Scientific computing on parallel machines |

This introduction to parallel programming and the message passing
interface (MPI) is part of the doctoral course in Mathematical
Research at Universidad Complutense de Madrid during the academic
year 2007-2008 (announcement in
pdf).

The handouts corresponding to the different lectures are available
here:

Lecture 1: | General introduction to parallel programming | 07.4.2008 | |

Lecture 2: | General introduction to MPI, hello-world | 08.4.2008 | |

Lecture 3: | MPI point-to-point communication, send/recv | 09.4.2008 | |

Lecture 4: | Exercise - parallel search problem | 10.4.2008 | |

Lecture 5: | MPI collective communication, pi | 11.4.2008 | |

Lecture 6: | Exercise - 2D Poisson solver | 14.4.2008 | |

Lecture 7: | Non-contiguous data & Mixed datatypes, search | 15.4.2008 | |

Lecture 8: | Virtual topologies & Communication subsets, search | 16.4.2008 | |

Lecture 9: | Use of linear algebra libraries, dense linear system solve | 17.4.2008 | |

Lecture 10: | Some examples of parallel applications | 18.4.2008 | |

All code studied in this course can be downloaded here.

The following tutorial is the suggested base of our course:
PACS *Introduction to MPI*
Course. You can sign
up to the online version and browse, take self tests, etc.

A complete reference of the MPI library standard is available at NETLIB.

A useful short summary of the syntax and use of each MPI command can be accessed here.

The user guides for SCALAPACK and BLACS are also available at NETLIB.

markus.uhlmann AT ifh.uka.de

Scientific computing on parallel machines |