RAO-SS: A Prototype of Run-time Auto-tuning Facility for Sparse Direct Solvers

TL;DR

This paper introduces RAO-SS, a run-time auto-tuning system for sparse direct solvers that optimizes performance parameters based on input matrices, achieving significant speedups with minimal overhead.

Contribution

The paper presents a novel auto-tuning method integrated into RAO-SS, utilizing fuzzy logic to adapt performance parameters at run-time for sparse solvers.

Findings

Speedup factors of 1.2 on average and 3.6 at maximum compared to default settings.

Autopilot overhead is negligible in RAO-SS.

Effective performance improvement for sparse direct solvers.

Abstract

In this paper, a run-time auto-tuning method for performance parameters according to input matrices is proposed. RAO-SS (Run-time Auto-tuning Optimizer for Sparse Solvers), which is a prototype of auto-tuning software using the proposed method, is also evaluated. The RAO-SS is implemented with the Autopilot, which is middle-ware to support run-time auto-tuning with fuzzy logic function. The target numerical library is the SuperLU, which is a sparse direct solver for linear equations. The result indicated that: (1) the speedup factors of 1.2 for average and 3.6 for maximum to default executions were obtained; (2) the software overhead of the Autopilot can be ignored in RAO-SS.