Topic outline

  • General

    Partnership for Advanced Computing in Europe

    Important note: The course agenda and materials shown below are those of the 2019 edition of the course. All materials will be updated right before the next course takes place at LRZ.

    Most HPC systems are clusters of shared memory nodes. Such SMP nodes can be small multi-core CPUs up to large many-core CPUs. Parallel programming may combine the distributed memory parallelization on the node interconnect (e.g., with MPI) with the shared memory parallelization inside of each node (e.g., with OpenMP or MPI-3.0 shared memory). This course analyzes the strengths and weaknesses of several parallel programming models on clusters of SMP nodes. Multi-socket-multi-core systems in highly parallel environments are given special consideration. MPI-3.0 has introduced a new shared memory programming interface, which can be combined with inter-node MPI communication. It can be used for direct neighbor accesses similar to OpenMP or for direct halo copies, and enables new hybrid programming models. These models are compared with various hybrid MPI+OpenMP approaches and pure MPI. Numerous case studies and micro-benchmarks demonstrate the performance-related aspects of hybrid programming.

    Hands-on sessions are included on both days. Tools for hybrid programming such as thread/process placement support and performance analysis are presented in a "how-to" section. This course provides scientific training in Computational Science, and in addition, the scientific exchange of the participants among themselves. The course is a PRACE Advanced Training Center event. It is organized by LRZ in cooperation with HLRS, RRZE, and VSC (Vienna Scientific Cluster).

    Agenda & Content (preliminary)

    1st day

    09:30   Registration
    10:00   Welcome
    10:05   Motivation
    10:15   Introduction
    10:45   Programming Models
                 - Pure MPI
    11:05   Coffee Break
    11:25    - Topology Optimization
    12:05       Practical (application aware Cartesian topology)
    12:45    - Topology Optimization (Wrap up)
    13:00   Lunch
    14:00    - MPI + MPI-3.0 Shared memory
    14:30      Practical (replicated data)
    15:00   Coffee Break
    15:20    - MPI Memory Models and Synchronization
    16:00      Practical (substituting pt-to-pt by shared memory)
    16:45   Coffee Break
    17:00      Practical (substituting barrier synchronization by pt-to-pt)
    18:00   End

    19:00   Social Event at Gasthof Neuwirt (self paying)

    2nd day

    09:00   Programming Models (continued)
                 - MPI + OpenMP
    10:30   Coffee Break
    10:50       Practical (how to compile and start)
    11:30       Practical (hybrid through OpenMP parallelization)
    13:00   Lunch
    14:00    - Overlapping Communication and Computation
    14:20       Practical (taskloops) 
    15:00   Coffee Break
    15:20    - MPI + OpenMP Conclusions
    15:30    - MPI + Accelerators
    15:45   Tools
    16:00   Conclusions
    16:15   Q&A
    16:30   End

    Date: Monday, January 28 09:30 - Tuesday, January 29, 2019 16:30
    Location:  LRZ Building, University campus Garching, near Munich, Kursraum 2 (H.U.010)


    Georg Hager (RRZE), Rolf Rabenseifner (HLRS), Claudia Blaas-Schenner and Irene Reichl (VSC Team, TU Wien)

    PRACE web page:

    PRACE evaluation:

    Course material (here ☺):