Performance Measurement and Support Software for the Message-Driven Model on Multiprocessors
Holm, John Gustaf
This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.
Permalink
https://hdl.handle.net/2142/81184
Description
Title
Performance Measurement and Support Software for the Message-Driven Model on Multiprocessors
Author(s)
Holm, John Gustaf
Issue Date
1997
Doctoral Committee Chair(s)
Banerjee, Prithviraj
Department of Study
Electrical Engineering
Discipline
Electrical Engineering
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Keyword(s)
Engineering, Electronics and Electrical
Language
eng
Abstract
It is important to study the characteristics of parallel applications in order to evaluate their impact on the architecture of multiprocessors and multicomputers, and their software support. We propose to evaluate some aspects of the usefulness of the message-driven model of computation for parallel programming. For certain scientific SPMD programs that exhibit a substantial amount of idle time, it will be shown that message-driven programming can help reduce idle time and increase speedup with minimal compiler support. Specifically, the message-driven model of computation will be shown to be an effective trade-off between full context-switched threading and no threading support. Message-driven code has been shown to be useful to program irregular, unstructured applications. This dissertation will use the ProperCAD II library as a testbed to analyze six large applications, and demonstrate the usefulness of some tools for parallel programming, developed with the message-driven model in mind. This dissertation reports on a detailed evaluation of message-driven parallel programs on two classes of parallel architectures, a shared-memory multiprocessor and a message-passing multicomputer. The parallelism structure, the communication characteristics, locality characteristics, grain sizes of computations, and detailed measurements of system time, idle time, and user time in these applications are examined in detail. Various problem sizes and system sizes are examined. After thorough examination of the codes, a case study of a specific application is done, and the application is optimized with the help of the tools and analysis presented in this dissertation.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
