BEGIN:VCALENDAR VERSION:2.0 PRODID:Linklings LLC BEGIN:VTIMEZONE TZID:Europe/Stockholm X-LIC-LOCATION:Europe/Stockholm BEGIN:DAYLIGHT TZOFFSETFROM:+0100 TZOFFSETTO:+0200 TZNAME:CEST DTSTART:19700308T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 TZNAME:CET DTSTART:19701101T020000 RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20241120T082410Z LOCATION:HG F 30 Audi Max DTSTART;TZID=Europe/Stockholm:20240604T094000 DTEND;TZID=Europe/Stockholm:20240604T101000 UID:submissions.pasc-conference.org_PASC24_sess158@linklings.com SUMMARY:Flash Poster Session - Part II DESCRIPTION:Poster\n\nThe aim of this rapid-fire session is to allow poste r presenters to introduce the topic of their poster and motivate the audie nce to visit them at the poster session scheduled later in the evening. Pr esenters will have 30 seconds to engage the audience.\n\nP39 / ACMP11 - Pe rformance Regression Unit Testing for High Performance Computing Packages in Julia\n\nThis research focuses on the integration of performance testin g into the unit testing phase for High-Performance Computing (HPC) softwar e, emphasizing its importance in ensuring optimal implementations and diag nosing performance regressions. In traditional unit testing, functional as pects are assess...\n\n\nDaniel Sergio Vega Rodríguez (Università della Sv izzera italiana), Samuel Omlin (ETH Zurich / CSCS), and Juraj Kardoš and O laf Schenk (Università della Svizzera italiana)\n---------------------\nP3 6 / ACMP08 - Mixed-Precision in High-Order Methods: Studying the Impact of Lower Numerical Precisions on the ADER-DG Algorithm\n\nWe study the impac t of using mixed and variable numerical precision in the high-order ADER-D G method for solving partial differential equations. The impact of precisi on on both the overall convergence order, as well as specific sections of the code are examined. This lets us judge how sensible each ...\n\n\nMarc Marot-Lassauzaie and Michael Bader (Technical University of Munich)\n----- ----------------\nP25 - GPU Benchmarking on Fully Occupied Accelerated Clu ster Nodes via Molecular Dynamics Software Packages\n\nNowadays, the usage of GPU accelerators in scientific computing is rapidly increasing. In var ious fields such as molecular and astrophysics simulations, geophysics, an d artificial intelligence models, modern graphic cards show significant im provement in computational speed and energy efficiency. Mul...\n\n\nPlamen Dobrev, Ivan Pribec, and Gerald Mathias (Leibniz Supercomputing Centre)\n ---------------------\nP20 - Fast Inference of Cosmology from High Resolut ion Maps Using Deep Learning\n\nOngoing galaxy surveys like the Dark Energ y Survey are designed to observe the large-scale structure of the Universe using a number of cosmological probes, such as weak gravitational lensing and galaxy clustering. Conventionally, constraints on the cosmological pa rameters are calculated by comparing...\n\n\nArne Thomsen (ETH Zurich); To masz Kacprzak (ETH Zurich, Swiss data science center); Peter Harrington (N ational Energy Research Scientific Computing Center); Agnes Ferte (SLAC); and Alexandre Refregier (ETH Zurich)\n---------------------\nP53 - Workflo w Automation for Verified High-Performance Molecular-Continuum Flow Simula tions\n\nThe Macro-Micro-Coupling Tool (MaMiCo) is an advanced CFD simulat ion framework enabling scientists to seamlessly integrate continuum mechan ics and molecular dynamics simulations for comprehensive fluid behavior an alyses on multiple scales. Due to the labor-intensive and error-prone proc ess of manual...\n\n\nJohannes Michaelis and Olaf Schenk (Università della Svizzera italiana)\n---------------------\nACMP06 - Finding Optimistic Up per Bounds for Task Graph Throughput on Heterogeneous Systems Using Linear Programming\n\nIn this extended abstract, we present a model -- inspired by previous work in the data flow community -- for finding optimistic uppe r bounds on the throughput of task graphs executed on heterogeneous system s. This model interprets the execution of such graphs as flow networks wit h additional resourc...