A refactored repository of gDel3D that works with recent CUDA architectures.

Original repo: https://github.com/ashwin/gDel3D

Original ReadMe:

This program constructs the Delaunay Triangulation of a set of points in 3D using the GPU. The algorithm used is a combination of incremental insertion, flipping and star splaying. The code is written using CUDA programming model of NVIDIA.

Programming authors

• Cao Thanh Tung
• Ashwin Nanjappa

Setup

gDel3D works on any NVIDIA GPU with hardware capability 1.1 onward. However, it works best on Fermi and higher architecture. The code has been tested on the NVIDIA GTX 450, GTX 460, GTX 470, GTX580 (using sm_20) on Windows OS; and GTX Titan on Linux (using sm_30).

To switch from double to single precision, simply define REAL_TYPE_FP32.

For more details on the input and output, refer to: CommonTypes.h (near the end) Demo.cpp DelaunayChecker.cpp.

Build and run

A Visual Studio 2012 project is provided for Windows user.

CMake is used to build gDel3D on Linux, as shown here:

mkdir build
cd build
cmake ..
make
./gflip3d # To run the demo executable

Note: Tested with Ubuntu20, and CUDA 11.4