Parallel SFC-based mesh partitioning and load balancing

TL;DR

This paper introduces improvements to a parallel mesh partitioner based on Hilbert Space-Filling Curves and applies it to dynamic load balancing in large-scale simulations, enhancing efficiency for Exascale computing.

Contribution

It presents novel enhancements to an SFC-based mesh partitioner and demonstrates its application for real-time load balancing in high-performance computing environments.

Findings

Improved partitioner performance on large meshes

Effective dynamic load balancing during simulations

Partitioning aligns with device performance for efficiency

Abstract

Modern supercomputers allow the simulation of complex phenomena with increased accuracy. Eventually, this requires finer geometric discretizations with larger numbers of mesh elements. In this context, and extrapolating to the Exascale paradigm, meshing operations such as generation, adaptation or partition, become a critical bottleneck within the simulation workflow. In this paper, we focus on mesh partitioning. In particular, we present some improvements carried out on an in-house parallel mesh partitioner based on the Hilbert Space-Filling Curve. Additionally, taking advantage of its performance, we present the application of the SFC-based partitioning for dynamic load balancing. This method is based on the direct monitoring of the imbalance at runtime and the subsequent re-partitioning of the mesh. The target weights for the optimized partitions are evaluated using a least-squares approximation considering all measurements from previous iterations. In this way, the final partition corresponds to the average performance of the computing devices engaged.