Characterizing and representing workloads for parallel computer architectures

Abstract

Experimental design of parallel computers calls for quantifiable methods to compare and evaluate the requirements of different workloads within an application domain. Such methods can help establish the basis for scientific design of parallel computers driven by application needs, to optimize performance to cost. In this paper, a framework is presented for representing and comparing workloads, based on the way they would exercise parallel machines. This workload characterization is derived from parallel instruction centroid and parallel workload similarity. The centroid is a workload approximation that captures the type and amount of parallel work generated by the workload on the average. The centroid is a simple measure that aggregates average parallelism, instruction mix, and critical path length. When captured with abstracted information about communication requirements, the result is a powerful tool in understanding the requirements of workloads and their potential performance on target machines. The workload similarity is based on measuring the normalized Euclidean distance (ned) between workload centroids. It will be shown that this workload representation method outperforms comparable ones in accuracy, as well as in time and space requirements. Analysis of the NAS Parallel Benchmark workloads and their performance will be presented to demonstrate some of the applications and insight provided by this framework. This will include the use of the proposed framework for predicting the performance of real-life workloads on target machines, with good accuracy.