TY - JOUR
AB - The rise of exascale supercomputers has fueled competition among GPU vendors, driving lattice QCD developers to write code that supports multiple APIs. Moreover, new developments in algorithms and physics research require frequent updates to existing software. These challenges have to be balanced against constantly changing personnel. At the same time, there is a wide range of applications for HISQ fermions in QCD studies. This situation encourages the development of software featuring a HISQ action that is flexible, high-performing, open source, easy to use, and easy to adapt. In this technical paper, we explain the design strategy, provide implementation details, list available algorithms and modules, and show key performance indicators for SIMULATeQCD, a simple multi-GPU lattice code for large-scale QCD calculations, mainly developed and used by the HotQCD collaboration. The code is publicly available on GitHub.
AU - Mazur, Lukas
AU - Bollweg, Dennis
AU - Clarke, David A.
AU - Altenkort, Luis
AU - Kaczmarek, Olaf
AU - Larsen, Rasmus
AU - Shu, Hai-Tao
AU - Goswami, Jishnu
AU - Scior, Philipp
AU - Sandmeyer, Hauke
AU - Neumann, Marius
AU - Dick, Henrik
AU - Ali, Sajid
AU - Kim, Jangho
AU - Schmidt, Christian
AU - Petreczky, Peter
AU - Mukherjee, Swagato
ID - 46120
JF - Computer Physics Communications
TI - SIMULATeQCD: A simple multi-GPU lattice code for QCD calculations
ER -
TY - JOUR
AU - Altenkort, Luis
AU - Eller, Alexander M.
AU - Francis, Anthony
AU - Kaczmarek, Olaf
AU - Mazur, Lukas
AU - Moore, Guy D.
AU - Shu, Hai-Tao
ID - 46119
IS - 1
JF - Physical Review D
SN - 2470-0010
TI - Viscosity of pure-glue QCD from the lattice
VL - 108
ER -
TY - JOUR
AU - Bruns, Julia
AU - Hagena, Maike
AU - Gasteiger, Hedwig
ID - 46155
JF - Teaching and Teacher Education
KW - Education
SN - 0742-051X
TI - Professional Development Enacted by Facilitators in the Context of Early Mathematics Education: Scaling up or Dilution of Effects?
VL - 132
ER -
TY - BOOK
ED - Biehler, Rolf
ED - Liebendörfer, Michael
ED - Gueudet, Ghislaine
ED - Rasmussen, Chris
ED - Winsløw, Carl
ID - 46157
SN - 1869-4918
TI - Practice-Oriented Research in Tertiary Mathematics Education
ER -
TY - CHAP
AU - Kostan, Anastassija
ED - Loos, Christian
ED - Stephan, Paul
ID - 46174
T2 - Widerständige Glieder. Der Leib als politischer Standort
TI - Feministische Neomaterialismen: Die ontologische Immanenz und posthumanistische Performativität von Natur, Kultur und Technik
ER -
TY - JOUR
AB - While FPGA accelerator boards and their respective high-level design tools are maturing, there is still a lack of multi-FPGA applications, libraries, and not least, benchmarks and reference implementations towards sustained HPC usage of these devices. As in the early days of GPUs in HPC, for workloads that can reasonably be decoupled into loosely coupled working sets, multi-accelerator support can be achieved by using standard communication interfaces like MPI on the host side. However, for performance and productivity, some applications can profit from a tighter coupling of the accelerators. FPGAs offer unique opportunities here when extending the dataflow characteristics to their communication interfaces.
In this work, we extend the HPCC FPGA benchmark suite by multi-FPGA support and three missing benchmarks that particularly characterize or stress inter-device communication: b_eff, PTRANS, and LINPACK. With all benchmarks implemented for current boards with Intel and Xilinx FPGAs, we established a baseline for multi-FPGA performance. Additionally, for the communication-centric benchmarks, we explored the potential of direct FPGA-to-FPGA communication with a circuit-switched inter-FPGA network that is currently only available for one of the boards. The evaluation with parallel execution on up to 26 FPGA boards makes use of one of the largest academic FPGA installations.
