Using intrinsic performance counters to assess efficiency in task-based parallel applications

Document Type

Conference Proceeding

Publication Date

7-18-2016

Abstract

The ability to measure performance characteristicsof an application at runtime is essential for monitoring the behavior of the application and the runtime system on the underlying architecture. Traditional performance measurement tools do not adequately provide measurements of asynchronous task-basedparallel applications, either in real-time or for post-mortem analysis. We propose that this capability is best performed directly by the runtime system for ease in use and to minimize conflicts and overheads potentially caused by traditional measurement tools. In this paper, we describe and illustrate the use of theperformance monitoring capabilities in the HPX runtimesystem. We describe and detail existing performance counters made available through HPX's performance counter framework and demonstrate how they are useful to understanding application efficiency and resource usage at runtime. This extensive framework provides the ability to asynchronously query softwareand hardware counters and could potentially be used as the basis for runtime adaptive resource decisions. We demonstrate the ease of porting the Inncabs benchmark suite to the HPX runtime system, the improved performance of benchmarks that employ fine-grained task parallelism when ported to HPX, and the capabilities and advantages of using thein-situ performance monitoring system in HPX to give detailed insight to the performance and behavior of the benchmarks and the runtime system.

Publication Source (Journal or Book title)

Proceedings - 2016 IEEE 30th International Parallel and Distributed Processing Symposium, IPDPS 2016

First Page

1692

Last Page

1701

This document is currently not available here.

Share

COinS