Scalable detection of MPI-2 remote memory access inefficiency patterns

Authors:
Marc-André Hermanns;Markus Geimer;Bernd Mohr;Felix Wolf
Affiliations:
Laboratory for Parallel Programming, German Research School for Simulation Sciences, Germany;Forschungszentrum Jülich, Jülich Supercomputing Centre, Germany;Forschungszentrum Jülich, Jülich Supercomputing Centre, Germany;Forschungszentrum Jülich, Jülich Supercomputing Centre, Germany, Laboratory for Parallel Programming, German Research School for Simulation Sciences, Germany, Department of Computer Scie ...
Venue:
International Journal of High Performance Computing Applications
Year:
2012

Citing 8
Cited 1

Automatic performance analysis of hybrid MPI/OpenMP applications

Journal of Systems Architecture: the EUROMICRO Journal - Special issue: Evolutions in parallel distributed and network-based processing
Performance Tool Support for MPI-2 on Linux

Proceedings of the 2004 ACM/IEEE conference on Supercomputing
A Scalable Implementation of a Finite-Volume Dynamical Core in the Community Atmosphere Model

International Journal of High Performance Computing Applications
The Tau Parallel Performance System

International Journal of High Performance Computing Applications
A scalable tool architecture for diagnosing wait states in massively parallel applications

Parallel Computing
The Scalasca performance toolset architecture

Concurrency and Computation: Practice & Experience - Scalable Tools for High-End Computing
Specification of inefficiency patterns for MPI-2 one-sided communication

Euro-Par'06 Proceedings of the 12th international conference on Parallel Processing
Event-Based measurement and analysis of one-sided communication

Euro-Par'05 Proceedings of the 11th international Euro-Par conference on Parallel Processing

Understanding the formation of wait states in applications with one-sided communication

Proceedings of the 20th European MPI Users' Group Meeting

Quantified Score

Hi-index	0.00

Visualization

Abstract

Wait states in parallel applications can be identified by scanning event traces for characteristic patterns. In our earlier work we defined such inefficiency patterns for MPI-2 one-sided communication, although still based on a serial trace-analysis scheme with limited scalability. In this article we show how wait states in one-sided communications can be detected in a more scalable fashion by taking advantage of a new scalable trace-analysis approach based on a parallel replay, which was originally developed for MPI-1 point-to-point and collective communication. Moreover, we demonstrate the scalability of our method and its usefulness for the optimization cycle with applications running on up to 32,768 cores.