The Geometric Tools Engine is a collection of source code for computing in the fields of mathematics, geometry, graphics, image analysis and physics. The engine is written in C++ 14 and supports high-performance computing using CPU multithreading and general purpose GPU programming (GPGPU). Portions of the code are described in various books as well as in PDF documents available at the Geometric Tools Website.
The Geometric Tools Engine is licensed under the Boost Software License 1.0.
The mathematics code is in a header-only library, GTMathematics. A mathematics library with GPU-based implementations is provided, GTMathematicsGPU. The CPU-based common graphics engine code is in its own library, GTGraphics. DirectX 11 wrappers are provided for graphics and applications, GTGraphicsDX11 and GTApplicationsDX11, on Microsoft Windows 10. OpenGL 4.5 wrappers are provided for graphics and applications, GTGraphicsGL45 and GTApplicationsGL45, on Microsoft Windows 10 and on Linux. A small number of files exist to use GLX and X-Windows on Linux.
On Microsoft Windows 10, the code is maintained using Microsoft Visual Studio 2015, 2017 and 2019 with Microsoft's compilers or with Intel C++ Compilers 17, 18 and 19.1.
On Ubuntu 20.04.1 LTS, the code is maintained using Visual Studio Code 1.49.2 and CMake 3.15.2, NVIDIA graphics drivers, OpenGL 4.5 and gcc 9.3.0,
On Fedora 34, the code is maintained using Visual Studio Code 1.49.2 and CMake 3.18.3, NVIDIA graphics drivers, OpenGL 4.5 and gcc 11.1.1.
The repository contains many sample applications to illustrate some features of the engine. Top-level solutions/makefiles exist to build everything in the repository. Please read the installation and release notes to understand what is expected of your development environment.
Here is a Bibtex entry you can use to cite this project in a publication.
@misc{GeometricToolsEngine5,
title = {{G}eometric {T}ools {E}ngine, Version 5},
author = {David Eberly},
howpublished = {\url{https://github.com/davideberly/GeometricTools}}
}
- Replacing the LCP approach by a geometric approach for the intersection of a finite cylinder and a box to fix the robustness problem reported by Seb Wouters. This is a longer-term project than I had planned. Wouters' approach uses convex hulls of 2D projections. I am implementing an approach that avoids this when the cylinder direction has 1 or 2 zero components. When the direction has no zero components, the explicit projection can be avoided and reduces computation time. The project includes writing a PDF that describes the ideas (of LCP and its robustness failure, of Wouters' approach, and my variation).
- Fixing a bug in the code for intersection of an infinite cylinder and a plane. The new code also supports finite cylinders. I am writing a PDF to describe the algorithms.
- Adding a minor improvement to the TriangulateEC algorithm as describe in my online PDF.
- The never ending attempts to finish the GTL implementation. Adding unit tests takes a lot of time, and technical support of GTE takes higher precedence when bugs are reported.