Andreas M. Hinz1, Ciril Petr2 1Mathematical Institute, Ludwig-Maximilians-Universität München 2University of Maribor
Eccentricities in Hanoi Graphs
Hanoi graphs H_n^p are the mathematical model of the Tower of Hanoi with p pegs and n discs. While for p = 3 virtually everything is known, the metrical properties of the graphs with p ≥ 4 become more complex and show some surprising phenomena. In order to understand the structures better, we computed eccentricities, i.e. maximum distances from the vertices, by means of breadth-first searches with delayed duplicate detection.
Keynote
Dmitry Krasnov Technische Universität Ilmenau
Tensor-product elliptic solver for liquid-metal magnetohydrodynamics
9:50
Keynote
Frauke Gräter Heidelberg Institute for Theoretical Studies Interdisciplinary Center for Scientific Computing, Heidelberg University
(Bond) Breaking news: Mechanoradicals in Collagen
10:10
Markus Klein Universität der Bundeswehr München Numerische Methoden in der Luft- und Raumfahrttechnik, LRT 1
Lewis number and pressure effects on weakly turbulent premixed Bunsen flames
Harald Köstler Friedrich-Alexander-Universität Erlangen-Nürnberg
Scalable multi-physics simulations of antidunes using the waLBerla framework
10:30
Mauricio Jentys Technical University of Munich Chair of Aerodynamics and Fluid Mechanics
Aerodynamic Analyses of Coupling Hybrid Laminar Flow Control and Variable Camber on a Transport Aircraft Wing
Computational fluid dynamics (CFD) results and modelling approaches, considering an aircraft wing with a combined hybrid laminar flow control and variable camber system are presented in this contribution. The goal of the technology coupling consists in increasing aerodynamic efficiency. Results with correlation-based turbulence models and linear stability for transition predicition are presented, where both reflect the intended synergy potential of the technology coupling.
Konstantin Lion Institute of Physics and IRIS Adlershof, Humboldt-Universität zu Berlin
Composition and Structure of β-Ga2O3 (001) under Realistic (T, p) Conditions
Ritama Kar Indian Institute of Technology Kanpur (IITK), India
Molecular Dynamics at the Fourth Rung of DFT Jacob's Ladder Made Two Orders of Magnitude Faster
10:50
Coffee break
11:10
Francesco Knechtli Uni Wuppertal
Charmonium and Confinement from Lattice QCD
Quantum ChromoDynamics (QCD) is the theory of strong interactions. It explains why quarks and gluons are not observed in isolation but are confined into hadrons. Confinement can be studied by means of Monte Carlo simulations of QCD discretized on a Euclidean space-time lattice. Among the hadrons we study in this project charmonium, which are states made of a charm quark and a charm anti-quark.
Patricio Muñoz Max Planck Institute for Solar System Research, Göttingen
Kinetic simulations of electron-scale magnetic reconnection in space plasma turbulence
Neutron star merger simulations for multi-messenger astrophysics
11:30
Jan Wilhelm Institute of Theoretical Physics, Uni Regensburg
Developing a GW algorithm for the electron band structure of complex 2D materials
Semiconducting two-dimensional materials are an ideal platform to study electronic excitations thanks to the good experimental accessibility and the strong binding energy of electron and hole. The GW+Bethe-Salpeter approach (GW+BSE) has been successful in analyzing electronic excitations in single-layer 2D materials, but the application of GW+BSE is challenging for 2D double layers and moiré structures. This is because the large unit cells in these structures contain hundreds to thousands of atoms, resulting in a high computational cost for GW+BSE calculations. In this talk, I will present a low-scaling GW algorithm for the electronic band structure of 2D materials. The GW algorithm is based on localized basis functions and potentially allows for the inclusion of more than a thousand atoms in the simulation. I will present benchmark calculations on Supermuc-NG.