\n\n\nStephen Nicholas Swatman (University of Amste rdam, CERN) and Ana-Lucia Varbanescu (University of Twente)\n------------- --------\nP24 - Fully Spectral Dynamo Simulations for Heterogeneous Comput ing\n\nOur CFD framework QuICC, based on a fully spectral method, has been successfully used for various dynamo simulations in spherical and Cartesi an geometries.\nIt runs efficiently on a few thousands of cores using a 2D data distribution based on a distributed memory paradigm (MPI).\nIn order to better ha...\n\n\nGiacomo Gastiglioni, Philippe Marti, and Dmitrii Tol machev (ETH Zurich); Daniel Ganellari (ETH Zurich / CSCS); and Andrew Jack son (ETH Zurich)\n---------------------\nP27 - A GT4Py-Based Multi-Node St andalone Python Implementation of the ICON Dynamical Core\n\nWe introduce a prototype atmospheric model, based on Icosahedral Non-hydrostatic (ICON) , which illustrates that indeed entire models can be ported to Python whil e still attaining performance portability. We reuse many of the stencils – written with a Python-based domain-specific language GT4Py...\n\n\nMagdal ena Luz, Abishek Gopal, and Chia Rui Ong (ETH Zurich); Christoph Müller, D aniel Hupp, and Nina Burgdorfer (MeteoSwiss); Nicoletta Farabullini (ETH Z urich); Fabian Bösch (ETH Zurich / CSCS); Anurag Dipankar (ETH Zurich); Ma uro Bianco and William Sawyer (ETH Zurich / CSCS); Samuel Kellerhals and J onas Jucker (ETH Zurich); Till Ehrengruber, Enrique González Paredes, and Hannes Vogt (ETH Zurich / CSCS); Peter Kardos (ETH Zurich); Rico Häuselman n (ETH Zurich / CSCS); and Xavier Lapillonne (MeteoSwiss)\n--------------- ------\nP23 - FraNetG – Fracture Network Growth\n\nThe phase-field method has emerged as a sophisticated technique for simulating crack initiation, propagation, and coalescence. This approach employs a damage field, termed the phase field, to represent the material's state from intact to fully f ractured. The phase-field approach is well known to yie...\n\n\nAlena Kopa nicakova (Brown University, Università della Svizzera italiana); Edoardo P ezzulli (ETH Zurich); Patrick Zulian (Università della Svizzera italiana, UniDistance Suisse); Hardik Kothari, Toby Simpson, and Maria Nestola (Univ ersità della Svizzera italiana); Thomas Driesner (ETH Zurich); and Rolf Kr ause (Università della Svizzera italiana, UniDistance Suisse)\n----------- ----------\nP34 - Ionbeam: Scalable IoT Streaming Infrastructure for Meteo rology\n\nThe European Centre for Medium-Range Weather Forecasts’ (ECMWF) relies on extensive meteorological observations, sourced from ground-based stations, aircraft, and satellites. Low-cost Internet-of-Things (IoT) dev ices present an opportunity to access observations at higher frequency, hi gher spa...\n\n\nThomas Hodson, Ulrike Falk, and Simon Smart (ECMWF)\n---- -----------------\nACMP10 - Optimized Finite Volume Methods Solver Allows for Real-Sized Tumor Simulations\n\nMulti-scale agent-based cell simulator s sets daunting computational challenges in bioinformatics, only feasible by supercomputing resources. These simulators consider evolving microenvir onmental conditions and cell interactions. By specifying rules at the cell level, researchers can explore complex ...\n\n\nJose Luis Estragués Muñoz , Carlos Álvarez, and Daniel Jimenez (Barcelona Supercomputing Center); Al fonso Valencia (Barcelona Supercomputing Center, ICREA); and Arnau Montagu d (Barcelona Supercomputing Center)\n---------------------\nP30 - Implemen tation and Benchmarking of a New Radiation Module in the WarpX Particle-In -Cell Code\n\nThe interaction of ultra-intense femtosecond lasers with pla smas is of interest for a variety of applications, including the accelerat ion of ultra-short, highly energetic electron bunches and the realization of compact secondary radiation sources. In all these scenarios, radiative processes are eith...