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:20260731T133323Z
LOCATION:Bldg. 8 - Entrance Hall
DTSTART;TZID=Europe/Stockholm:20260630T173000
DTEND;TZID=Europe/Stockholm:20260630T194500
UID:submissions.pasc-conference.org_PASC26_sess135@linklings.com
SUMMARY:Poster Session and Reception
DESCRIPTION:P01 - Accelerating Lattice QCD Dirac GCR Solvers with Multiple
  Right-Hand Sides (MRHS)\n\nLattice QCD simulations are often limited by t
 he memory-bandwidth bottlenecks of solving the Dirac equation for numerous
  source vectors. We present an optimised Multiple Right-Hand Side (MRHS) G
 eneralised Conjugate Residual (GCR) solver in the openQxD framework that a
 ddresses these limitations. By t...\n\n\nJingJing Li and Roman Gruber (Uni
 versity of Bern) and Marina Krstic Marinkovic (ETH Zurich)\n--------------
 -------\nP02 - Advancing The Data Assimilation Research Testbed (DART) as 
 an Early-Career Software Engineer\n\nThe Data Assimilation Research Testbe
 d (DART) is an open-source software facility for ensemble data assimilatio
 n that combines information from numerical model predictions with measurem
 ents of the Earth system to enhance the value of both. It has supported a 
 diverse community of users for over 20 ye...\n\n\nMarlena Smith (NSF Natio
 nal Center for Atmospheric Research)\n---------------------\nP03 - Algebra
 ic Multi-Level Methods for Lattice Dirac Operators in LQCD\n\nThe main com
 putational challenge in Lattice QCD is the efficient and scalable approxim
 ate solution of the Dirac equation Dz = b, where D denotes the Dirac matri
 x on a four-dimensional space-time lattice. Modern solvers for this case a
 re based on Adaptive Multigrid. Among them, Domain Decomposition A...\n\n\
 nPauline Schauerte (University of Bonn, Fraunhofer SCAI) and Jaime Fabian 
 Nieto Castellanos (Forschungszentrum Jülich, University of Bonn)\n--------
 -------------\nP04 - Algorithms and Optimizations for Global Non-Linear Hy
 brid Fluid-Kinetic Finite Element Stellarator Simulations\n\nPredictive mo
 deling of stellarator plasmas is crucial for advancing nuclear fusion ener
 gy, yet it faces unique computational difficulties. A primary challenge is
  accurately simulating the dynamics of specific particle species not well 
 captured by fluid models, necessitating the use of hybrid fluid-k...\n\n\n
 Luca Venerando Greco and Matthias Hoelzl (Max Planck Institute for Plasma 
 Physics); Guido Huijsmans (CEA, IRFM); and Edoardo Carrà (Max Planck Insti
 tute for Plasma Physics)\n---------------------\nP05 - Bridging Python Fle
 xibility and GPU Performance with Aithon: Kernel-Level Optimization, Scali
 ng, and Extreme-Resolution MHD Turbulence Simulations\n\nWe present Aithon
 , a GPU-accelerated incompressible flow solver for hydrodynamics and magne
 tohydrodynamics, designed for extreme-scale supercomputing. Optimized for 
 AMD MI250X GPUs and deployed on the Frontier system, Aithon combines kerne
 l-level GPU optimizations, CUDA/HIP-aware MPI, and Python int...\n\n\nMant
 han Verma (Indian Institute of Technology kanpur) and Gina Sitaraman and P
 aul Mullowney (AMD)\n---------------------\nP06 - Co-designing Regional Hi
 gh-Performance Computing Ecosystems in Africa: A Pilot Focus on Kenya and 
 West Africa\n\nHigh-performance computing is becoming more important for b
 ioinformatics, genomics, and public health research. However, in Africa, i
 ts growth and use remain uneven, scattered, and poorly documented. This st
 udy examines current HPC capacity and access models in West, East, and Sou
 thern Africa, drawi...\n\n\nPauline Karega (University of Manchester, Bioi
 nformatics Hub of Kenya initiative)\n---------------------\nP07 - Correlat
 ed Electrons on Accelerated Architectures from Frequency-Dependent Respons
 e Functions\n\nUnderstanding, characterizing and engineering spectral prop
 erties of correlated materials is crucial for next-generation technologies
