Local discovery of system architecture - application parameter sensitivity: an empirical technique for adaptive grid applications

I.R. Corey1, J.R. Johnson1, J.S. Vetter1
1Computing Applications & Research Department, Lawrence Livemore National Laboratory, USA

Tóm tắt

This study presents a technique that can significantly improve the performance of a distributed application by allowing the application to locally adapt to architectural characteristics of distinct resources in a distributed system. Application performance is sensitive to system architecture-application parameter pairings. In a distributed or Grid enabled application, a single parameter configuration for the whole application will not always be optimal for every participating resource. In particular, some configurations can significantly degrade performance. Furthermore, the behavior of a system may change during the course of the run. The technique described here provides an automated mechanism for run-time adaptation of application parameters to the local system architecture. Using a scaled-down simulation of a Monte Carlo physics code, we demonstrate that this technique can conservatively achieve speedups up to 65% on individual resources and may even provide order of magnitude speedup in the extreme case.

Từ khóa

#Computer architecture #Distributed computing #Runtime #Laboratories #Performance evaluation #Application software #Processor scheduling #Computer applications #Degradation #Monte Carlo methods

Tài liệu tham khảo

10.1109/SC.2000.10024 10.1177/109434200001400407 vetter, 2002, Performance Assertions: a Performance Diagnosis Tool, submitted whaley, 2001, Automated Emprical Optimizations of Software and the ATLAS Project, Parallel Computing, 27 10.1016/S0167-739X(99)00025-4 10.1177/109434209701100205 foster, 1998, The Grid Blueprint for a New Computing Infrastructure hollingsworkth, 1999, Prediction and Adaptation in Active Harmony, Cluster Computing, 2 10.1109/ICASSP.1998.681704 10.1016/0167-8191(93)90055-P 10.1109/IPDPS.2002.1016570 10.1177/109434209701100304 bilmes, 1997, Optimizing Matrix Multiply using PHiP AC: a Portable, High-Performance, ANSI C Coding Methodology, International Conference on Supercomputing mirkovic, 2001, Automatic Performance Tuning in the UHFFT Library, International Conference on Parallel Computing