\n\n\nLuca Fedeli, Thomas Clark, and Pierre Bartoli ( CEA); Axel Huebl, Rémi Lehe, and Jean-Luc Vay (Lawrence Berkeley National Laboratory); and Henri Vincenti (CEA)\n---------------------\nACMP04 - Eff icient Compression for Weather and Climate Data\n\nWeather and climate sim ulations produce petabytes of high-resolution data that are later analyzed by researchers to understand climate change or severe weather. We propose an efficient data compression method dedicated to weather and climate dat a. It consists of two components: one base compressor u...\n\n\nLangwen Hu ang and Torsten Hoefler (ETH Zurich)\n---------------------\nP46 - Scaling Laws for Machine-Learned Reconstruction\n\nMachine Learning (ML) methods have been successfully applied to various High Energy Physics (HEP) proble ms, such as particle identification, event reconstruction, jet tagging, an d anomaly detection. However, the relationship between the model size, i.e ., the number of model parameters, and the physi...\n\n\nEric Wulff (CERN) , Joosep Pata (National Institute of Chemical Physics and Biophysics), and Maria Girone (CERN)\n---------------------\nP32 - InterTwin - An Interdis ciplinary Digital Twin Engine for Science\n\nThe interTwin project, funded by the European Commission, is at the forefront of leveraging 'Digital Tw ins' across various scientific domains, with a particular emphasis on eart h observation and physics. This initiative encompasses core modules design ed to address the intricacies of data-driven and ...\n\n\nMatteo Bunino, K alliopi Tsolaki, Alexander Zoechbauer, Maria Girone, and Sofia Vallecorsa (CERN)\n---------------------\nP47 - Sculpting Precision: Unveiling the Im pact of eXplainable Features and Magnitudes in Neural Network Pruning\n\nI n the domain of Machine Learning (ML), models are celebrated for their hig h accuracy, however, integrating them into resource-constrained embedded s ystems poses a formidable challenge. This study empirically demonstrates t hat traditional magnitude-based pruning techniques, though effective in co mpr...\n\n\nJamil Gafur (The University of Iowa, National Renewable Energy Laboratory) and Steve Goddard (The University of Iowa)\n----------------- ----\nP29 - ICON-HAM: Modelling Aerosol-Cloud Interactions at High Resolut ion on GPUs\n\nAtmospheric aerosols are a key component to understand the Earth’s climate, as they have a strong influence on clouds, which in turn strongly impact the global radiative budget. The ICON-HAM model couples IC ON (Zängl et al., 2015) to the aerosol module HAM (Tegen et al., 2019). Th e aerosols ar...\n\n\nMikael Stellio (ETH Zurich, MeteoSwiss); Sylvaine Fe rrachat (ETH Zurich); Xavier Lapillonne (MeteoSwiss); and Ulrike Lohmann ( ETH Zurich)\n---------------------\nP37 - Parallel Implementation of Mesh- Free Operators for 2D and 3D PDEs on a Sphere for Atmospheric Dynamics\n\n This project explores a mesh-free method for the approximations of Numeric al Operators for 2D and 3D partial differential equations used in atmosphe ric dynamics. Compactly Supported Radial Basis Functions were chosen as th e category of Mesh Free method for discretization. The primary objective o f th...\n\n\nLakshmi Aparna Devulapalli (Università della Svizzera italian a) and William Barton Sawyer (ETH Zurich / CSCS)\n---------------------\nP 52 - Waveform Relaxation for Atmosphere-Ocean-Sea Ice Coupling in the EC-E arth Single Column Model\n\nEarth system models and general circulation mo dels couple many submodels in time and space. As such, numerical errors ar e introduced at the geometrical interfaces between components. The magnitu de of the numerical error in time can be estimated using iterative couplin g algorithms, so-called Schwarz ...\n\n\nValentina Schüller and Philipp Bi rken (Lund University); Eric Blayo (Université Grenoble Alpes); and Floria n Lemarié (INRIA, Université Grenoble Alpes)\n---------------------\nACMP0 2 - Demographic Aware Hyperparameter Optimization for Cancer\n\nThe prelim inary work presented here evaluates the robustness and bias of ML models b y probing HPO behavior under different demographic conditions, which is cr itical for the development of clinically-usable methods. By examining the different hyperparameter distributions for a transformer based model...