This is the open-source version of the Brownian Dynamics code BDX that I developed for the University of Puerto Rico - Mayagüez UPRM. The demo is a minimalistic version and has been implemented in an entirely different way and inspired by LAMMPS source code. The proprietary version of the BDX code leverages AVX512 instructions for performance in HPC CPUs while this demo does not aim to be a HPC grade code. It does aim to be fun to tinker with and it has been like that for me.
The purpose of this demo is to showcase my ability to write C/C++ code, it's not intended for research work because this code does not even commit the simulation data to disk and also it does not perform data processing while the simulation is executing. If you wish to perform Brownian dynamics simulations I encourage you to explore open-source projects such as LAMMPS.
Hopefully the demo can be understood as is without comments all over the place even for developers who are not familiar with Brownian dynamics simulations.
We do not intend to rely on the STL because there's no real need for developing this
application.
We know how many entities we will simulate beforehand and that number will not change
throughout the simulation. We are also able to optimize the application (via domain
partitioning) with a simple user-defined data structure backed by C-style arrays
so that we don't see the need to introduce the std::vector. If we ever need more complex
data structures we shall seriously consider leveraging the STL to develop this
application.
And last but not least, this is a demo meant to showcase my ability to solve problems with just core language features. Just going to add here that the creator of C++, Bjarne Stroupstrup, knows that developers are free to use his programming language as they see fit, he has also said that mastering C++ is not about using the latest features but to know when and how to use them. He also admits it is not possible to implement the most performant features of the STL without (carefully) using all the core features of C++ ( source: Lex Fridman's podcast ). This is enough reason for me to be on the path of mastering the use of C-style pointers; only then I shall be able to appreciate the facilities of the STL such as smart pointers.
I have spent a considerable amount of time studying Quake's source code because I would like to create some games on my spare time. I have learned that I need to know a lot more such as graphics programming in order to understand Quake's source code but found out about how they did memory management and thought that would be useful for some of my projects like this one.
Thus, I borrowed from there (which is fine because it's GPL'd source code like this one) and made some modifications to implement a simple garbage collector GC. The GC makes it possible to abort execution of the simulation if we encounter an error without incurring on memory leaks. This is also the reason why I did not implement destructors since the GC will handle that at the end of the simulation.
Sample Video of a BDX Simulation (click on the image below to watch on youtube):
The video was generated with the actual BDX code (which is the proprietary and fully featured version owned by the University of Puerto Rico - Mayagüez UPRM) not this demo.
I might add videos produced by this open-source version later.