An Introduction to Performance Debugging for Parallel Computers
- 575 Downloads
Programming parallel computers for performance is a difficult task that requires careful attention to both single-node performance and data exchange between processors. This chapter discusses some of the sources of poor performance, ways to identify them in an application, and a few ways to address these issues.
Unable to display preview. Download preview PDF.
- Message Passing Interface Forum, 1994. “MPI: A message-passing interface standard,”, Int. J. Supercomput. Applics. 8, pp. 159–416. http://www.mcs.anl.gov/mpi/mpi-report/mpi-report.htmlGoogle Scholar
- Gropp, W., Lusk, E., and Skjellum, A., 1994. Using MPI, MIT Press.Google Scholar
- Herrarte, V. and Lusk, E., 1991. “Studying parallel program behavior with Upshot,” Technical Report ANL-91/15, Argonne National Laboratory, Mathematics and Computer Science Division.Google Scholar
- Nicol, D. M., 1990. “Inflated speedups in parallel simulations via malloc,” ICASE Technical Report 90-63.Google Scholar
- Reed, D. A., Aydt, R. A., Noe, R. J., Roth, P. C., Shields, K. A., Schwartz, B., and Tavera, L. F., 1993. “Scalable performance analysis: The Pablo performance analysis environment,” in Proc. Scalable Parallel Libraries Conference, Anthony Skjel-lum, ed., IEEE Computer Society, Los Alamitos.Google Scholar
- Yan, J., Hontalas, P., and Listgarten, S., 1993. The Automated Instrumentation and Monitoring System (AIMS) Reference Manual NASA TM-108795.Google Scholar