AU - Meyer, Marius
AU - Kenter, Tobias
AU - Plessl, Christian
ID - 38041
JF - ACM Transactions on Reconfigurable Technology and Systems
KW - General Computer Science
SN - 1936-7406
TI - Multi-FPGA Designs and Scaling of HPC Challenge Benchmarks via MPI and Circuit-Switched Inter-FPGA Networks
ER -
TY - CHAP
AU - Topalović, Elvira
AU - Drepper, Laura
AU - Härtel, Kira
ED - Bangel, Melanie
ED - Rautenberg, Iris
ID - 46180
T2 - Schriftspracherwerb im Spannungsfeld zwischen Wissen und Können
TI - Lerngegenstand „Doppelkonsonanz“ in der Grundschule: Explizites und implizites Lernen im Längsschnitt
ER -
TY - CHAP
AU - Hansmeier, Tim
AU - Kenter, Tobias
AU - Meyer, Marius
AU - Riebler, Heinrich
AU - Platzner, Marco
AU - Plessl, Christian
ED - Haake, Claus-Jochen
ED - Meyer auf der Heide, Friedhelm
ED - Platzner, Marco
ED - Wachsmuth, Henning
ED - Wehrheim, Heike
ID - 45893
T2 - On-The-Fly Computing -- Individualized IT-services in dynamic markets
TI - Compute Centers I: Heterogeneous Execution Environments
VL - 412
ER -
TY - CONF
AU - Opdenhövel, Jan-Oliver
AU - Plessl, Christian
AU - Kenter, Tobias
ID - 46190
T2 - Proceedings of the 13th International Symposium on Highly Efficient Accelerators and Reconfigurable Technologies
TI - Mutation Tree Reconstruction of Tumor Cells on FPGAs Using a Bit-Level Matrix Representation
ER -
TY - CONF
AU - Faj, Jennifer
AU - Kenter, Tobias
AU - Faghih-Naini, Sara
AU - Plessl, Christian
AU - Aizinger, Vadym
ID - 46188
T2 - Proceedings of the Platform for Advanced Scientific Computing Conference
TI - Scalable Multi-FPGA Design of a Discontinuous Galerkin Shallow-Water Model on Unstructured Meshes
ER -
TY - CONF
AU - Prouveur, Charles
AU - Haefele, Matthieu
AU - Kenter, Tobias
AU - Voss, Nils
ID - 46189
T2 - Proceedings of the Platform for Advanced Scientific Computing Conference
TI - FPGA Acceleration for HPC Supercapacitor Simulations
ER -
TY - CHAP
AB - We present a concept for quantifying evaluative phrases to later compare rating texts numerically instead of just relying on stars or grades. We achievethis by combining deep learning models in an aspect-based sentiment analysis pipeline along with sentiment weighting, polarity, and correlation analyses that combine deep learning results with metadata. The results provide new insights for the medical field. Our application domain, physician reviews, shows that there are millions of review texts on the Internet that cannot yet be comprehensively analyzed because previous studies have focused on explicit aspects from other domains (e.g., products). We identify, extract, and classify implicit and explicit aspect phrases equally from German-language review texts. To do so, we annotated aspect phrases representing reviews on numerous aspects of a physician, medical practice, or practice staff. We apply the best performing transformer model, XLM-RoBERTa, to a large physician review dataset and correlate the results with existing metadata. As a result, we can show different correlations between the sentiment polarity of certain aspect classes (e.g., friendliness, practice equipment) and physicians’ professions (e.g., surgeon, ophthalmologist). As a result, we have individual numerical scores that contain a variety of information based on deep learning algorithms that extract textual (evaluative) information and metadata from the Web.
AU - Kersting, Joschka
AU - Geierhos, Michaela
ED - Cuzzocrea, Alfredo
ED - Gusikhin, Oleg
ED - Hammoudi, Slimane
ED - Quix, Christoph
ID - 46205
SN - 1865-0929
T2 - Data Management Technologies and Applications
TI - Towards Comparable Ratings: Quantifying Evaluative Phrases in Physician Reviews
VL - 1860
ER -
TY - JOUR
AU - Ficara, Elena
AU - Beall, Jc
ID - 46209
IS - 2
JF - History and Philosophy of Logic
TI - Hegelian Conjunction, Hegelian Contradiction
VL - 44
ER -
TY - JOUR
AU - Ficara, Elena
AU - Priest, Graham
ID - 46210
IS - 2
JF - History and Philosophy of Logic
TI - The Formalization of Dialectics
VL - 44
ER -
TY - JOUR
AU - Weber, Daniel
AU - Schenke, Maximilian
AU - Wallscheid, Oliver
ID - 46213
JF - IEEE Access
KW - General Engineering
KW - General Materials Science
KW - General Computer Science
KW - Electrical and Electronic Engineering
SN - 2169-3536
TI - Steady-State Error Compensation for Reinforcement Learning-Based Control of Power Electronic Systems
VL - 11
ER -
TY - CONF
AU - Weber, Daniel
AU - Schenke, Maximilian
AU - Wallscheid, Oliver
ID - 46212
T2 - 2023 International Conference on Future Energy Solutions (FES)
TI - Safe Reinforcement Learning-Based Control in Power Electronic Systems
ER -
TY - JOUR
AB - Ernährungsarmut wird als Problem in Deutschland weitgehend vernachlässigt. Von materieller Ernährungsarmut betroffen sind Menschen, denen es aufgrund fehlender finanzieller Mittel nicht möglich ist, sich gesundheitsfördernd zu ernähren. Etwa 3,5 % der Bevölkerung (rund drei Millionen Menschen) sind durch materielle Ernährungsarmut gefährdet. Damit einher geht oft auch soziale Ernährungsarmut, die Menschen von der sozialen Teilhabe ausschließt, die z. B. durch gemeinsames Essen ermöglicht wird. Essen ist nicht nur von Bedeutung für die körperliche Gesundheit, sondern auch zentral für die psychische Gesundheit und das Wohlbefinden sowie für die soziale Zugehörigkeit, was sich wiederum auf die körperliche Gesundheit auswirkt. Die sozialen Funktionen des Essens werden vielfach übersehen – und bei der Berechnung des Regelbedarfs ausgeblendet. Gerade die Corona-Pandemie hat gezeigt, wie fehlende soziale Kontakte