instr (1158309), страница 3

Просмтор этого файла доступен только зарегистрированным пользователям. Но у нас супер быстрая регистрация: достаточно только электронной почты!

Текст из файла (страница 3)

This can be done with the toolse_merge. If the target machine does not have a hardware global clock, se_mergewill also establish a global timescale for the event traces by correcting timestamps ifnecessary.(2)Trace Conversion, Analysis, and Visualization: With the help of the utility toolse_convert and the EDF file, the traces can be converted to SDDF format to be usedwith the Pablo performance analysis environment [ROA+91, ROM+92] or toALOG format to be used in the Upshot event display tool [HL91]. It also canproduce a simple, user-readable ASCII dump of the binary trace.The trace files can also processed with the SIMPLE event trace analysis environmentor other TDL/POET based performance analysis tools [mohr90, mohr93a].

Thesetools use the generated TDL description of the trace files to access them.Future versions of the pC++ parallel programming environment will have their ownsophisticated trace analysis and visualization tools.References[BBG+92a]F. Bodin, P. Beckman, D. Gannon, S. Narayana, S. Srinivas, “Sage++: AClass Library for Building Fortran 90 and C++ Restructuring Tools”,Technical Report, CICA, Indiana University, 1992.[BBG+92b]F.

Bodin, P. Beckman, D. Gannon, S. Narayana, S. Yang, “DistributedpC++: Basic Ideas for an Object Parallel Language”, Technical Report,CICA, Indiana University, 1992.[BBG+93]F. Bodin, P. Beckman, D. Gannon, S. Yang, S. Kesavan, A. Malony, B.Mohr, “Implementing a Parallel C++ Runtime System for ScalableParallel Systems”, to appear in: Proceedings of the Supercomputing ‘93Conference, Portland, Oregon, November 15-19, 1993.[gann93]D. Gannon, “Libraries and Tools for Object Parallel Programming”, in:Proc. CNRS-NSF Workshop on Environments and Tools For ParallelScientific Computing, Saint Hilaire du Touvet, France, Elsevier SciencePublisher, Advances in Parallel Computing, Vol.

6, pp. 231-246, 1993.D R A F T 0.211Instrumenting pC++September 8, 1993[GKK82]S.L. Graham, P.B. Kessler, M.K. McKusick, “gprof: a call graph executionprofiler”, in: Proc. SIGPLAN’82 Symp. on Compiler Construction, 1982,pp. 120-126.[HL91]V. Herrarte, E. Lusk, “Studying Parallel Program Behavior with Upshot”,Technical Report ANL-91/15, Mathematics and Computer ScienceDivision, Argonne National Laboratory, August 1991.[mohr90]B. Mohr, “Performance Evaluation of Parallel Programs in Parallel andDistributed Systems”, in: H. Burkhart (Eds.), Proc. CONPAR 90–VAPP IV,Joint International Conference on Vector and Parallel Processing, Zürich,Lecture Notes in Computer Science 457, pp. 176-187, Berlin, Heidelberg,New York, London, Paris, Tokio, Springer Verlag, 1990.[mohr93a]B.

Mohr, “Standardization of Event Traces Considered Harmful or Is anImplementation of Object-Independent Event Trace Monitoring andAnalysis Systems Possible?”, in: Proc. CNRS-NSF Workshop onEnvironments and Tools For Parallel Scientific Computing, Saint Hilairedu Touvet, France, Elsevier Science Publisher, Advances in ParallelComputing, Vol.

6, pp. 231-246, 1993.[mohr93b]B. Mohr, “A Portable Software Event Tracing Module for ParallelMachines”, not yet written.[ROA+91]D. A. Reed, R. D. Olson, R. A. Aydt, T. M. Madhyasta, T. Birkett, D. W.Jensen, B. A. A. Nazief, B. K. Totty, Scalable Performance Environmentsfor Parallel Systems. In: Proceedings of the 6th Distributed MemoryComputing Conference, IEEE, Computer Society Press, pp. 562-569,1991.[RAM+92]D.

A. Reed, R. A. Aydt, T. M. Madhyastha, Roger J. Noe, Keith A. Shields,B. W. Schwartz, An Overview of the Pablo Performance AnalysisEnvironment. Department of Computer Science, University of Illinois,November 1992.D R A F T 0.212.

Характеристики

Список файлов учебной работы