Different (and naive) implements of data structure for collision detection.

This project is inspired by https://github.com/ttvd/spatial-collision-datastructures. I have used his code in "Core.cpp" (on visualizing the scene).

The performance is much better than the above one. Mainly because I removed the polymorphism of "point". After all, "you don't pay for what you don't use". You can extend and reuse the code by using PointTraits<T>. I believe that template is the correct way to write a cpp program.

(need C++17, opengl, openmp)

mkdir build && cd build
cmake ..
make

And we will get two executable files. One is TestMain, for testing the correctness and speed of the implements. One is GlutMain, for visualizing the scene. For the latter one, you can use different cmd args to launch it. For description of the args, type

./GlutMain --help

The precision of float32 is really unacceptable in terms of collision detection between shapes... So maybe I will have to change all the api related to float