Scalable Metropolis Monte Carlo for simulation of hard shapes

TL;DR

The paper introduces HPMC, a scalable, open-source Monte Carlo simulation toolkit for hard particle shapes, optimized for CPU and GPU parallelism, enabling efficient large-scale simulations of diverse particle geometries.

Contribution

HPMC is a new, highly optimized, parallel Monte Carlo toolkit supporting a wide variety of shapes, with significant performance improvements over serial methods.

Findings

HPMC achieves 10 million sweeps in 10 minutes on 96 CPU cores.

The toolkit scales efficiently to large systems, completing 16.8 million particles in 1.4 hours on 2048 GPUs.

Supports diverse particle shapes and ensembles, enabling advanced simulations.

Abstract

We design and implement HPMC, a scalable hard particle Monte Carlo simulation toolkit, and release it open source as part of HOOMD-blue. HPMC runs in parallel on many CPUs and many GPUs using domain decomposition. We employ BVH trees instead of cell lists on the CPU for fast performance, especially with large particle size disparity, and optimize inner loops with SIMD vector intrinsics on the CPU. Our GPU kernel proposes many trial moves in parallel on a checkerboard and uses a block-level queue to redistribute work among threads and avoid divergence. HPMC supports a wide variety of shape classes, including spheres / disks, unions of spheres, convex polygons, convex spheropolygons, concave polygons, ellipsoids / ellipses, convex polyhedra, convex spheropolyhedra, spheres cut by planes, and concave polyhedra. NVT and NPT ensembles can be run in 2D or 3D triclinic boxes. Additional integration schemes permit Frenkel-Ladd free energy computations and implicit depletant simulations. In a benchmark system of a fluid of 4096 pentagons, HPMC performs 10 million sweeps in 10 minutes on 96 CPU cores on XSEDE Comet. The same simulation would take 7.6 hours in serial. HPMC also scales to large system sizes, and the same benchmark with 16.8 million particles runs in 1.4 hours on 2048 GPUs on OLCF Titan.