some code that implements the interface specified in common/.
    • main/: A directory that contains the executables.
    • materials/: Some material implementations that have the common interface.
    • optix/: OptiX backend
  • assets/: Some scene settings for both mitsuba and this renderer.
  • tests/: Unit tests

TODO

  • Integrate Embree as leaf nodes to accelerate intersection.
  • Support XML scene loading
  • Integrate OpenImageDenoise
  • Add low-discrepancy sampler
  • Add naive OptiX backend(WIP)
  • Replace OpenVDB with NanoVDB to implement other tracking methods.(WIP)
  • Add vectorized code generated by ispc to optimize extremity calculations, e.g., EstimateDirect(), BRDF evaluations, Light Sampling.
  • Implement some BVH algos.
  • Add Path Guiding Integrator

About

A personal renderer project aiming to have fun and simultaneously practice my coding skills.

License:MIT License


Languages

Language:C++ 93.6%Language:CMake 4.2%Language:C 0.8%Language:Makefile 0.6%Language:Cuda 0.5%Language:Python 0.3%