Blog 2 ​

I started working on a particle simulator, with the final goal of simulating basic molecules. So far I've implemented particles, charged particles, rendering, and some basic physics calculations using Couloumb's law. I knew the final product would have thousands of particles, so I started with a quadtree to store them instead of a list. I also made a basic profiler, but it sort of fails if there's too many particles due to the way I keep track of time spent by a certain task.

Most of the time spent by the program is in locating nearby charged particles. Even compressing the quadtree to reduce children uses a tiny amount of time, but so many comparisons for each charged particle takes forever. Currently the performance is pretty good after a few optimizations, but if I need to add significantly more particles I'll need to switch to something like a k-d tree or R-tree.

One of the major decisions I've made with the simulator is that the laws of physics will be very different for the most part. Simulating actual molecules is both difficult and expensive, since quantum mechanics exists and applies at that level. I'll have to create my own kind of unique chemistry, but I'll start with some analogues for real life atoms and similar bonding mechanics caused by electrons.