High-performance computing for quantum field theory: a case for Rust in Monte Carlo simulations

Luca Ciucci (University of Pisa)
Serena Bruzzesi (University of Pisa)
Friday session 1 (Zoom) (13:0014:30 BST)
Watch a recording of this talk on YouTube

Monte Carlo methods, particularly Markov chain Monte Carlo (MCMC), play a fundamental role in numerical simulations of quantum field theory (QFT), especially in lattice quantum chromodynamics (Lattice QCD). These simulations require high-performance computing (HPC) techniques due to their extreme computational complexity. However, traditional implementations often rely on legacy languages like C and Fortran, leading to challenges in extensibility, maintainability, safety, and performance optimization.

In this talk, I will explore how Rust, with its modern idioms, strong concurrency model, and a still immature but growing numerical computing ecosystem, can be a compelling alternative for high-performance QFT simulations. I will discuss the advantages Rust offers over traditional approaches, the challenges of integrating it into existing scientific computing workflows, and its potential for Rust-based HPC in theoretical physics.