🤖 AI Summary
High-order radial basis function-generated finite differences (RBF-FD) methods for solving the Poisson equation on distributed-memory systems face an inherent trade-off between numerical accuracy and communication overhead.
Method: This work systematically quantifies how approximation order affects compute-communication efficiency, and proposes a target-accuracy-driven adaptive order selection strategy. The approach integrates RBF-FD derivative approximation, OpenMPI-based inter-node communication, and OpenMP-based intra-node parallelism, implemented atop an explicit iterative solver on CPU clusters.
Contribution/Results: We establish, for the first time, a quantitative relationship between problem size and optimal approximation order. Our strategy significantly reduces total time-to-solution for prescribed accuracy—achieving 1.8×–2.5× speedup over fixed-order baselines in representative benchmarks. The work delivers a reusable performance modeling and tuning paradigm for deploying high-order meshfree methods in distributed environments.
📝 Abstract
Meshless methods approximate operators in a specific node as a weighted sum of values in its neighbours. Higher order approximations of derivatives provide more accurate solutions with better convergence characteristics, but they come at the cost of including more neighbours. On the accuracy-per-compute time basis we know that increasing the approximation order is beneficial for a shared memory computer, but there is additional communication overhead when problems become too large and we have to resort to distributed memory systems. Meshless nodes are divided between systems in spatially coherent subdomains with approximations at their edges requiring neighbouring value exchange. Performance optimization is then a balancing act between minimizing the required number of communicated neighbours by lowering the approximation order or increasing it to enable faster convergence. We use the radial basis function-generated finite difference method (RBF-FD) to approximate the derivatives that we use to solve the Poisson equation with an explicit iterative scheme. Inter-system communication is provided by Open MPI, while OpenMP is used for intra-system parallelisation. We perform the analysis on a homogenous CPU-based cluster where we examine the behaviour and attempt to determine the optimal parameterisation with the goal of minimizing the computational time to reach a desired accuracy.