\n\ n\nRylie Weaver (Argonne National Laboratory, Claremont Graduate Universit y)\n---------------------\nP42 - A Python Dynamical Core for Operational N umerical Weather Prediction\n\nNumerical weather prediction is vital for a pplications like population warnings and energy predictions. However, adap ting forecasts to diverse hardware poses challenges. MeteoSwiss relies on the ICON model up to a one km resolution, initially ported to GPUs using O penACC. While enabling GPU use, Ope...\n\n\nChristoph Müller, Daniel Hupp, and Nina Burgdorfer (MeteoSwiss); Abishek Gopal and Nicoletta Farabullini (Center for Climate Systems Modeling (C2SM)); Till Ehrengruber (ETH Zuric h / CSCS); Samuel Kellerhals and Magdalena Luz (Center for Climate Systems Modeling (C2SM)); William Sawyer (ETH Zurich / CSCS); Matthias Röthlin (M eteoSwiss); Enrique G. Paredes (ETH Zurich / CSCS); Benjamin Weber (MeteoS wiss); Hannes Vogt and Mauro Bianco (ETH Zurich / CSCS); Carlos Osuna (Met eoSwiss); Christina Schnadt and Anurag Dipankar (Center for Climate System s Modeling (C2SM)); and Xavier Lapillonne (MeteoSwiss)\n------------------ ---\nP40 - A Performance-Portable All-Scale Atmospheric Model Framework\n\ nWe provide an overview of activities and results in the development of a performance-portable atmospheric model for research applications and numer ical weather prediction. The model framework considers a full Python imple mentation with the GT4Py (GridTools for Python) domain-specific library en compa...\n\n\nNicolai Krieger (ETH Zurich), Christian Kühnlein (ECMWF), St efano Ubbiali (ETH Zurich), Till Ehrengruber (ETH Zurich / CSCS), Lukas Pa pritz and Sara Faghih-Naini (ECMWF), Gabriel Vollenweider (ETH Zurich), Lo ïc Maurin (Meteo-France), and Heini Wernli (ETH Zurich)\n----------------- ----\nP28 - GT4Py: A Python Framework for the Development of High-Performa nce Weather and Climate Applications\n\nGT4Py is a Python framework for we ather and climate applications simplifying the development and maintenance of high-performance codes in prototyping and production environments. GT4 Py separates model development from hardware architecture dependent optimi zations, instead of intermixing both togethe...\n\n\nMauro Bianco and Till Ehrengruber (ETH Zurich / CSCS); Nina Burgdorfer (MeteoSwiss); Nicoletta Farabullini, Abishek Gopal, Samuel Kellerhans, and Peter Kardos (ETH Zuric h); and Enrique Paredes, Rico Häuselmann, Felix Thaler, Hannes Vogt, Phili p Müller, and Christos Kotsalos (ETH Zurich / CSCS)\n--------------------- \nP35 - Machine Learning Emulator of the Radiation Solver in the ICON Clim ate Model\n\nThe computationally demanding radiative transfer parameteriza tion is a prime candidate for machine learning (ML) emulation.\nIn this pr oject, we develop an ML-based radiative parameterization.\nA random forest (RF) is used as a baseline method, with the European Centre for Medium-Ra nge Weather Forecas...\n\n\nGuillaume Bertoli (ETH Zurich)\n-------------- -------\nP21 - Fast Simulations of Next-Generation Radio Cosmological Surv eys: A Forward-Modeling Pipeline of Neutral Hydrogen Maps for SKA and HIRA X\n\nIn the last century, astronomical breakthroughs, particularly on Dark Matter and Dark Energy, have reshaped our understanding of the Universe. Despite comprising 95% of the Universe, these enigmatic components remain mysterious. Upcoming radio astronomical surveys, such as SKA and HIRAX, wi ll promis...\n\n\nLuis Fernando Machado Poletti Valle (ETH Zurich)\n------ ---------------\nP44 / ACMP03 - Scalable Simulations of Resistive Memory D evices: A Dynamical Monte Carlo Approach\n\nResistive random access memori es (ReRAM) are expected to play a prominent role in modern computer archit ectures due to their low cost, simple structure, and unique functionality. The long-range atomic movements inside these devices, which occur over ex tended timescales under applied fields, can be a...