Scalable Parallel Numerical CSP Solver

TL;DR

This paper introduces a parallel solver for numerical constraint satisfaction problems that effectively scales across multiple cores, significantly improving computational efficiency through cooperative search space management.

Contribution

The paper presents a novel parallel approach for NCSPs that achieves high scalability and efficiency by coordinating multiple worker solvers.

Findings

Achieved up to 119-fold speedup with 256 cores

Demonstrated effective load balancing among workers

Scalable performance on parallel computing architectures

Abstract

We present a parallel solver for numerical constraint satisfaction problems (NCSPs) that can scale on a number of cores. Our proposed method runs worker solvers on the available cores and simultaneously the workers cooperate for the search space distribution and balancing. In the experiments, we attained up to 119-fold speedup using 256 cores of a parallel computer.