BLACS
Basic Linear Algebra Communication Subprograms

The BLACS, or "Basic Linear Algebra Communication Subprograms", form a linear algebra-oriented message passing interface that may be implemented efficiently and uniformly across a large range of distributed memory platforms.

The length of time required to implement efficient distributed memory algorithms makes it impractical to rewrite programs for every new parallel machine. The BLACS exist in order to make linear algebra applications both easier to program and more portable. It is for this reason that the BLACS are used as the communication layer of the distributed memory linear algebra package SCALAPACK, for instance.

MPI is one example of a distributed memory system. A program written at the BLACS level can run on under MPI. The same program should run correctly on systems that use other distributed memory systems. The key is that on each system, the installation of the BLACS library takes into account the interfact between the standard BLACS routines and the local distributed memory system.

Related Data and Programs

MPI is a FORTRAN90 example directory which demonstrates the use of MPI for parallel computing in distributed memory systems.

PBLAS is a FORTRAN90 example directory which demonstrates the use of the PBLAS, or "parallel BLAS" routines.

SCALAPACK is a FORTRAN90 example directory which demonstrates the use of SCALAPACK.

Reference:

• Jack Dongarra, Clinton Whaley,
  A User's Guide to the BLACS, v1.1,
  LAPACK Working Note 94,
  lawn94.pdf
• blacs_install.pdf, Installing and Testing the BLACS, v1.1;
• Susan Blackford, Jaeyoung Choi, Andrew Cleary, Eduardo D'Azevedo, James Demmel, Inderjit Dhillon, Jack Dongarra, Sven Hammarling, Greg Henry, Antoine Petitet, Ken Stanley, David Walker, Clinton Whaley,
  The ScaLAPACK User's Guide,
  SIAM, 1997,
  ISBN13: 978-0-898713-97-8.
• http://www.netlib.org/blacs/index.html the BLACS website on NETLIB.
• http://www.netlib.org/scalapack/index.html the ScaLAPACK website on NETLIB.

Examples and Tests:

BLACS_PRB is a simple test in which a grid is set up, and each process reports in to the master process:

• blacs_prb.f90, the sample problem;
• blacs_prb.cmd, commands to compile, link and run the sample problem on the IBM LoadLeveler queueing system;
• blacs_prb.csh, commands to compile, link and run on an MPI distributed memory system that uses MPIRUN for execution;
• blacs_prb_output.txt, the output file;

You can go up one level to the FORTRAN90 source codes.