[cover image]

April 1997
ISBN 0-262-72026-4
224 pp.
$27.50 (paper)


by Robert A. van de Geijn

PLAPACK is a library infrastructure for the parallel implementation of linear algebra algorithms and applications on distributed memory supercomputers such as the Intel Paragon, IBM SP2, Cray T3D/T3E, SGI PowerChallenge, and Convex Exemplar. This infrastructure allows library developers, scientists, and engineers to exploit a natural approach to encoding so-called blocked algorithms, which achieve high performance by operating on submatrices and subvectors. This feature, as well as the use of an alternative, more application-centric approach to data distribution, sets PLAPACK apart from other parallel linear algebra libraries, allowing for strong performance and significanltly less programming by the user.

This book is a comprehensive introduction to all the components of a high-performance parallel linear algebra library, as well as a guide to the PLAPACK infrastructure.

Scientific and Engineering Computation series, The MIT Press

with contributions by

Philip Alpatov, Greg Baker, Carter Edwards, John Gunnels, Greg Morrow, and James Overfelt.

Links to The MIT Press:

HTML version


Codes in text

Page 24: Figure 2.3

Page 40: Figure 2.4

Page 69 Figure 4.1
Note: minor corrections in code

Page 70 Figure 4.2
Note: minor corrections in code

Page 146 Figure 8.4
Note: minor corrections in code:

PLA_Pmvector_create_conf_to( a, PLA_PROJ_ONTO_COL, PLA_DUPLICATED , nb_alg, &c1dup );
PLA_Pmvector_create_conf_to( a, PLA_PROJ_ONTO_ROW, PLA_DUPLICATED , nb_alg, &b1dup );

should be

PLA_Pmvector_create_conf_to( a, PLA_PROJ_ONTO_COL, PLA_ALL_COLS , nb_alg, &c1dup );
PLA_Pmvector_create_conf_to( a, PLA_PROJ_ONTO_ROW, PLA_ALL_ROWS , nb_alg, &b1dup );


Page 176: Appendix A (Summary of PLAPACK Routines and their Arguments)

PLA_Obj_set ( obj )

should be

PLA_Obj_set ( obj, datatype, *value )
Last Updated: Feb. 6, 1998