\n\n\nAlexander Maeder, Manasa Kaniselvan, Marko Mladenović, Mathieu Luisier, and Alexandros Niko laos Ziogas (ETH Zurich)\n---------------------\nP38 - Performance Charact erisation of Software for Lattice Quantum Field Theory Beyond the Standard Model\n\nLattice Quantum Chromodynamics (QCD) is a computationally demand ing field that has driven many innovations in the High-Performance Computi ng space. Beyond the Standard Model (BSM) physics introduces additional de grees of freedom that significantly increase the complexity of software an d the difficul...\n\n\nEd Bennett (Swansea University); Luigi Del Debbio a nd Ryan Hill (University of Edinburgh); Jong-Wan Lee (Institute for Basic Sciences); Julian Lenz, Biagio Lucini, and Maurizio Piai (Swansea Universi ty); Andrew Sunderland (Science and Technology Facilities Council); and Da vide Vadacchino (University of Plymouth)\n---------------------\nP41 - Pro babilistic Weather Forecasting through Latent Space Perturbations of Machi ne Learning Emulators\n\nThe intrinsic variability of the atmospheric syst em is historically reproduced by ensembles of forecasts based on numerical weather prediction. However, the computational cost of running such ensem bles based on perturbed initial conditions is prohibitive. Recent advances in machine learning (ML)-bas...\n\n\nSimon Adamov (ETH Zurich, MeteoSwiss ); Sebastian Schemm (ETH Zurich); Oliver Fuhrer (MeteoSwiss); and Reto Knu tti (ETH Zurich)\n---------------------\nP43 - Quo Vadis: Helping Applicat ions Manage On-Node Resources on Modern Systems\n\nScientific discovery is increasingly enabled by heterogeneous\nhardware that includes multiple pr ocessor types, deep memory\nhierarchies, and heterogeneous memories. To ef fectively utilize this\nhardware, computational scientists must compose th eir applications\nusing a combination of programming models...\n\n\nEdgar A. Leon (Lawrence Livermore National Laboratory) and Samuel K. Gutierrez ( Los Alamos National Laboratory)\n---------------------\nP22 - FFT-Accelera ted Polynomial Transforms for Fully Spectral Simulations\n\nOne of the mos t time-consuming parts of our CFD framework QuICC is the computation of th e physical to spectral space transformations. In spherical geometry, this transformation can be decomposed into three main parts: Fourier Transform and Spherical Harmonics Transform for angular parts and Jones-Wo...\n\n\nD mitrii Tolmachev, Philippe Marti, and Giacomo Castiglioni (ETH Zurich); Da niel Ganellari (ETH Zurich / CSCS); and Andrew Jackson (ETH Zurich)\n----- ----------------\nP49 - The Task-Based GPU-Enabled Distributed Eigensolver available in DLA-Future\n\nDLA-Future implements an efficient GPU-enabled distributed eigenvalue solver using asynchronous methods based on the C++ std::execution API. Using a task-based approach reduces the number of syn chronization points and allows for simple overlapping of communication and computation which helps improve...\n\n\nJohn Biddiscombe, Alberto Inverni zzi, Rocco Meli, Auriane Reverdell, Mikael Simberg, and Raffaele Solcà (ET H Zurich / CSCS)\n---------------------\nP33 - Inviscid Dynamo Simulation Using QuICC\n\nEarth’s magnetic field is believed to be generated in the m etallic outer core through a process known as the geodynamo. Direct numeri cal simulation (DNS) of the geodynamo has successfully reproduced many fea tures of the Earth’s field. However, even the state-of-the-art simulations have a...\n\n\nLonghui Yuan, Andrew Jackson, Philippe Marti, and Jiawen L uo (ETH Zurich)\n---------------------\nP31 - Improving Chest X-ray Image Classification via Parallelized Generative Neural Architecture Search\n\nE xplore GenNAS for chest X-ray classification in lung diseases, leveraging novel parallel training methods for enhanced accuracy and efficiency. Medi cal image classification for pulmonary pathologies from chest X-rays is tr aditionally time-consuming. GenNAS, using GPT-4's generative capabilities, au...