 , including energy harvesting and quantum technologies. These properties e
 ncode a material's response to external stimuli, and while important in ge
 neral, they are eve...\n\n\nPaolo Settembri and Nicola Colonna (Paul Scher
 rer Institute); Anton Kozhevnikov (ETH Zurich / CSCS); and Nicola Marzari 
 (EPFL, Paul Scherrer Institute)\n---------------------\nP08 - Coupling km-
 Scale Earth System Model to Hierarchical Output for Analysis-Ready Dataset
 \n\nKilometer-scale Earth System Model (ESM) simulations produce petabyte-
 scale outputs that are difficult to access, analyse, and share due to thei
 r size, heterogeneity, and the overhead of ad-hoc workflows.\nWe introduce
  **Hiopy** (Hierarchical Output in Python), a lightweight in-situ output c
 omponent,...\n\n\nNils-Arne Dreier (DKRZ) and Siddhant Tibrewal (Max Planc
 k Institute for Meteorology)\n---------------------\nP09 - Developing and 
 Evaluating Performance-Portable Physical Parametrization Codes\n\nWe prese
 nt results and ongoing work in the porting of physical parametrizations to
  Python using the GridTools for Python (GT4Py) library. Our basis is the F
 ortran code from the Integrated Forecasting System (IFS) which is run oper
 ationally at the European Centre for Medium-Range Weather Forecasts (E...\
 n\n\nGabriel Vollenweider and Stefano Ubbiali (ETH Zurich), Christian Kühn
 lein (ECMWF), and Heini Wernli (ETH Zurich)\n---------------------\nP10 - 
 Discretization Error Quantification in Plane-Wave Density Functional Theor
 y\n\nDensity functional theory (DFT) has become a workhorse of computation
 al materials science. DFT computations in materials typically use a plane 
 wave basis set, truncated at a so-called kinetic energy cutoff Ecut. Estim
 ates for the truncation error of the basis set open opportunities for erro
 r balanci...\n\n\nBruno Ploumhans and Michael Herbst (EPFL)\n-------------
 --------\nP11 - Estimation of Global Surface Carbon Fluxes at the Grid Sca
 le Using Machine Learning Techniques\n\nMachine learning (ML) techniques h
 ave recently been applied in the field of geoscience as in other fields, a
 nd has shown significant progress. One of the major advantages of ML is it
 s remarkable effectiveness in overcoming the problem of realistic computat
 ional costs from a computational science per...\n\n\nJi-Sun Kang (Korea In
 stitute of Science and Technology Information)\n---------------------\nP12
  - Evaluating Open-Source Infrastructure-As-Code Virtual Clusters against 
 SuperMUC-NG Phase 1\n\nTraditional high-performance computing (tHPC) infra
 structure requires weeks to months for hardware procurement, network confi
 guration and software integration, which limits agility for short-term pro
 jects and hampers reproducibility through non-standardized configurations.
  Infrastructure-as-Code (Ia...\n\n\nPrasanth Babu Ganta, Elmira Birang, Pl
 amen Dobrev, Birkan Emrem, Matteo Foglieni, and Ferdinand Jamitzky (Leibni
 z Supercomputing Centre)\n---------------------\nP13 - Exploring Performan
 ce and Efficiency of State-of-the-Art Deep Learning Protein Structure Pred
 iction Frameworks on the Frontier Exascale Supercomputer\n\nAccurately pre
 dicting the structure of a protein has been a long standing and extremely 
 challenging problem in biology. In recent years, the rapid evolution and a
 doption of artificial intelligence have made the prediction of protein str
 uctures leveraging deep learning frameworks with accuracy rivali...\n\n\nV
 erónica G. Melesse Vergara, Elijah MacCarthy, Asim YarKhan, John Holmen, M
 anesh Shah, Érica Teixeira Prates, and Dan Jacobson (Oak Ridge National La
 boratory)\n---------------------\nP14 - A Flexible Interface for Neural Ne
 twork Potentials in GROMACS\n\nWe present a new interface for hybrid machi
 ne learning/molecular mechanics (ML/MM) simulations implemented in the mol
 ecular dynamics engine GROMACS. The interface enables neural network poten
 tials (NNPs) trained in the PyTorch framework to contribute energies and f