Lukas Krenz Department of Computer Science, TUM School of Computation, Information and Technology
Elastic-acoustic coupling for large scale earthquake simulations
In this talk, we discuss elastic-acoustic coupling and its implementation in SeisSol. We discuss aspects of the modeling and show two examples: the 2018 Palu, Sulawesi earthquake-tsunami and sound generation from induced earthquakes caused by an enhanced geothermal system in the Helsinki metropolitan area.
Clotilde Cucinotta Department of Chemistry, Imperial College London
Towards a Realistic Modelling of Electrified Interfaces the Nanoscale
In this presentation, I will discuss the challenges associated with simulating electrified interfaces at the nanoscale from first principles. Specifically, I will focus on simulating the effects of an applied potential to an electrochemical cell, using realistic models for the charged electrode-electrolyte interface. I will also share recent progress made in simulating the double layer of the fundamental Pt-water interface and its response to changes in the potential applied to the cell. To achieve this, we have developed a general ab initio electrode-charging approach.
11:50
Sergio Hoyas Escuela Técnica Superior de Ingeniería del DiseñoUniversitat Politècnica de València
Symmetry theory and turbulence
Johannes Kleinert Arbeitsgruppe Luftfahrzeugaerodynamik Institut für Aerodynamik und Gasdynamik Uni Stuttgart
Interactions between wing wake and horizontal tail plane flow in high-speed stall conditions
Patrick Wollny, Dominik Meller Lehrstuhl Fluiddynamik, Energie- und Material-Prozess-Institut (EMPI), Universität Duisburg-Essen
Detailed, time-resolving simulations of turbulent reacting multi-phase flows
A wide variety of processes in energy and process engineering require mathematical investigations using simulations that fully resolve the physical domain in time and space. Often, these processes involve phase transitions from at least one phase to the other (i.e., solid to gas phase and vice versa). Our work presents advanced models and resulting massively parallel simulations developed and performed by the authors at the Chair of Fluid Dynamics at the University of Duisburg-Essen. The presentation focuses on novel approaches for modeling nanoparticle synthesis in a top-to-bottom process (gas to particle), as well as coal combustion for low-emission flames and the validation of the presented models.
The first case presents the modeling of heterogeneous nanoparticles (platinum on alumina support) in a turbulent spray flame reactor, which is done by introducing a multi-population balance model that handles the size spectra of dispersed (gas phase) and deposited (on alumina carriers) nanoparticle synthesis.
The second case presents FPV-LES of the co-firing of coal and ammonia. A four-mixture fraction FPV approach is developed, considering all necessary mixture fractions to accurately describe coal and ammonia combustion. The influence of char combustion is found to be negligible in the investigated domain.
The presented simulations, using the newly developed models, show the advantage of new model developments in this field, together with state-of-the-art high-performance computing, as the results shown provide new insights that contribute to the general understanding of these delicate processes.
12:10
Jan Benáček Institute of Physics and Astronomy, University of Potsdam, Centre for Astronomy and Astrophysics, Technical University of Berlin
Revealing pulsar radio emissions in electron-positron plasmas by massively parallel PIC-code simulations
Though decades of studies of coherent pulsar radiation, the physical mechanism of the emission at kinetic micro-scales is still under investigation. One of the proposed mechanisms is the linear acceleration emission that is coherent antenna-type of radiation. The mechanism is based on plasma particles oscillating along magnetic field lines and producing linear acceleration emission.
We studied how plasma bunches/clouds of electron-positron pairs created during spark events in the polar gap region evolve as a function of the plasma temperature and drift velocity between electrons and positrons and how the bunches radiate by the linear acceleration emission. We utilized particle-in-cell simulations of relativistically hot bunches to investigate the non-linear evolution of the bunches. Also, we have implemented a novel treatment by utilizing the plasma current in the simulations to obtain properties of the coherent radio emission mechanism. Our treatment does not require tracking individual plasma macro-particles but can directly utilize aggregated information from the currents about the collective particle motion.
