Side projects

Fast utility functions for R

R language, however great, is slow to handle calculations done on billions of rows of data. In such cases, it is faster to call an external compiled function directly from R to streamline the data analysis without needing to switch to another language. In that spirit, I implemented a set of utility functions in Rust and made it available as a R package. Please see the README for how to use it, and the man pages for more details on the implemented functions.

One-dimensional interacting stochastic many-body particle system (GPU & CPU)

The studying interacting stochastic many-body particle systems analytically is limited to a few simple cases. As a result, it is imperative to study such complex systems numerically. However, these simulations can take up days if not weeks, due to the sheer number of particles that might be needed in the simulations to approximate the thermodynamic limit. Here, I started implementing one such codebase in Rust that can be extended and optimised for such purposes. Moreover, I’m now developing a GPU-accelerated version of this codebase in order to simulate 10M particles under a day using Apple or Nvidia GPUs.

This work is still in progress.