 orces during molecular dynamics (MD...\n\n\nLukas Müllender and Berk Hess 
 (KTH Royal Institute of Technology) and Erik Lindahl (KTH Royal Institute 
 of Technology, Stockholm University)\n---------------------\nP15 - A Flux-
 Form Semi-Lagrangian WENO Scheme on Triangular Meshes\n\nThe icosahedral m
 odel for weather and climate simulations utilises flux-form semi-Lagrangia
 n (FFSL) schemes for the transport of species. The motivation is the highe
 r Courant-Friedrich-Lewy (CFL) number compared to Eulerian approaches. The
  schemes are implemented on the triangular mesh on a sphere w...\n\n\nAndr
 eas Jocksch (ETH Zurich / CSCS); Daniel Reinert (Deutscher Wetterdienst (D
 WD)); Christoph Müller (MeteoSwiss); David Strassmann (ETH Zurich); Nina B
 urgdorfer (MeteoSwiss); Anurag Dipankar (ETH Zurich); Mauro Bianco (ETH Zu
 rich / CSCS); and Thomas Schulthess (ETH Zurich, ETH Zurich / CSCS)\n-----
 ----------------\nP16 - FPGA-Specific Optimizations for Multi-Device Shall
 ow Water Simulations with SYCL\n\nThe shallow water equations are an essen
 tial tool for modeling tides, tsunamis, and storm surges. At PASC 24, we p
 resented an implementation of the shallow water equations running on CPUs,
  GPUs and FPGAs. While the numerical code is shared across the different a
 rchitectures, the implementation uses ...\n\n\nChristoph Alt (Paderborn Un
 iversity, Friedrich-Alexander-Universität Erlangen-Nürnberg); Markus Büttn
 er (University of Bayreuth); Tobias Kenter (Paderborn University); Harald 
 Köstler (Friedrich-Alexander-Universität Erlangen-Nürnberg); Christian Ple
 ssl (Paderborn University); and Vadym Aizinger (University of Bayreuth)\n-
 --------------------\nP17 - Generalization of Long-Range Machine Learning 
 Potentials in Complex Chemical Spaces\n\nThe vastness of chemical space ma
 kes generalization a fundamental challenge for machine learning interatomi
 c potentials (MLIPs). Although MLIPs enable near–quantum-accuracy atomisti
 c simulations at greatly reduced computational cost, their practical relia
 bility is often limited by poor transfe...\n\n\nMichał Sanocki (Technical 
 University of Munich)\n---------------------\nP18 - GPU-Accelerated Method
 s for Numerically Stable Resampling in Fluid-Structure Interaction\n\nFlui
 d-structure interaction simulations require accurate transfer of scalar fi
 elds between overlapping meshes with different topologies. We address the 
 problem of transferring fields from unstructured tetrahedral to structured
  hexahedral meshes.\n\nThis problem is challenging because direct quadratu
 re...\n\n\nSimone Riva (Università della Svizzera italiana) and Patrick Zu
 lian (UniDistance Suisse, Università della Svizzera italiana)\n-----------
 ----------\nP19 - Graph Neural Network Potentials for Million-Atom Molecul
 ar Dynamics Simulations of Aluminum Solidification\n\nSolidification is ub
 iquitous in the fabrication of metal parts. Molecular dynamics simulations
  can predict the microstructure and the corresponding mechanical propertie
 s. However, both high accuracy of interatomic potential energy and scalabi
 lity to millions of atoms are required to capture physical...\n\n\nIan Stö
 rmer and Julija Zavadlav (Technical University of Munich)\n---------------
 ------\nP20 - A High-Performance, GPGPU-Enabled Discontinuous Galërkin Sol
 ver Using OpenMP Offloading and MPI\n\nWe present a GPGPU-enabled modal Di
 scontinuous Galërkin solver that uses OpenMP+MPI. Device code is generated
  by offloading OpenMP pragmas, and inter-device/inter-node communication i
 s enabled by MPI.\nOur test case implements a diffusion-advection solver w
 ith a Runge-Kutta-Chebyshev time stepping sc...\n\n\nMarco Scarpelli, Paol
 a Francesca Antonietti, Carlo De Falco, and Luca Formaggia (Politecnico di
  Milano) and Giovanni Viciconte (ENI S.p.A.)\n---------------------\nP21 -
  Hybrid Block-Structured Grids for Coastal Ocean Domains\n\nAchieving high
