High Performance Computing
Hosted by Prof. Andrea Walther, at the Universitaet Paderborn
Location: Building N, Room N2.206, Pohlweg 55, 33098 Paderborn
see: http://www.uni-paderborn.de/universitaet/anreise-lageplan
Monday 5th to Wednesday 7th February 2018
Presentations by Dr. Rolf Rabenseifner HLRS, http://www.hlrs.de/home/ and Kshitij Kulshreshtha, Institute for Mathematics, UPB
Monday 5th February
- Parallel Architectures and Programming Models
- MPI Introduction to the Message Passing Interface, Overview & Process model
- Messages and Point-to-Point Communication
- Non-Blocking Communication
- Collective Communication
- Error handling
- Probe, Cancel, Persistent Requests
- The new Fortran module
Tuesday 6th February
- Groups & Communicators, Environment Management
- Virtual Topologies
- Derived Datatypes
- OpenMP – Overview and execution model, work sharing and Data Environment
- OpenMP – Summary and Pitfalls, Heat example and Extensions
- Verifying an OpenMP Parallelization with the Intel Inspector XE
Wednesday 7th February
- Parallelization of Explicit and Implicit Solvers
- One-sided Communication and Shared memory
- Parallel file, MPI and Threads
- Process Creation and Management
- Algorithmic Differentiation with parallelism using Open MP & MPI