We found that the initial drift velocity between electrons and positrons is the main parameter influencing the bunch evolution. For zero drift, the bunches can expand and overlap in the phase space and form relativistic streaming instability. Otherwise, the bunches are constrained from expansion by ambipolar diffusion effects, oscillating electrostatic fields are formed, and the plasma is strongly heated. Furthermore, we calculated the radio emission properties of both types of the bunch evolutions. We found that bunches constrained from expansion have similar observational characteristics as observed for pulsars. The radiation power oscillates at micro-second time scales, and the spectrum contains a flat part for low frequencies and power-law profiles for higher frequencies. Also, the emitted radiation is relativistically beamed along the pulsar dipole axis.
James Beattie Australian National University
The world’s largest compressible magnetohydrodynamic turbulence simulation
The interstellar medium in galaxies is magnetised and turbulent. The turbulence is driven on large scales by stochastic supernova detonations, gravitational instabilities in the galactic disc, or even mergers between galaxies. Understanding how the energy is transported to smaller scales, where it affects the dynamics of relativistic particles, the structure of the interstellar medium, and processes such as star formation, is a problem that involves a deep understanding of the turbulent energy cascade in the presence of strong magnetic fields and compressions. We have presently run the largest compressible, magnetised turbulence simulations in the world using our modified version of the FLASH code, a finite volume MHD code. These simulations utilise uniform grids of up to 10,080^3 and have been run on roughly 140,000 computing cores for a total of 80 M-core hours. Our kinetic energy spectra, E(k), reveal a dichotomy of turbulent dynamics separated by a transition scale between two cascades: (1) on large scales (low k modes), we find Burgers turbulence E(k) ~ k^{-2}, which is the turbulence of saw-tooth shocks, and (2) on small scales (high k modes), we find a Boldyrev (2006) dynamically aligned energy spectrum E(k) ~ k^{-3/2}, which we corroborate with direct measurements of the cross-helicity. This is the first calculation to demonstrate such a dichotomy. Similar to our previous calculations for supersonic hydrodynamical turbulence in Federrath et al. (2021), we show that incompressible theories of turbulence, which are modified in the presence of a magnetic field, are nested within highly-compressible, supersonic turbulence.
Anna Neuweiler Institute of Physics and Astronomy, University of Potsdam
Analyzing and Interpreting Compact Binary Mergers
About one hundred gravitational wave signals from mergers of compact objects such as black holes and neutron stars have been detected to date. The correct interpretation of the observational data depends on cross-correlation with theoretical models. For this purpose, we have developed a framework that allows us to extract the system parameters from the observational data using Bayes' theorem. Our research focuses on the analysis of the signals, but also on the development of accurate theoretical models. The latter requires numerical-relativity simulations that solve Einstein's field equations together with the equations of general relativistic hydrodynamics. Our resources on SuperMUC-NG have allowed us to significantly improve our infrastructure for analyzing the signals and our in-house numerical-relativistic code. In the presentation, we will discuss some of our recent research highlights.
12:30
Lunch
14:00
User forum 1
HPC at LRZ@GCS
User forum 5
Quantum Computing@LRZ
User forum 8
Education and Training@LRZ
LRZ provides first-class education and training opportunities for the national and European HPC community. We regularly offer education and training on topics such as Application Software, Data Analytics, Deep Learning and AI, HPC, Optimisation, Programming Languages, Quantum Computing, System and Internet Security.
The presentation provides an overview of the education and training activities offered by LRZ and GCS.
14:30
User forum 2
GCS: The Gauss Centre for Supercomputing
Roadmap to Exascale
User forum 6
Mentoring of HPC projects@LRZ
User and Project Administration
Project types: Test, Regular, Large Scale. How to apply for projects on SuperMUC-NG How we can help you with your project
User forum 9
LRZ Data Science Archive
Tutorial on how to store, retrieve, transfer and share SuperMUC-NG long-term data in our LRZ Data Science Archive.
15:00
User forum 3
Supporting AI and Data Analytics on HPC Systems
User forum 7
Application Labs
Software Provisioning and programming support
User forum 10
Research Data Management
FAIR data / dataset publication in HPC: status update on planned services at LRZ and beyond. Additional focus: activities with the National Research Data Infrastructure (NFDI, in particular NFDI4Ing and NFDI4Earth).