  performance and performance portability is critical for next-generation c
 limate and ocean modelling on heterogeneous computing systems. Ocean model
 s face complex, fractal-like coastlines and rapidly varying bathymetry, ma
 king unstructured triangular meshes attractive for their flexibi...\n\n\nJ
 onathan Schmalfuß and Vadym Aizinger (University of Bayreuth)\n-----------
 ----------\nP22 - Hypergraph Partitioning for Sparse Matrix Reordering\n\n
 Fill-in during sparse matrix factorization remains a critical bottleneck i
 n scientific computing. We present an efficient hypergraph partitioning ap
 proach for sparse matrix reordering based on the Clique-Node Hypergraph (C
 NH) representation, building on prior work by Çatalyürek et al. and Selvit
 opi ...\n\n\nRitvik Ranjan, Vincent Maillou, Alexandros Nikolaos Ziogas, a
 nd Mathieu Luisier (ETH Zurich)\n---------------------\nP23 - An Integrate
 d HPC Workflow for AI-Driven Immunogenic Peptide Prediction\n\nImmunogenic
  peptides play important roles as drivers for the adaptive immune response
  - our bodies' ultimate protection against infections and cancers. Parts o
 f these peptides, called epitopes, are recognized by either major histocom
 patibility complexes or antibodies, which then interact with T-cell...\n\n
 \nCathrine Bergh (KTH Royal Institute of Technology), Leonardo Salicari (C
 INECA), Danai Kotzampasi and Victor Reys (Utrecht University), Narendra Ku
 mar (National Institute of Immunology), Archana Achalere and Sunitha Manja
 ri Kasibhatla (Center for the Development of Advanced Computing), Alessand
 ra Villa (KTH Royal Institute of Technology), Uddhavesh Sonavane (Center f
 or the Development of Advanced Computing), and Alexandre Bonvin (Utrecht U
 niversity)\n---------------------\nP24 - A Machine Learning Framework for 
 CFD Applications\n\nIn the present study, an automated framework is prepar
 ed that contains two modules, Computational Fluid Dynamics (CFD) simulatio
 ns and surrogate modelling. CFD simulations are performed to model and mak
 e thermal assessment of battery air cooling in different air stream condit
 ions (i.e. stream veloci...\n\n\nMasumeh Gholamisheeri, Harry Durnberger, 
 Tim Powell, and Jony Castagna (STFC)\n---------------------\nP25 - Maintai
 nable, Sustainable, and Generalizable Data Layouts and Vectorization for R
 igid-Body Molecular Dynamics\n\nls1-MarDyn (ls1) is an MD simulator design
 ed for large-scale simulations of multi-site molecules and has been succes
 sfully used in a variety of scientific studies. It represents molecules as
  rigid bodies composed of multiple interaction sites that each exert force
 s on their neighbours, which are det...\n\n\nSamuel James Newcome, Luis Ga
 ll, David Martin, Markus Mühlhäußer, and Hans-Joachim Bungartz (Technical 
 University of Munich)\n---------------------\nP26 - Optimizing the ICON Dy
 namical Core for GPUs Utilizing GT4Py and DaCe\n\nNumerical weather predic
 tions are based on a numerical model running on a large super computer. Im
 proving the performance of these models is an active field of research whi
 ch benefits society. The ICON model is a finite volume model running on an
  icosahedral mesh.\nFinite volume stencil computations ...\n\n\nChristoph 
 Müller (MeteoSwiss) and Magdalena Luz, Nicoletta Farabullini, Till Ehrengr
 uber, Chia Rui Ong, Daniel Hupp, Philip Müller, Edoardo Paone, Ioannis Mag
 kanaris, Christos Kotsalos, Yilu Chen, Jacopo Canton, Hannes Vogt, Enrique
  González Paredes, Rico Häuselmann, Anurag Dipankar, Mauro Bianco, William
  Sawyer, and Mikael Simberg (ETH Zurich / CSCS)\n---------------------\nP2
 7 - Parallel Tempering on Boundary Conditions with Normalizing Flows to So
 lve Topological Freezing\n\nIn particle physics, Lattice Quantum Chromodyn
 amics (LQCD) studies the strong interaction, responsible, for example, for
  the binding of atomic nuclei, through computational methods.\nAn essentia
