1st Block: Introduction to Hybrid Programming in HPC
Date and Time: November 04, 2016, 09:00 - 17:00
Location: TU Wien, Kontaktraum, EI, Gußhausstraße 27-29, 6th floor
Lecturers: Georg Hager (RRZE / HPC, Uni. Erlangen) and
Rolf Rabenseifner (HLRS, Uni. Stuttgart)
Details and Registration see: http://vsc.ac.at/training/2016/HY-VSC
Abstract:
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 the Message Passing Interface - MPI) with the shared memory parallelization inside of each node (e.g., with OpenMP or MPI-3.0 shared memory). This course analyses the strengths and weaknesses of several parallel programming models on clusters of SMP nodes. 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.
2nd Block: Shared memory parallelization with OpenMP
Date and Time: November 21, 2016, 08:45 - 18:15
Location: TU Wien, FH Internet-Raum FH1, Wiedner Hauptstraße 8-10, ground floor, red
Lecturers: Lukas Einkemmer
(lectures+practicals; Department of Mathematics, University of Innsbruck),
Irene Reichl and Claudia Blaas-Schenner
(practicals only; VSC Team, ZID, TU Wien)
Details and Registration see: http://vsc.ac.at/training/2016/OpenMP
Abstract:
The focus of this 1 day course is on shared memory parallelization with OpenMP for dual-core, multi-core, shared memory, and ccNUMA platforms. This course teaches OpenMP starting from a beginners level. Hands-on sessions (in C and Fortran) will allow users to immediately test and understand the OpenMP directives, environment variables, and library routines. Race-condition debugging tools are also presented.
3rd Block: Parallelization with MPI
Date and Time: November 22 - 24, 2016, 09:00 - 17:30 (24.11. until 16:00)
Location: TU Wien, FH Internet-Raum FH1, Wiedner Hauptstraße 8-10, ground floor, red
Lecturers: Irene Reichl and Claudia Blaas-Schenner (VSC Team, ZID, TU Wien)
Details and Registration see: http://vsc.ac.at/training/2016/MPI-Nov
Abstract:
On clusters and distributed memory architectures, parallel programming with the Message Passing Interface (MPI) is the dominating programming model. In this 3 days course parallel programming with MPI is taught starting from a beginners level. Hands-on sessions (in C and Fortran) will allow users to immediately test and understand the basic constructs of the Message Passing Interface (MPI).
4th Block: A short introduction to the PGAS programming paradigm
Date and Time: Friday, December 16, 2016, 09:00 - 13:00
Location: TU Wien, Freihaus HS 2, Wiedner Hauptstraße 8-10, 2nd floor, yellow
Lecturer: Martina Prugger (Department of Mathematics, University of Innsbruck)
Details and Registration see: http://vsc.ac.at/training/2016/PGAS
Abstract:
The Partitioned Global Address Space (PGAS) paradigm is a programming model for parallel computing that combines ease of use with the possibility to program distributed memory systems. Programming with PGAS is straightforward resembling the user-friendliness of OpenMP that is a standard for shared memory parallelization. With PGAS parallelism is part of the programming language itself, any processor can directly address memory/data on any other processor, hence parallelism can be expressed more easily as compared to library based approaches such as Message Passing Interface (MPI).
In this course I will describe the PGAS programming model and discuss the two dominant PGAS languages (Unified Parallel C (UPC) and Co-array Fortran (CAF) being PGAS extensions to C and Fortran, respectively). Furthermore, I investigate the ease of use as well as the performance of UPC for some problems that have been tackled during my work on the VSC School Project "High-resolution numerical schemes for hyperbolic conservation laws, and their performance on modern HPC architectures".
The goal of this course is to provide participants with an idea how parallelism is expressed in PGAS languages as well as to help code developers decide if PGAS is an option to consider.
This talk is based on work supported by the VSC Research Center funded by the Austrian Federal Ministry of Science, Research and Economy (bmwfw).
5th Block: Molecular Photodynamics Simulations on HPC Systems
Date and Time: Friday, January 20, 2017, 15:00 - 17:00 (NEW TIME)
Location: TU Wien, Freihaus HS 7(NEW LOCATION),, Wiedner Hauptstraße 8-10, 2nd floor, yellow
Lecturer: Felix Plasser (Department of Theoretical Chemistry, University of Vienna)
Details and Registration see: http://vsc.ac.at/training/2017/Plasser
Abstract:
Many areas of current research, such as photovoltaics, photobiology, and phototherapy are characterized by ultrafast processes in the excited state manifold. The computational description of such processes requires explicit dynamical simulations of the nuclear and electronic degrees of freedom, as well as their interplay. Here, on-the-fly surface hopping dynamics offer themselves as an attractive option as they allow to simulate dynamical processes without any bias toward specific nuclear degrees of freedom while still incorporating essential quantum effects. Despite the high potential of these simulations, the computational effort spent becomes a limiting factor, and high-performance computing (HPC) becomes imperative for many applications of current interest. The dynamics simulations require the computation of different terms related to the electronic energy, the nuclear forces, and nonadiabatic state-to-state interactions. Depending on the molecular system studied and the computational infrastructure used, either one of these steps may act as the computational bottleneck. This heterogeneity of the computational steps causes severe challenges for the dynamics simulations as the different computational steps have to be optimized individually.
In this talk, first the field of photodynamics and applications in our group are reviewed. Subsequently, specific method developments carried out within the VSC School Project "Parallelizing the Gradient and Overlap Calculation for MRCI Wavefunctions on HPC Clusters" are discussed. A major achievement is concerned with the development of a flexible and highly efficient code of the computation of wavefunction overlaps. This code was implemented to compute nonadiabatic interaction terms for the multi-reference configuration interaction method and has been extended to various other electronic structure methods. After presenting the technical details of the implementation, a number of high-profile applications currently on their way on the VSC are presented. Due to the flexibility of the developed code, also other application areas have been explored, such as the comparison of general many-electron wavefunctions and the computation of Dyson orbitals. Methods for the computation of atomic forces will be discussed, as well. In this case progress, that has been achieved in- and outside of this project, is explained in detail.
This talk is based on work supported by the VSC Research Center funded by the Austrian Federal Ministry of Science, Research and Economy (bmwfw).
Further topics/events wil be announced as soon as they are scheduled.
