Effects of Partitioning and Scheduling Sparse Matrix Factorization on Communication and Load Balance

Abstract

We present a block-based, automatic partitioning and scheduling methodology for sparse matrix factorization on distributed memory systems. Using experimental results, we analyze this technique for communication and load imbalance overhead. To study the performance effects, we compare these overheads with those obtained from a straightforward 'wrap-mapped' column assignment scheme. All experimental results were obtained using test sparse matrices from the Harwell-Boeing data set. The results show that there is a communication and load balance trade-off. The block-based method results in lower communication cost whereas the wrap-mapped scheme gives better load balance.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Oct 01, 1991
Accession Number
ADA244296

Entities

People

  • Sesh Venugopal
  • Vijay K. Naik

Tags

Communities of Interest

  • C4I
  • Energy and Power Technologies
  • Ground and Sea Platforms

DTIC Thesaurus Topics

  • Algebra
  • Algorithms
  • Automatic
  • Classification
  • Computations
  • Computer Science
  • Data Sets
  • Engineering
  • Grain Size
  • Hot Spots
  • Linear Algebra
  • Linear Systems
  • Load Distribution
  • Numbers
  • Scheduling (Production)
  • Sparse Matrix
  • Triangles

Fields of Study

  • Computer science
  • Engineering

Readers

  • Linear Algebra
  • Radio communications and signal processing.
  • Tribology (the study of the boundary interaction between sliding surfaces, lubrication, wear and friction).