 l part of LQCD consists on being able to sample high-dimensional multi-mod
 al distributions, for which direc...\n\n\nVictor Granados (University of B
 ern)\n---------------------\nP28 - Performance-Portable and Highly Scalabl
 e Spectral Transforms with ecTrans\n\nThe continued increase in the skill 
 of weather forecasts observed over the past decades depends crucially on t
 he efficient exploitation of the next generation of high-performance compu
 ters by Earth-system models. The European Centre for Medium-Range Weather 
 Forecast (ECMWF)'s model, the IFS, is one ...\n\n\nSam Hatfield (ECMWF)\n-
 --------------------\nP29 - Perspectives on Teamwork and AI in Scientific 
 Computing\n\nThe development and use of high-quality software—a primary me
 chanism for sustained collaboration and progress in scientific computing—i
 s undergoing profound change, driven by increasing complexity in scientifi
 c drivers and computing architectures and the rapid adoption of AI, placin
 g dem...\n\n\nOlivia B. Newton (University of Montana), Anshu Dubey (Argon
 ne National Laboratory), Denice Ward Hood (University of Illinois Urbana-C
 hampaign), Lois Curfman McInnes (Argonne National Laboratory), and Santiag
 o Ospina Tabares (University of Illinois Urbana-Champaign)\n--------------
 -------\nP30 - The Portable Model for Multi-Scale Atmospheric Prediction (
 PMAP): Towards Sub-Kilometer Scale and Large-Eddy Simulation of Real Weath
 er\n\nThe Portable Model for multi-scale Atmospheric Prediction (PMAP) is 
 an advanced high-resolution numerical model. Written entirely in Python, i
 t leverages the GT4Py domain-specific language to achieve high performance
  and portability – running straightforwardly on laptops and GPU-accelerate
 d HP...\n\n\nLukas Papritz and Nicolai Krieger (ETH Zurich); Christian Küh
 nlein (ECMWF); Till Ehrengruber (ETH Zurich / CSCS); Sara Faghih-Naini (EC
 MWF); and Stefano Ubbiali, Gabriel Vollenweider, Heini Wernli, and Jan Zib
 ell (ETH Zurich)\n---------------------\nP31 - Predictive Alerts and Atmos
 pheric Data for Airport Windshear by CPAS 200m Weather Model\n\nLow-level 
 wind shear (LLWS) at Hong Kong International Airport (HKIA) poses signific
 ant aviation risks, primarily driven by terrain-induced turbulence from La
 ntau Island and sea-breeze interactions. While current mitigation relies o
 n real-time detection and short-term forecasting, for operational pu...\n\
 n\nSai Lun Tin, Chi Chiu Cheung, Ka Ki Ng, and Wai Pang Sze (ClusterTech L
 imited)\n---------------------\nP32 - Profile-Guided-Optimisation of Latti
 ce QCD Contractions on CPU and GPU\n\nWe present a performance optimisatio
 n study of the 2+2 disconnected component bottleneck of Lattice QCD comput
 ation of the hadronic light-by-light contribution to the muon's anomalous 
 magnetic moment. Optimization on CPU and GPU architectures were guided by 
 popular profiling tools perf, valgrind, ns...\n\n\nJingJing Li, Urs Wenger
 , and Roman Gruber (University of Bern)\n---------------------\nP33 - RDQ:
  A Zero-Copy Remote Data Queue for In-Situ Machine Learning in HPC\n\nThe 
 integration of machine learning into the computational sciences is increas
 ingly pursued to reduce time-to-solution, alleviate I/O bottlenecks, and e
 nable adaptive analysis during simulation. We present RDQ (Remote Data Que
 ue), a library for coupling HPC simulations and machine learning training 
 ...\n\n\nMaximilian Sander (TU Dresden) and Jens Domke (RIKEN)\n----------
 -----------\nP34 - Scaling Linear Algebra: Eigenvalue Solvers and Performa
 nce Trends on Contemporary HPC Systems\n\nLarge-scale eigenvalue problems 
 are a fundamental component of modern scientific simulations in fields suc
 h as materials science, computational chemistry, and theoretical physics, 
 where they often represent a dominant computational bottleneck. The effici
 ency and scalability of linear algebra and eig...\n\n\nMaria Montagna, Ser