\n\n\nFelix Mejia (Industrial University of Santander), John Anderso n Garcia Henao (University of Bern), Carlos Barrio (Industrial University of Santander), and Michell Riveill (Université Côte d’Azur)\n------------- --------\nACMP09 - On-Line Tracking of Tropical Cyclones in a Climate Mode l on a Given Radius Around its Path\n\nWe discuss the implementation of tw o added functionalities to the ICON weather and climate model: time depend ent output regions for variables stored on grid cell centers and an on-lin e version of an existing off-line tracking algorithm for tropical cyclones (TCs).\n\n\nThibault Meier (ETH Zurich)\n---------------------\nP50 - Tow ards Linear-Scaling Density Functional Theory on Real Space Grids\n\nDesig n structures of semiconductor circuits have shrunk to the lengthscale of a few nanometers. Despite that, systems so far have been too large to predi ct the electronic structure of realistic nano devices with an atom model d escription as accurate as density functional theory (DFT). DFT eigenvalue. ..\n\n\nPaul F. Baumeister and Shigeru Tsukamoto (Forschungszentrum Jülich )\n---------------------\nP51 - Tuning Atmospheric Turbulence Parameters w ith Machine Learning Surrogates\n\nParameterizations of subgrid-scale (SGS ) processes, like cloud microphysics, radiation, or turbulence, cause cons iderable uncertainty in numerical climate and weather models at various s patiotemporal scales. Tuning the involved model parameters is challenging, given the immense computational cost o...\n\n\nDana Grund, Sebastian Sche mm, and Siddhartha Mishra (ETH Zurich) and Oliver Fuhrer (MeteoSwiss)\n--- ------------------\nACMP05 - Efficient, Portable, Massively Parallel Free- Space Solvers for the Poisson Equation\n\nVico et al. (2016) suggest a fas t algorithm for computing volume potentials which is of benefit to the bea m and plasma physics communities, as they require the solution of Poisson’ s equation with free-space boundary conditions. The standard method to sol ve the free-space Poisson equation is to...\n\n\nSonali Mayani (Paul Scher rer Institute, ETH Zurich); Antoine Cerfon (New York University, Type One Energy); Matthias Frey (University of St Andrews); Veronica Montanaro (ETH Zurich); Sriramkrishnan Muralikrishnan (Forschungszentrum Jülich); and An dreas Adelmann (Paul Scherrer Institute, ETH Zurich)\n-------------------- -\nACMP07 - High Performance Computing Derived Biological Multiplex Networ k Uncovers Distinct Pathways Underlying Opioid and Nicotine Addiction\n\nL everaging High-Performance Computing (HPC) for biological network generati on, key insights into the genetic and epigenetic mechanisms supporting opi oid and nicotine addition have been uncovered. Using distributed network g eneration software on the Frontier supercomputer, the authors processed 70 0 s...\n\n\nMatthew Lane (University of Tennessee, Oak Ridge National Labo ratory)\n---------------------\nP26 - GPU-Accelerated Linear-Response for DFT+Hubbard Using the SIRIUS Library\n\nElectronic-structure methods have been indispensable in materials science, especially on the study of existi ng and the discovery of novel materials. Linear-response (LR) algorithms, a computationally intensive step compared to the self-consistent cycle, ar e widely present in electronic-structure cod...\n\n\nGiannis D. Savva (EPF L), Iurii Timrov and Nicola Colonna (Paul Scherrer Institute), Anton Kozhe vnikov (ETH Zurich / CSCS), and Nicola Marzari (EPFL)\n------------------- --\nP48 - Simulations of Giant Impacts: The Importance of High Resolution\ n\nGiant impacts (GI) form the last stage of planet formation and play a k ey role in determining many aspects like the final structure of planetary systems and the masses and compositions of its constituents. A common choi ce for numerically solving the equations of motion is the Smoothed Particl e Hydro...\n\n\nThomas Meier (University of Zurich); Christian Reinhardt ( University of Zurich, University of Bern); and Douglas Potter and Joachim Stadel (University of Zurich)\n\nSession Chair: Iva Kavcic (Met Office) END:VEVENT END:VCALENDAR