und sozialer Ausschluss sich auf die physische, psychische und soziale Gesundheit auswirken können. Aus Sicht des WBAE (Wissenschaftliche Beirat für Agrarpolitik, Ernährung und gesundheitlichen Verbraucherschutz) ist deshalb der integrative Blick auf die materielle und soziale Ernährungsarmut zentral, da sie sich gegenseitig verstärken können. Diese Stellungnahme befasst sich mit der Frage, wie sich die Corona-Pandemie auf von Ernährungsarmut gefährdete Menschen ausgewirkt hat und was daraus gelernt werden kann. Ein wichtiger Aspekt ist der zeitweilige Wegfall von Essensangeboten in Kitas, Schulen, Tafeln und anderen karitativen Essensangeboten, die ein Sicherheitsnetz für Menschen in Ernährungsarmut darstellen. Die Stellungnahme zieht auf dieser Basis auch Schluss-folgerungen zur Ernährungsarmut außerhalb von Pandemiebedingungen. Im Hinblick auf die Notwendigkeit der Vermeidung von Fehlernährung in Haushalten, die von Ernährungsarmut bedroht sind, kommt der WBAE zu dem Schluss, dass diese Risikogruppe während der Pandemie nicht ausreichend im Fokus der Maßnahmen zur Abmilderung der Pandemiefolgen stand. Dies zeigt sich schon daran, dass kaum Daten erhoben wurden. Deutschland verfügt über kein klares Indikatoren- und Zielsystem und kein Monitoring zur Ernährungsarmut. Die wenigen vorhandenen Daten weisen auf problematische Entwicklungen hin. Insgesamt wurden durch den Bund sehr umfangreiche Mittel zur Bewältigung der wirtschaftlichen und sozialen Folgen der Corona-Pandemie und zur Einkommenssicherung während der Pandemie eingesetzt. Dies ist unter dem Blickwinkel „Ernährungsarmut” positiv zu bewerten. Es ist davon auszugehen, dass diese Maßnahmen verhindert haben, dass noch mehr Haushalte von Ernährungsarmut betroffen waren. Insgesamt zeigt die Analyse jedoch, dass es kaum gezielte Maßnahmen gab, um den Ausfall von Kita- und Schulernährung durch alternative Formen der Bereitstellung von Lebensmitteln oder einer warmen Mahlzeit aufzufangen. Maßnahmen auf Bundesebene fehlten, um Länder, Kommunen und nicht-staatliche Organisationen dabei zu unterstützen. Auswertungen der Pandemiemaßnahmen vergleichbarer Länder zeigen, dass dort schnelle und flexible Maßnahmen zur Bereitstellung von Lebensmitteln eine wichtige Rolle gespielt haben, um armutsgefährdete Haushalte zu unterstützen. In Deutschland wurden vorwiegend finanzielle Transfers genutzt, z. T. zeitlich verzögert und wenig zielgruppenspezifisch. Vor diesem Hintergrund hat der WBAE mögliche Maßnahmen gegen Ernährungsarmut – in Krisen, aber auch generell – untersucht: (1) finanzielle Transfers (einmalig oder fortlaufend), (2) Lebensmittelgutscheine („Voucher“), (3) Lebensmittelpakete und (4) warme Mahlzeiten, wobei die letzten beiden Optionen jeweils zur Abholung oder zur Auslieferung angeboten werden können. Die untersuchten internationalen Beispiele zeigen Möglichkeiten, wie z. T. mit schnellen und teilweise innovativen Ansätzen auf Lockdown-Maßnahmen reagiert wurde. Im Grundsatz spricht aus Sicht des WBAE aber vieles für das deutsche System zur Bekämpfung der Ernährungsarmut, das im Kern auf einem Bürgergeld beruht. Ein solches Bürgergeld muss jedoch ausreichend sein, um materielle und soziale Ernährungsarmut zu vermeiden. Die aktuellen, im Bürgergeld für Essen und Getränke zur Verfügung stehenden Beträge entsprechen allerdings nicht diesem Anspruch. Bei der Einführung des Bürgergelds im Jahr 2023 wurde zwar zeitnah auf die hohe Inflation reagiert. Die Berechnungsmethodik für die Bedarfsermittlung wurde aber nicht angepasst, sodass der Regelsatz nach wie vor nicht für eine gesundheitsfördernde Ernährung ausreicht. Die soziale Funktion von Ernährung wird bei der Berechnung des Regelsatzes nicht berücksichtigt. Um bei akutem Handlungsbedarf, wie z. B. auch bei der derzeitigen Inflation, schneller reagieren zu können, sollten außerdem Systeme zur Auszahlung sozial differenzierter „Krisentransfers” an vulnerable Haushalte mit geringem Verwaltungsaufwand ermöglicht werden. Darüber hinaus sollte für den zentralen Bereich der Kinder-ernährung der Schutz einer adäquaten Grundsicherung durch einen Systemwechsel zu einer integrativen, beitragsfreien und somit nicht-diskriminierenden Kita- und Schulverpflegung ergänzt werden, die zudem den DGE-Qualitätsstandards entspricht. Schließlich sollte der Staat privates, ehrenamtliches Engagement ergänzend fördern, z. B. durch Infrastrukturunterstützung bei karitativen Essensangeboten. Eine Herausforderung der Gestaltung von karitativen Lebensmittelangeboten besteht darin, dass sie oft als stigmatisierend und beschämend wahrgenommen werden und der Nachweis eines Anspruches teils schwierig ist. Zudem bestehen oft nur wenige Wahlmöglichkeiten, sodass individuelle Präferenzen oder Unverträglichkeiten nur sehr eingeschränkt berücksichtigt werden können. Im Hinblick auf die soziale Dimension der Ernährung kommt hinzu, dass wichtige soziale Funktionen des Essens, wie das gemeinsame Essen, also Kommensalität, nicht oder nur sehr eingeschränkt ermöglicht wird. Auch bleiben diejenigen, die versorgt werden, dabei meist unter sich. Vor dem Hintergrund dieser Gesamteinschätzung empfiehlt der WBAE die folgenden, in Abbildung Z 1 dargestellten Maßnahmen (siehe Beitrag). Zusammenfassend betont der WBAE die in Deutschland weiterhin unterschätzten Risiken materieller und sozialer Ernährungsarmut. Im Rahmen der geplanten Nationalen Ernährungsstrategie der Bundesregierung sollte dem Thema Ernährungsarmut aus Sicht des WBAE eine herausgehobene Bedeutung zugemessen werden.
