Presented by the National Center for Supercomputing Applications for University of Illinois Urbana-Champaign students and staff.
Registration: https://forms.illinois.edu/sec/1130423126
Description
This workshop will be composed of three sessions that explore parallel programming and computing on HPC systems at the introductory level. Attendees will learn how to parallelize and run C/C++/Fortran applications using shared memory (OpenMP) and distributed memory (MPI) frameworks with hands-on exercises and access to the Illinois Campus Cluster. General terminology, parallel computing concepts and ideas, measuring performance, identifying bottlenecks, and other relevant topics will also be discussed. The learning topics for each session and other details are below.
Intro to Parallel Computing on HPC Systems - 4/13 at 11 AM
- Parallel computing terminology
- Programming models
- High-performance systems architecture
- Data and task parallelism
- Performance measurement
- Basic OpenMP parallelization
Shared-memory Parallel Computing with OpenMP - 4/20 at 11 AM
- Balance workloads to threads in parallel loops using different schedules (static, dynamic, guided) and tune your loop parallelization for optimal performance
- Explicitly declare data contexts (private, shared) to avoid race conditions and improve the quality of your code, facilitating collaborations
- Use reduction clauses (additions, maximum and minimum values) to calculate properties without thread concurrency
- Use OpenMP library functions to span teams of threads, mix data and task parallelism
Distributed-memory Parallel Computing with MPI - 4/27 at 11 AM
- The message-passing interface paradigm
- Core components of an MPI message: body, envelope
- MPI processes and communicators
- Collective communications: broadcast messages and reductions
Prerequisites
- Basic C/C++ or Fortran programming skills to work on the exercises
- Familiarity with a supercomputing cluster (e.g., command line interface, submitting batch scripts, etc.)
Date and location
This workshop will be presented virtually (Zoom) with two-hour-long sessions on 4/13, 4/20, and 4/27 starting at 11 AM. Registrations will close on 4/6 at 11:59 AM CT.