GOVERNMENT

STUTTGART, GERMANY -- By Uwe Harms, Harms-Supercomputing-Consulting --HLRS (Höchstleistungsrechenzentrum, High-Performance Computer Center Stuttgart) has put its 6-day parallel programming workshop into the Internet. It is availbale in three versions: ONLINE, Online presentation of the slides without audio information; ONLINE-English, Online presentation of the slides with english audio information. (Further audio-based presentations will be available soon.) The audio information is coded as RealMedia Playlists. They require at least the free RealPlayer 8 Basic. PDF, a full copy of the slides with 2 slides per sheet. Based on HLRS Parallel Programming Workshops we made the total course material online available. It includes the full audio information. The course contains for example: - Message Passing Interface (MPI, part 1) - Parallel Architectures and Programming Models - MPI-1 - Introduction to the Message Passing Interface - MPI Overview, MPI Process Model, Point-to-Point, Non-Blocking Communication - Writing Message-Passing Parallel Programs with MPI, Course Notes from the EPCC, - Parallel debugging and TotalView - Message Passing Interface (MPI, part 2) and OpenMP - Collective Communication - Other MPI-1 features MPI-1 Exercises - OpenMP - Overview and execution model, Work sharing directives, Data environment, Pitfalls, Exercises - Assure - detection of OpenMP race - MPI-2 and advanced MPI programming - MPI-2 overview, parallel file I/O, parallel file I/O, one-sided Communication - Optimization of MPI applications - MPI Performance Studies on Cray T3E and SGI Origin - Tools for MPI Performance Analysis and Debugging - VAMPIR and other tools for performance analysis - Advanced OpenMP programming and Vectorization - OpenMP - Enhancments in OpenMP 2.0 - OpenMP - Cluster extensions, Performance tuning - MPI on hybrid systems / MPI + OpenMP (talk) - NEC HPC platforms (April 2001) - NEC SX-5 - Vectorization on NEC Supercomputers - NEC SX-5 - Shared Memory parallelization on NEC Supercomputers - NEC psuite (online demonstration) - Hitachi SR8000 - Introduction and Hardware Architecture (April 2001) - Hitachi SR8000 - Programming models and tuning - Domain Decomposition and Load Balancing - Numerical libraries - Parallel numerics - Object oriented parallel programming with C++ - High Performance Fortran (HPF), Exercises - Skeletons and Solutions For more information visit www.hlrs.de/organization/