AU - Birner, Regina
AU - Linseisen, Jakob
AU - Arens-Azevedo, Ulrike
AU - Balmann, Alfons
AU - Biesalski, Hans Konrad
AU - Bosy-Westphal, Anja
AU - Buyken, Anette
AU - Döring, Thomas
AU - Feindt, Peter
AU - Götz, Kay-Uwe
AU - Iweala, Sarah
AU - Martínez, José
AU - Nieberg, Hiltrud
AU - Nöthlings, Ute
AU - Pischetsrieder, Monika
AU - Renner, Britta
AU - Spiller, Achim
AU - Voget-Kleschin, Lieske
AU - Weingarten, Peter
AU - Wesseler, Justus
AU - Wieck, Christine
ID - 46214
JF - Berichte über Landwirtschaft
TI - Ernährungsarmut unter Pandemiebedingungen
ER -
TY - JOUR
AU - Renner, Britta
AU - Buyken, Anette
AU - Gedrich, Kurt
AU - Lorkowski, Stefan
AU - Watzl, Bernhard
AU - Linseisen, Jakob
AU - Daniel, Hannelore
AU - Conrad, Johanna
AU - Ferrario, Paola G.
AU - Holzapfel, Christina
AU - Leitzmann, Michael
AU - Richter, Margrit
AU - Simon, Marie-Christine
AU - Sina, Christian
AU - Wirsam, Jan
ID - 46217
JF - Advances in Nutrition
KW - Nutrition and Dietetics
KW - Medicine (miscellaneous)
KW - Food Science
SN - 2161-8313
TI - Perspective: A Conceptual Framework for Adaptive Personalized Nutrition Advice Systems
ER -
TY - CONF
AU - Kaufhold, Marc-André
AU - Bayer, Markus
AU - Bäumler, Julian
AU - Reuter, Christian
AU - Stieglitz, Stefan
AU - Basyurt, Ali Sercan
AU - Mirbabaie, Milad
AU - Fuchß, Christoph
AU - Eyilmez, Kaan
ID - 46216
T2 - Mensch und Computer
TI - CYLENCE: Strategies and Tools for Cross-Media Reporting, Detection, and Treatment of Cyberbullying and Hatespeech in Law Enforcement Agencies
ER -
TY - GEN
AU - Lienen, Christian
AU - Nowosad, Alexander Philipp
AU - Platzner, Marco
ID - 46229
TI - Mapping and Optimizing Communication in ROS 2-based Applications on Configurable System-on-Chip Platforms
ER -
TY - GEN
AU - N., N.
ID - 46221
TI - Improving the End-of-Line Test of Custom-Built Geared Motors using Clustering based on Neural Networks
ER -
TY - JOUR
AU - Demir, Caglar
AU - Ngonga Ngomo, Axel-Cyrille
ID - 46243
JF - ECML-PKDD
TI - Clifford Embeddings – A Generalized Approach for Embedding in Normed Algebras
ER -
TY - JOUR
AU - Demir, Caglar
AU - Ngonga Ngomo, Axel-Cyrille
ID - 46251
JF - International Joint Conference on Artificial Intelligence
TI - Neuro-Symbolic Class Expression Learning
ER -
TY - JOUR
AU - Ma, Yulai
AU - Mattiolo, Davide
AU - Steffen, Eckhard
AU - Wolf, Isaak Hieronymus
ID - 46256
IS - 3
JF - SIAM Journal on Discrete Mathematics
KW - General Mathematics
SN - 0895-4801
TI - Pairwise Disjoint Perfect Matchings in r-Edge-Connected r-Regular Graphs
VL - 37
ER -
TY - CONF
AU - Meier, Jana
AU - Vogelsang, Christoph
AU - Küth, Simon
AU - Scholl , Daniel
AU - Watson, Christina
AU - Seifert, Andreas
ID - 46261
TI - Welche Rolle spielt eine reflexive Haltung für eine qualitätsvolle Unterrichtsreflexion? – Zusammenhänge zwischen einer quasi-experimentellen Einstellung und der Reflexionsperformanz von Lehramtsstudierenden [Einzelbeitrag].
