In-memory integration of existing software components for parallel adaptive unstructured mesh workflows

Abstract

Reliable mesh-based simulations are needed to solve complex engineering problems. Mesh adaptivity can increase reliability by reducing discretization errors, but requires multiple software components to exchange information. Often, components exchange information by reading and writing a common file format. This file-based approach becomes a problem on massively parallel computers where filesystem bandwidth is a critical performance bottleneck. Our data stream and component interface approaches avoid the filesystem bottleneck. In this paper we present our approaches and their use within the PHASTA computational fluid dynamics solver and Albany multiphysics framework. Information exchange performance results are reported on up to 2048 cores of a BlueGene/Q system.