Program winter term 2019/2020
Monday, October 14, 2019, 4:15pm, Hörsaal 28 D 001
Claudia Felser (MPI Dresden)
Topological materials science
Topology, a mathematical concept, recently became a hot and truly transdisciplinary topic in condensed matter physics, solid state chemistry and materials science. Since there is a direct connection between real space: atoms, valence electrons, bonds and orbitals, and reciprocal space: bands, Fermi surfaces and Berry curvature, a simple classification of topological materials in a single particle picture should be possible. Binary phosphides are an ideal material class for a systematic study of Dirac, Weyl and new Fermion physics, since these compounds can be grown as high-quality single crystals. A new class of topological phases that have Weyl points was also predicted in the family that includes NbP, NbAs. TaP, MoP and WP2. Beyond Weyl and Dirac, new fermions can be identified in compounds that have linear and quadratic 3-, 6- and 8- band crossings that are stabilized by space group symmetries. Crystals of chiral topological materials CoSi, AlPt and RhSi were investigated by angle resolved photoemission and show giant unusual helicoid Fermi arcs with topological charges of ±2. In agreement with the chiral crystal structure two different chiral surface states are observed. In magnetic materials the Berry curvature and the classical anomalous Hall (AHE) and spin Hall effect (SHE) helps to identify potentially interesting candidates. As a consequence, the magnetic Heusler compounds have already been identified as Weyl semimetals: for example, Co2YZ, and Co3Sn2S2. The Anomalous Hall angle also helps to identify materials in which a QAHE should be possible in thin films. Heusler compounds with non-collinear magnetic structures also possess real-space topological states in the form of magnetic antiskyrmions, which have not yet been observed in other materials.
Monday, October 28, 2019, 4:15pm, Hörsaal 28 D 001
Thomas Lippert (FZ Jülich)
Taming Complexity in Computational Physics Simulations
In the last ten years, simulation and data analysis have become increasingly complex in computational physics, chemistry and biology. Examples for such multi-physics workflows are found in particle physics, astronomy, earth system research, civil security, neuroscience, medicine, etc. The advent of Deep Learning in Data Analysis and its integration into simulation processes has added a further dimension of complexity.
In contrast, all supercomputers in the Top 500 are based on a monolithic architecture. They assemble heterogeneous nodes consisting of elements such as CPUs and GPUs sharing I/O. They are significantly underutilized as complex nodes make the overall system prone to inefficiencies. Secondly, scalability is costly, as a node needs to perform very complex computations for problems often not being scalable, and the same node must perform scalable computations for problems being easily scalable. A third difficulty is the inclusion of future technologies in workflows, such as quantum computers.
To master these challenges, we propose a disaggregation of resources and their dynamic recomposition through a programming paradigm called modular supercomputing. It gives a new degree of freedom in supercomputing, the dynamical optimal adaptation of program parts to different architectures within a joint high-speed network. Modularity is motivated by a computer theoretical generalization of Amdahl's Law.
Modular supercomputing helps to tame the increasing complexities in computational science: It offers energy-efficient exascale computing, optimized workflows in data analysis or interactive supercomputing, but also the simple integration of future quantum computers and neuromorphic computers.
Monday, November 11, 2019, 4:15pm, Hörsaal 28 D 001
Matthias Wuttig (RWTH Aachen University)
Monday, November 25, 2019, 4:15pm, Hörsaal 28 D 001
Tommaso Calarco (FZ Jülich)
Monday, December 9, 2019, 4:15pm, Hörsaal 28 D 001
Christian Weinheimer (Universität Münster)
Monday, December 16, 2019, 4:15pm, Hörsaal 28 D 001
Yifang Wang (IHEP Chinese Academy of Sciences Beijing)
Monday, January 13, 2020, 4:15pm, Hörsaal 28 D 001
Heino Falcke (Radboud-Universität Nijmegen)