Visualization of do-loop performance
Performance visualization is the use of graphical display techniques for the analysis of performance data in order to improve understanding of complex performance phenomena. Performance visualization systems for parallel programs have been helpful in the past and they are commonly used in order to improve parallel program performance. However, despite the advances that have been made in visualizing scientific data, techniques for visualizing performance of parallel programs remain ad hoc and performance visualization becomes more difficult as the parallel system becomes more complex.
The use of scientific visualization tools (e.g. AVS, Application Visualization System) to display performance data is becoming a very powerful alternative to support performance analysis of parallel programs. One advantage of this approach is that no tool development is required and that every feature of the data visualization tool can be used for further data analysis.
In this paper the Do-Loop-Surface (DLS) display, an abstract view of the performance of a particular do-loop in a program implemented using AVS, is presented as an example on how a data visualization tool can be used to define new abstract representations of performance, helping the user to analyze complex data potentially generated by a large number of processors.
Unable to display preview. Download preview PDF.
- 1.Advanced Visual Systems Inc. AVS User's Guide, Release 4, May 1992.Google Scholar
- 2.Bangalore, P. Private electronic mail communication. Subject: MPI version of LINPACK, March 1995.Google Scholar
- 3.Ken Brodlie. Scientific Visualization — past, present and future. Nuclear Instruments and Methods in Physics Research — A, 354:104–111, 1995.Google Scholar
- 4.Alva Couch. Categories and Context in Scalable Execution Visualization. Journal of Parallel and Distributed Computing, 18:195–204, 1993.Google Scholar
- 5.Cwik, T. and van de Geijn, R. and Patterson, J. Application of Massively Parallell Computation to Integral Equation Models of Electromagnetic Scattering (Invited Paper). J. Opt. Soc. Am. A, 11(4), April 1994.Google Scholar
- 6.Dongarra, Jack and van de Geijn, Robert and Walker, David. LAPACK Working Note 43: A look at Scalable Dense Linear Algebra Libraries. Technical report, University of Tennessee, Oak Ridge National Laboratory, and University of Texas, May 1992.Google Scholar
- 7.Gropp, William and Lusk, Ewing. Users Guide for the ANL IBM SPx. MCS, Argonne National Laboratory, January 1995. Draft.Google Scholar
- 8.Gropp, William and Lusk, Ewing and Skjellum, Anthony. USING MPI: Portable Parallel Programming with the Message-Passing Interface. Scientific and ENgineering Computation Series, MIT Press, 1994.Google Scholar
- 9.M. Heath and J. Etheridge. Visualizing the performance of parallel programs. IEEE Software, pages 29–39, September 1991.Google Scholar
- 10.Michael Heath. Recent Developments and Case Studies in Performance Visualization using ParaGraph. In Workshop on Performance Measurement and Visualization of Parallel Systems, Moravany, Czecho-Slovakia, October 1992.Google Scholar
- 11.T. LeBlanc, J. Mellor-Crummey, and R. Fowler. Analyzing parallel program executions using multiple views. Journal of Parallel and Distributed Computing, 9(2):203–217, 1990.Google Scholar
- 12.Allen Malony and Gregory Wilson. Future directions in parallel performance environments. In Workshop on Performance Measurement and Visualization of Parallel Systems, Moravany, Czecho-Slovakia, October 1992.Google Scholar
- 13.Barton Miller, M. Callaghan, J. Cargille, J. Hollingsworth, B. Irvin, K. Karavanic, K. Kunchithapadam, and T. Newhall. The Paradyn Parallel Performance Measurement Tools. IEEE Computer, 28(11):37–46, November 1995.Google Scholar
- 14.Oscar Naím and Tony Hey. Do-Loop-Surface: An Abstract Performance Data Visualization. Lecture Notes in Computer Science, 797:367–372, April 1994.Google Scholar
- 15.Oscar Naím, Tony Hey, and Ed Zaluska. Do-Loop-Surface: An Abstract Representation of Parallel Program Performance. Concurrency-Practice and Experience, 8(3):205–234, 1996.Google Scholar
- 16.Parasoft Corporation. ParaSoft Express. User's Guide, 1990.Google Scholar
- 17.Daniel Reed, Ruth Aydt, Tara Madhyastha, Roger Noe, Keith Shields, and Bradley Schwartz. The PABLO performance analysis environment. Department of Computer Science, University of Illinois at Urbana-Champaign, 1992.Google Scholar
- 18.Diane Rover and Abdul Waheed. Multiple Domain Analysis Methods. In Proc. of the 3rd ACM/ONR Workshop on Parallel and Distributed Debugging, pages 53–63, May 1993.Google Scholar
- 19.Diane Rover and Charles Wright. Visualizing the Performance of SPMD and Data-Parallel Programs. Journal of Parallel and Distributed Computing, 18:129–146, 1993.Google Scholar
- 20.Sekhar Sarukkai, Doug Kimelman, and Larry Rudolph. A methodology for visualizing performance of loosely synchronous programs. In Scalable High Performance Computing Conference, SHPCC-92, pages 424–432. IEEE Computer Society, April 1992.Google Scholar
- 21.Margaret Simmons and Rebecca Koskela. Performance Instrumentation and Visualization. ACM Press, Frontier Series, 1990.Google Scholar
- 22.Abdul Waheed, Bernd Kronmüller, Roomi Sinha, and Diane Rover. A Toolkit for Advanced Perofrmance Analysis. In International Workshop on Modeling, Analysis, and Simulation of Computers and Telecommunication Systems (MASCOTS'94), Durham NC, January 1994.Google Scholar
- 23.Abdul Waheed and Diane Rover. Performance Visualization of Parallel Programs. In Visualization '93, San Jose, California, October 1993.Google Scholar