ScaLAPACK: A Linear Algebra Library for Message-Passing Computers

L. S. Blackford, J. Choi, A. Cleary, E. D'Azevedo, J. Demmel,   I. Dhillon, J. Dongarra,
S. Hammarling,   G. Henry,   A. Petitet, K. Stanley, D. Walker,   and R. C. Whaley

Abstract:

This article outlines the content and performance of some of the ScaLAPACK software. ScaLAPACK is a collection of mathematical software for linear algebra computations on distributed-memory computers. The importance of developing standards for computational and message-passing interfaces is discussed. We present the different components and building blocks of ScaLAPACK and provide initial performance results for selected PBLAS routines and a subset of ScaLAPACK driver routines.

 
• Introduction
• Factors Affecting ScaLAPACK Performance
  • Processor Performance and Total Performance
  • ScaLAPACK and the BLAS
  • Data Distribution
  • Installation
• Performance, Portability, and Scalability
  • The BLAS as the Key to (Trans)portable Efficiency
  • Block Cyclic Data Layout as the Key to Load Balancing and Software Reuse
  • The BLACS as an Efficient, Portable, and Adequate Message-Passing Interface
  • Parallel Efficiency
• ScaLAPACK Strategy
  • Software Hierarchy
  • Performance of Selected PBLAS Routines
• Solution of Common Numerical Linear Algebra Problems
  • Factorizations for Solving Linear Equations
  • QR Factorization
• Conclusions
• References
• About this document ...