ER -
TY - CONF
AU - Scholl, Daniel
AU - Küth, Simon
AU - Vogelsang, Christoph
AU - Meier, Jana
AU - Watson, Christina
AU - Seifert, Andreas
ID - 46262
TI - Das Unterrichtsplanungsprinzip der Interdependenz – Eine netzwerkanalytische Untersuchung der Begründungsstrukturen beim Planungsentscheiden [Einzelbeitrag].
ER -
TY - CONF
AU - Scholl, Daniel
AU - Vogelsang, Christoph
AU - Küth, Simon
AU - Meier, Jana
AU - Watson, Christina
AU - Seifert, Andreas
ID - 46263
TI - Eine reflexive Haltung als Grundlage einer hochwertigen Unterrichtsreflexion? Zusammenhänge zwischen einer quasi-experimentellen Einstellung zur Reflexion und der Reflexionsperformanz von Lehramtsstudierenden [Einzelbeitrag].
ER -
TY - GEN
ED - Meier, Jana
ED - Küth, Simon
ED - Scholl , Daniel
ED - Vogelsang, Christoph
ED - Watson, Christina
ID - 46260
TI - Der Zyklus von Planung und Reflexion. Zusammenhänge zwischen der generischen Unterrichtsplanungsfähigkeit und der Reflexionskompetenz angehender Lehrkräfte.
ER -
TY - CONF
AB - State-of-the-art LLC resonant converters use MOSFETs in their inverter stage, which allows high switching frequencies and thus the use of compact magnetic components. The large parasitic output capacitance and the poor reverse-recovery behaviour of the inherent body diode of high-voltage (600 V) silicon MOSFETs require soft switching, i.e. zero-voltage switching (ZVS). Otherwise, the high turn-on switching losses would lead to excessive heating and ultimately to the destruction of the switch. Therefore, MOSFET-based LLC converters are operated in the so-called inductive region only, which enables ZVS. The use of robust and cost-effective IGBTs instead of MOSFETs is particularly advantageous for automotive applications, since in addition to high reliability low costs are an important objective here. Since IGBTs are characterized by dominant turn-off losses and generally higher switching losses compared to MOSFETs, the aim is to operate them with zero-current switching (ZCS) and at low switching frequencies below the resonance frequency. In this region also the voltage transfer characteristic is steeper, which qualifies for applications with a strongly varying input-to-output voltage ratio, such as given for automotive on-board DC-DC converters connecting the (high-voltage) traction battery with the (12 V) auxiliary battery. In this paper, a stress value analysis based on a switched-model simulation is used to design a ZCS LLC converter and take advantage of the mentioned benefits of IGBTs as well as of the steeper voltage transfer characteristic. Within this operation region below the resonance frequency, however, a new phenomenon of several current pulses occurring during a single switching period through the rectifier components may appear. Generally, in applications with high output currents a synchronous rectifier (SR) is often used to keep the conduction losses of the rectifier stage at a moderate level: Low-voltage MOSFETs, which actively need to be gated synchronously to the polarity of the current pulses, are employed then instead of more lossy rectifier diodes. However, standard SR driver ICs have been shown to be unable to properly rectify the multi-pulse output currents of the proposed LLC operation, resulting in high conduction losses of the rectifier stage. A cost-effective hardware concept is presented which ensures proper rectification by using standard SR-ICs that are actively overdriven by the converter’s central microcontroller. A 2 kW prototype for an EV on-board DC-DC converter was built to show the effectiveness of the method, documenting an increase in efficiency by up to 4.1 % compared to a purely SR-IC-based solution. Overall efficiency is very similar to that of a conventional (MOSFET-based) LLC converter so that the ZCS-operated LLC IGBT-converter represents a cost-effective alternative, which even shows 10 % less worst-case losses.
AU - Urbaneck, Daniel
AU - Schafmeister, Frank
AU - Böcker, Joachim
ID - 46269
KW - LLC Converter
KW - IGBT
KW - ZCS
KW - Synchronous Rectification
T2 - PCIM Europe 2023
TI - Advanced Synchronous Rectification for an IGBT-Based ZCS LLC Converter with High Output Currents for a 2 kW Automotive DC-DC Stage
ER -
TY - CONF
AB - The computation of electron repulsion integrals (ERIs) over Gaussian-type orbitals (GTOs) is a challenging problem in quantum-mechanics-based atomistic simulations. In practical simulations, several trillions of ERIs may have to be
computed for every time step.
