Want to accelerate your distance, angle, and dihedral calculations with blazing speed? You are in luck! As part of an MDAnalysis CSI EOSS project, Richard Gowers and I developed distopia, a high-performance Python library designed to efficiently compute these geometric measures under periodic boundary conditions (PBCs) using explicitly vectorised SIMD. Distopia is compatible with common PBC types, is highly performant on modern CPUs at single and double precision with minimal code changes, and is usable either as a standalone package or from within MDAnalysis’ distance backend. We think we have one of, if not the fastest package for calculating distances, angles and dihedrals under PBCs on a CPU but would love to be proved wrong! ...
TLDR: Molecular simulation can enter its big-data era if we solve some software challenges and go cloud-native. Geoscience has faced many of these challenges and we can learn from their approach. Atomistic molecular simulations (otherwise known as molecular dynamics) have advanced significantly from the first simulations of liquid argon with a Lennard-Jones potential in 1964. With Moore’s law driving compute ever cheaper and software and expertise becoming democratised, more people than ever are using molecular simulation to drive insight and innovation in their scientific domain. ...
Old blog is gone, new year new me? Hoping to share some of my thoughts on technical and some personal matters here.