HPCTOOLKIT: tools for performance analysis of optimized parallel programs

被引:401
作者
Adhianto, L. [1 ]
Banerjee, S. [1 ]
Fagan, M. [1 ]
Krentel, M. [1 ]
Marin, G. [2 ]
Mellor-Crummey, J. [1 ]
Tallent, N. R. [1 ]
机构
[1] Rice Univ, Dept Comp Sci, Houston, TX 77251 USA
[2] Oak Ridge Natl Lab, Oak Ridge, TN 37831 USA
关键词
performance tools; call path profiling; tracing; binary analysis; execution monitoring; VISUALIZATION;
D O I
10.1002/cpe.1553
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
HPCTOOLKIT is an integrated suite of tools that supports measurement, analysis, attribution, and presentation of application performance for both sequential and parallel programs. HPCTOOLKIT can pinpoint and quantify scalability bottlenecks in fully optimized parallel programs with a measurement overhead of only a few percent. Recently, new capabilities were added to HPCTOOLKIT for collecting call path profiles for fully optimized codes without any compiler support, pinpointing and quantifying bottlenecks in multithreaded programs, exploring performance information and source code using a new user interface, and displaying hierarchical space time diagrams based on traces of asynchronous call path samples. This paper provides an overview of HPCTOOLKIT and illustrates its utility for performance analysis of parallel applications. Copyright (C) 2009 John Wiley & Sons, Ltd.
引用
收藏
页码:685 / 701
页数:17
相关论文
共 45 条
[1]  
AMMONS G, 1997, SIGPLAN C PROGR LANG, P85
[2]  
ANDERSON TE, 1990, PERF E R SI, V18, P115, DOI 10.1145/98460.98518
[3]   An api for runtime code patching [J].
Buck, B ;
Hollingsworth, JK .
INTERNATIONAL JOURNAL OF HIGH PERFORMANCE COMPUTING APPLICATIONS, 2000, 14 (04) :317-329
[4]  
Caubet J, 2001, LECT NOTES COMPUT SC, V2104, P53
[5]  
COARFA C, 2007, ICS 07, P13
[6]  
CORTESI D, 2001, 0073430003 SIL GRAPH
[7]  
DEROSE L, 2001, P INT PAR DISTR PROC
[8]  
FRIGO M, 1998, PLDI 98, P212, DOI DOI 10.1145/277652.277725
[9]  
Froyd N., 2005, P 19 ANN INT C SUPER, P81
[10]  
Furlinger K., 2007, P 13 INT EUR C PAR P, P62