Nova

High performance GPU accelerated ray tracer using OpenCL/CUDA

Examples

Supported Operating Systems

• Linux
• Any operating system capable of running Linux Docker containers (though exposing the GPU and GUI may be tricky)

Supported Hardware

Either:

• Any CPU or GPU supporting OpenCL 2.1+
• A CUDA-enabled GPU

Dependencies For Building

Common:

• CMake 3.5+
• C++17
• OpenMP
• OpenGL 3.3+
• GLFW dependencies

OpenCL:

• OpenCL runtime for your device
• OpenCL C and C++ headers
• Clang 9+ with LLVM/SPIR-V Translator

CUDA:

• NVIDIA drivers
• CUDA Toolkit

Included Dependencies

• Assimp
• GLFW
• Intel Open Image Denoise
• GLAD
• GLM
• Dear ImGui
• ImGuiFileDialog
• toml11
• CLI11
• stb

Build and Run

$ git clone https://github.com/wchang22/Nova.git
$ cd Nova && mkdir build && cd build

# Either
$ export CC=/path/to/clang && export CXX=/path/to/clang++
$ cmake .. -DBACKEND=OpenCL
# Or
$ export CC=/path/to/gcc && export CXX=/path/to/g++
$ cmake .. -DBACKEND=CUDA

$ cmake --build .
$ ./nova # Add --headless to run without the GUI

Suggested - Docker (For OpenCL)

$ cd Nova
$ docker build -t nova .
$ xhost local:root
$ docker run -it --rm --name nova -w /root/Nova/ -e DISPLAY \
  -v /tmp/.X11-unix:/tmp/.X11-unix -v $(pwd):/root/Nova \
  --device /dev/dri:/dev/dri nova
$ cd Nova && mkdir build && cd build
$ cmake .. -DBACKEND=OpenCL
$ cmake --build .
$ ./nova