In this work, we investigate FPGAs as accelerators for the ERI computation. We use template parameters, here within the Intel oneAPI tool flow, to create customized designs for 256 different ERI quartet classes, based on their orbitals. To maximize data reuse, all intermediates are buffered in FPGA on-chip memory with customized layout. The pre-calculation of intermediates also helps to overcome data dependencies caused by multi-dimensional recurrence
relations. The involved loop structures are partially or even fully unrolled for high throughput of FPGA kernels. Furthermore, a lossy compression algorithm utilizing arbitrary bitwidth integers is integrated in the FPGA kernels. To our
best knowledge, this is the first work on ERI computation on FPGAs that supports more than just the single most basic quartet class. Also, the integration of ERI computation and compression it a novelty that is not even covered by CPU or GPU libraries so far.
Our evaluation shows that using 16-bit integer for the ERI compression, the fastest FPGA kernels exceed the performance of 10 GERIS ($10 \times 10^9$ ERIs per second) on one Intel Stratix 10 GX 2800 FPGA, with maximum absolute errors around $10^{-7}$ - $10^{-5}$ Hartree. The measured throughput can be accurately explained by a performance model. The FPGA kernels deployed on 2 FPGAs outperform similar computations using the widely used libint reference on a two-socket server with 40 Xeon Gold 6148 CPU cores of the same process technology by factors up to 6.0x and on a new two-socket server with 128 EPYC 7713 CPU cores by up to 1.9x.
AU - Wu, Xin
AU - Kenter, Tobias
AU - Schade, Robert
AU - Kühne, Thomas
AU - Plessl, Christian
ID - 43228
T2 - 2023 IEEE 31st Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM)
TI - Computing and Compressing Electron Repulsion Integrals on FPGAs
ER -
TY - JOUR
AB - The non-orthogonal local submatrix method applied to electronic structure–based molecular dynamics simulations is shown to exceed 1.1 EFLOP/s in FP16/FP32-mixed floating-point arithmetic when using 4400 NVIDIA A100 GPUs of the Perlmutter system. This is enabled by a modification of the original method that pushes the sustained fraction of the peak performance to about 80%. Example calculations are performed for SARS-CoV-2 spike proteins with up to 83 million atoms.
AU - Schade, Robert
AU - Kenter, Tobias
AU - Elgabarty, Hossam
AU - Lass, Michael
AU - Kühne, Thomas
AU - Plessl, Christian
ID - 45361
JF - The International Journal of High Performance Computing Applications
KW - Hardware and Architecture
KW - Theoretical Computer Science
KW - Software
SN - 1094-3420
TI - Breaking the exascale barrier for the electronic structure problem in ab-initio molecular dynamics
ER -
TY - JOUR
AB - Site-controlled Ga droplets on AlGaAs substrates are fabricated using area-selective deposition of Ga through apertures in a mask during molecular beam epitaxy (MBE). The Ga droplets can be crystallized into GaAs quantum dots using a crystallization step under As flux. In order to model the complex process, including the masked deposition of the droplets and a reduction of their number during a thermal annealing step, a multiscale kinetic Monte Carlo (mkMC) simulation of self-assembled Ga droplet formation on AlGaAs is expanded for area-selective deposition. The simulation has only two free model parameters: the activation energy for surface diffusion and the activation energy for thermal escape of adatoms from a droplet. Simulated droplet numbers within the opening of the aperture agree quantitatively with the experimental results down to the perfect site-control, with one droplet per aperture. However, the model parameters are different compared to those of the self-assembled droplet growth. We attribute this to the presence of the mask in close proximity to the surface, which modifies the local process temperature and the As background. This approach also explains the dependence of the model parameters on the size of the aperture.
AU - Feddersen, Stefan
AU - Zolatanosha, Viktoryia
AU - Alshaikh, Ahmed
AU - Reuter, Dirk
AU - Heyn, Christian
ID - 46278
IS - 3
JF - Nanomaterials
KW - General Materials Science
KW - General Chemical Engineering
SN - 2079-4991
TI - Modeling of Masked Droplet Deposition for Site-Controlled Ga Droplets
VL - 13
ER -
TY - JOUR
AU - Sieland, Benedikt
AU - Stahn, Marcel
AU - Schoch, Roland
AU - Daniliuc, Constantin
AU - Spicher, Sebastian
AU - Grimme, Stefan
AU - Hansen, Andreas
AU - Paradies, Jan
ID - 46277
JF - Angewandte Chemie International Edition
KW - General Chemistry
KW - Catalysis
SN - 1433-7851
TI - Dispersion Energy‐Stabilized Boron and Phosphorus Lewis Pairs
ER -
TY - CONF
AU - Clausing, Lennart
AU - Guetattfi, Zakarya
AU - Kaufmann, Paul
AU - Lienen, Christian
AU - Platzner, Marco
ID - 45913
T2 - Proceedings of the 19th International Symposium on Applied Reconfigurable Computing (ARC)
TI - On Guaranteeing Schedulability of Periodic Real-time Hardware Tasks under ReconOS64
ER -
TY - CHAP
AU - Tophinke, Doris
ED - Feilke, Helmuth
ID - 46099
T2 - Zwischen Fake und Fakt
TI - "Tankrabatt" und "Freiheitsenergie". Wie der Blog "neusprech.org" politische Euphemismen entlarvt
VL - 300
ER -
TY - JOUR
AU - Merten, Marie-Luis
AU - Wever, Marcel
AU - Tophinke, Doris
AU - Geierhos, Michaela
AU - Hüllermeier, Eyke
ID - 45673
JF - International Journal of Corpus Linguistics
TI - Annotation uncertainty in the context of grammatical change
ER -
TY - THES
AU - Tornede, Alexander
ID - 45780
TI - Advanced Algorithm Selection with Machine Learning: Handling Large Algorithm Sets, Learning From Censored Data, and Simplyfing Meta Level Decisions
ER -
TY - THES
AU - Spiess-Bru, Clarissa
ID - 46296
TI - Information Asymmetry and Gendered Behavior in Organizations and Digital Markets
ER -
TY - CONF
AB - A frequency-flexible Nyquist pulse synthesizer is presented with optical pulse bandwidths up to fopt=100 GHz and repetition rates equal to fopt/9, fabricated in an electronic-photonic co-integrated platform utilizing linear on-chip drivers.
