Summary: Introduction to OpenMP

This workshop introduces the OpenMP compiler directives to scientists who are interested in writing programs for shared-memory parallel computers, or who want to convert existing serial code to parallel. No previous knowledge about parallel programming is required, but we assume some basic background in programming, preferably with the Fortran or C programming languages. The use of OpenMP has become the de facto industry standard for parallel programming on shared-memory machines, such as HPCVL's Sun servers. It also makes implicit use of the multi-core nature of current CPU's, for instance the Sun Niagara 2 systems. Most of the examples are in Fortran, but the C/C++ languages are considered as well. Here is a short outline of the contents:

• Introduction to parallel programming, especially on SMP machines
• OpenMP compiler directives
• Problems and Pitfalls of shared-memory programming and how to avoid them
• Loop parallelism
• Explicit parallel regions
• Synchronization

Summary: Introduction to MPI

The "Introduction to MPI" workshop is directed at programmers and scientists, with a basic background in programming, who want to acquire basic skills in "parallelizing" code for distributed-memory systems such as the HPCVL Victoria Falls and Beowulf clusters. No prior knowledge of MPI or other message-passing systems is required. However, some background in Unix operating systems and programming in Fortran, C, or other languages would be helpful. The following subjects will be addressed:

• MPI Basics (Programming Environments, Data Types, Communication)
• Runtime Environments
• Parallel Principles and Programming Steps
• Combination of MPI with OpenMP
• Parallel Scheduling
• User-Defined Data Types
• All examples are in Fortran, C and C++

Summary: Using the NAG Numerical Library

HPCVL in collaboration with Numerical Algorithms Group (NAG) is hosting two 2-day workshops on Using the NAG Numerical Library, which is a comprehensive collection of mathematical and statistical algorithms for programmers and application developers. These routines are accessible from a variety of environments, including standard languages such as Fortran, C and C++, as well as packages like MATLAB, Maple and Excel, and provide a straightforward way to incorporate powerful, reliable and accurate algorithms into developers' applications. The training session will include an introduction to the functionality of the library, along with several examples of its use. In particular, it will highlight those routines which have been specially developed and tuned for use on Symmetric Multi-Processor (SMP) systems. Here is a brief outline:

Day 1:

1. Overview of the NAG Library, with notes on speed, accuracy,building a solution, working with the Library
2. An example from each chapter of the NAG documentation, with special emphasis on linear algebra

1. Overview of NAG SMP Library, Parallel Library, with notes on special considerations for cluster computing
2. Brief introductions to other interfaces to NAG Library (.NET,JAVA, C++, MATLAB, EXCEL, R)