Modeling pre-Exascale AMR Parallel I/O Workloads via Proxy Applications

TL;DR

This paper develops a simple proxy model to simulate and analyze pre-exascale AMR simulation I/O workloads, aiding in autotuning strategies for future exascale systems.

Contribution

It introduces a linear regression-based proxy model that translates AMReX simulation inputs into I/O workload estimates, validated on Summit supercomputer data.

Findings

MACSio can approximate AMReX I/O workloads with reasonable confidence

The model links simulation parameters to data production rates

Provides a foundation for autotuning in exascale I/O management

Abstract

The present work investigates the modeling of pre-exascale input/output (I/O) workloads of Adaptive Mesh Refinement (AMR) simulations through a simple proxy application. We collect data from the AMReX Castro framework running on the Summit supercomputer for a wide range of scales and mesh partitions for the hydrodynamic Sedov case as a baseline to provide sufficient coverage to the formulated proxy model. The non-linear analysis data production rates are quantified as a function of a set of input parameters such as output frequency, grid size, number of levels, and the Courant-Friedrichs-Lewy (CFL) condition number for each rank, mesh level and simulation time step. Linear regression is then applied to formulate a simple analytical model which allows to translate AMReX inputs into MACSio proxy I/O application parameters, resulting in a simple "kernel" approximation for data production at each time step. Results show that MACSio can simulate actual AMReX non-linear "static" I/O workloads to a certain degree of confidence on the Summit supercomputer using the present methodology. The goal is to provide an initial level of understanding of AMR I/O workloads via lightweight proxy applications models to facilitate autotune data management strategies in anticipation of exascale systems.