AU - Kress, Christian
AU - Schwabe, Tobias
AU - Silberhorn, Christine
AU - Scheytt, J. Christoph
ID - 45578
T2 - Conference on Lasers and Electro-Optics (CLEO) 2023
TI - Generation of 100 GHz Periodic Nyquist Pulses using Cascaded Mach-Zehnder Modulators in a Silicon Electronic-Photonic Platform
ER -
TY - JOUR
AU - Homt, Martina
AU - Bloh, Bea
ID - 46354
IS - 1
JF - Lehrerbildung auf dem Prüfstand
TI - Der formelle Diskurs Forschenden Lernens in der Lehrerbildung. Ein Blick zurück nach vorn
VL - 16
ER -
TY - CHAP
AU - Ksouri-Gerwien, Christoph
AU - Vorbohle, Christian
ED - Schallmo, D.R.A.
ED - Kundisch, Dennis
ED - Lang, K.
ID - 37704
T2 - Digitale Plattformen und Ökosysteme im B2B-Bereich
TI - Anwendung von System Dynamics zur Geschäftsmodellinnovation in einem B2B-Ökosystem
ER -
TY - CONF
AB - Effect chain modelling is a method for creating information
models for impact analyses of changes in system elements. For
the estimation of change propagation, dependencies between
requirements must be detected. The high number of require-
ment dependencies in the engineering of complex technical
systems results in the need for automation. In a study, it was
shown that transformer models (BERT) are suitable for the
automated dependency analysis of requirements. However,
there are currently deficits in the applicability of the models
for different projects without an extensive and heterogeneous
training database. This paper investigates how active learning
can be used to train BERT models (active-BERT) in order to
increase the performance of the models for classifying requi-
rement dependencies of projects with heterogeneous require-
ments. The results show that the performance of the models
increases significantly through active learning. Through active-
BERT, engineers are enabled to model effect chains efficiently
and to handle requirement changes effectively.
AU - Gräßler, Iris
AU - Preuß, Daniel
ED - Hölzle, Katharina
ED - Kreimeyer, Matthias
ED - Roth, Daniel
ED - Maier, Thomas
ED - Riedel, Oliver
ID - 45661
SN - 2364-4885
T2 - Stuttgarter Symposium für Produktentwicklung SSP 2023
TI - Automatisierte Abhängigkeitsanalyse von Anforderungen zur Wirkkettenmodellierung
ER -
TY - BOOK
AU - Fuchs, Christian
ID - 42949
SN - 9781032362724
TI - Digital Democracy and the Digital Public Sphere
VL - 6
ER -
TY - GEN
AB - Effective photoinduced charge transfer makes molecular bimetallic assemblies attractive for applications as active light induced proton reduction systems. For a more sustainable future, development of competitive base metal dyads is mandatory. However, the electron transfer mechanisms from the photosensitizer to the proton reduction catalyst in base metal dyads remain so far unexplored. We study a Fe-Co dyad that exhibits photocatalytic H2 production activity using femtosecond X-ray emission spectroscopy, complemented by ultrafast optical spectroscopy and theoretical time-dependent DFT calculations, to understand the electronic and structural dynamics after photoexcitation and during the subsequent charge transfer process from the FeII photosensitizer to the cobaloxime catalyst. Using this novel approach, the simultaneous measurement of the transient Kalpha X-ray emission at the iron and cobalt K-edges in a two-colour experiment is enabled making it possible to correlate the excited state dynamics to the electron transfer processes. The methodology, therefore, provides a clear and direct spectroscopic evidence of the Fe->Co electron transfer responsible for the proton reduction activity.
AU - Nowakowski, Michał
AU - Huber-Gedert, Marina
AU - Elgabarty, Hossam
AU - Kubicki, Jacek
AU - Kertem, Ahmet
AU - Lindner, Natalia
AU - Khakhulin, Dimitry
AU - Lima, Frederico Alves
AU - Choi, Tae-Kyu
AU - Biednov, Mykola
AU - Piergies, Natalia
AU - Zalden, Peter
AU - Kubicek, Katerina
AU - Rodriguez-Fernandez, Angel
AU - Salem, Mohammad Alaraby
AU - Kühne, Thomas
AU - Gawelda, Wojciech
AU - Bauer, Matthias
ID - 40982
T2 - arxiv
TI - Ultrafast two-colour X-ray emission spectroscopy reveals excited state landscape in a base metal dyad
ER -
TY - CONF
AB - The article shows how to learn models of dynamical systems from data which are governed by an unknown variational PDE. Rather than employing reduction techniques, we learn a discrete field theory governed by a discrete Lagrangian density $L_d$ that is modelled as a neural network. Careful regularisation of the loss function for training $L_d$ is necessary to obtain a field theory that is suitable for numerical computations: we derive a regularisation term which optimises the solvability of the discrete Euler--Lagrange equations. Secondly, we develop a method to find solutions to machine learned discrete field theories which constitute travelling waves of the underlying continuous PDE.
AU - Offen, Christian
AU - Ober-Blöbaum, Sina
ED - Nielsen, F
ED - Barbaresco, F
ID - 42163
KW - System identification
KW - discrete Lagrangians
KW - travelling waves
T2 - Geometric Science of Information
TI - Learning discrete Lagrangians for variational PDEs from data and detection of travelling waves
VL - 14071
ER -
TY - JOUR
AB - The principle of least action is one of the most fundamental physical principle. It says that among all possible motions connecting two points in a phase space, the system will exhibit those motions which extremise an action functional. Many qualitative features of dynamical systems, such as the presence of conservation laws and energy balance equations, are related to the existence of an action functional. Incorporating variational structure into learning algorithms for dynamical systems is, therefore, crucial in order to make sure that the learned model shares important features with the exact physical system. In this paper we show how to incorporate variational principles into trajectory predictions of learned dynamical systems. The novelty of this work is that (1) our technique relies only on discrete position data of observed trajectories. Velocities or conjugate momenta do not need to be observed or approximated and no prior knowledge about the form of the variational principle is assumed. Instead, they are recovered using backward error analysis. (2) Moreover, our technique compensates discretisation errors when trajectories are computed from the learned system. This is important when moderate to large step-sizes are used and high accuracy is required. For this,
we introduce and rigorously analyse the concept of inverse modified Lagrangians by developing an inverse version of variational backward error analysis. (3) Finally, we introduce a method to perform system identification from position observations only, based on variational backward error analysis.
AU - Ober-Blöbaum, Sina
AU - Offen, Christian
ID - 29240
JF - Journal of Computational and Applied Mathematics
KW - Lagrangian learning
KW - variational backward error analysis
KW - modified Lagrangian
KW - variational integrators
KW - physics informed learning
SN - 0377-0427
TI - Variational Learning of Euler–Lagrange Dynamics from Data
VL - 421
ER -
TY - JOUR
AB - The numerical solution of an ordinary differential equation can be interpreted as the exact solution of a nearby modified equation. Investigating the behaviour of numerical solutions by analysing the modified equation is known as backward error analysis. If the original and modified equation share structural properties, then the exact and approximate solution share geometric features such as the existence of conserved quantities. Conjugate symplectic methods preserve a modified symplectic form and a modified Hamiltonian when applied to a Hamiltonian system. We show how a blended version of variational and symplectic techniques can be used to compute modified symplectic and Hamiltonian structures. In contrast to other approaches, our backward error analysis method does not rely on an ansatz but computes the structures systematically, provided that a variational formulation of the method is known. The technique is illustrated on the example of symmetric linear multistep methods with matrix coefficients.
AU - McLachlan, Robert
AU - Offen, Christian
ID - 29236
IS - 1
JF - Journal of Geometric Mechanics
KW - variational integrators
KW - backward error analysis
KW - Euler--Lagrange equations
KW - multistep methods
KW - conjugate symplectic methods
TI - Backward error analysis for conjugate symplectic methods
VL - 15
ER -
TY - JOUR
AB - Recently, Hamiltonian neural networks (HNN) have been introduced to incorporate prior physical knowledge when
learning the dynamical equations of Hamiltonian systems. Hereby, the symplectic system structure is preserved despite
the data-driven modeling approach. However, preserving symmetries requires additional attention. In this research, we
enhance the HNN with a Lie algebra framework to detect and embed symmetries in the neural network. This approach
allows to simultaneously learn the symmetry group action and the total energy of the system. As illustrating examples,
a pendulum on a cart and a two-body problem from astrodynamics are considered.
AU - Dierkes, Eva
AU - Offen, Christian
AU - Ober-Blöbaum, Sina
AU - Flaßkamp, Kathrin
ID - 37654
IS - 6
JF - Chaos
SN - 1054-1500
TI - Hamiltonian Neural Networks with Automatic Symmetry Detection
VL - 33
ER -
TY - CHAP
AU - Dehmel, Lukas
ED - Kondratjuk, Maria
ID - 46473
T2 - Grundlagen der Erwachsenen- und Weiterbildungsforschung
TI - 6.3.2 Medienpädagogische Professionalisierung in der beruflichen Weiterbildung. Eine Studie aus Perspektive der biografischen Medienforschung. Steckbrief zur Dissertation.
ER -
TY - JOUR
AU - Unterstell, Rembert
ID - 46471
IS - 1
JF - german research – Magazine of the Deutsche Forschungsgemeinschaft
TI - „Allowing the Economy to Breathe Even During the Crisis“ – Interview with Tax Expert Caren Sureth-Sloane
ER -