 gio Orlandini, and Fabio Affinito (CINECA)\n---------------------\nP35 - S
 hedding Light on the Solar Dynamo Using Bayesian Data Science\n\nSolar mag
 netic activity exhibits an approximately 11-year cycle that is far from st
 able, showing strong long-term variability and recurrent episodes of stron
 gly reduced activity known as Grand Minima. Understanding the origin of th
 ese features remains a fundamental challenge in solar physics and is ...\n
 \n\nSimone Ulzega (Zurich University of Applied Sciences, Institute of Com
 putational Life Sciences) and Carlo Albert (Swiss Federal Institute of Aqu
 atic Science and Technology (EAWAG))\n---------------------\nP36 - Stencil
  Computation on Tenstorrent Wormhole\n\nThe rapid ascent of large language
  models (LLMs) has prioritized domain-specific accelerators (DSAs) optimiz
 ed for dense matrix-based deep learning. However, the suitability of these
  architectures for traditional high-performance computing (HPC) kernels, l
 ike stencil-based partial differential equat...\n\n\nLorenzo Piarulli and 
 Daniele De Sensi (Sapienza University)\n---------------------\nP37 - Struc
 tured Reinforcement Learning for Loop Transformation in MLIR\n\nOptimizing
  polyhedral kernels for modern multicore architectures is a high-dimension
 al, non-convex problem where small structural changes often yield orders-o
 f-magnitude runtime variation. While traditional compilers rely on rigid s
 tatic heuristics and autotuners require prohibitive search times, th...\n\
 n\nAbrar Hossain (University of Toledo)\n---------------------\nP38 - Trac
 king Mechanistic Evolution Across Brain Tissues and Cell Types Using Multi
 plex Networks\n\nTracking the evolution of biological function remains a m
 ajor challenge in computational biology, as existing approaches are often 
 limited to sequence conservation, gene presence, or predefined pathways. T
 hese methods can fail to identify conserved functional mechanisms even tho
 ugh constituent genes...\n\n\nKenneth Smith (Oak Ridge National Laboratory
 ); Matthew Lane (University of Tennessee); Alice Townsend and Jean Merlet 
 (Oak Ridge National Laboratory, University of Tennessee); Anna Vlot and Al
 ana Wells (Oak Ridge National Laboratory); and Daniel Jacobson (Oak Ridge 
 National Laboratory, University of Tennessee)\n---------------------\nP39 
 - Tuning the Performance of Three-Body Interactions in Molecular Dynamics\
 n\nMolecular Dynamics (MD) simulations predict thermophysical properties, 
 yet standard pair potentials can lack the desired accuracy for certain app
 lications. Introducing three-body potentials, such as the Axilrod-Teller-M
 uto model, improves results but poses significant computational challenges
 . This ...\n\n\nMarkus Mühlhäußer, Samuel James Newcome, Fabio Gratl-Gaßne
 r, Manish Kumar Mishra, and Hans-Joachim Bungartz (Technical University of
  Munich)\n---------------------\nP40 - Uncertainty Quantification for Ener
 gy Efficiency Analysis of Scientific Applications at Exascale\n\nEnergy ef
 ficiency has become a critical constraint in high-performance computing (H
 PC) as systems scale toward larger node counts. In modern HPC platforms, e
 nergy consumption is influenced by complex interactions among hardware and
  software parameters, including operating frequencies, concurrency le...\n
 \n\nMatheus Machado, Mariana Costa, Matheus Costa, Philippe Navaux, and Ar
 thur Lorenzon (UFRGS) and Antigoni Georgiadou and Bronson Messer (Oak Ridg
 e National Laboratory)\n---------------------\nP41 - Using Generative Mach
 ine Learning to Produce High-Resolution Weather Data\n\nGenerative  machin
 e learning techniques show promise for performing atmospheric downscaling 
 (super-resolution for meteorological data) to produce high-resolution weat
 her and climate simulations. Previous work has evaluated the quality of th
 ese models using standard error scores such as RMSE, absolut...\n\n\nPetar
  Stamenkovic and Mary McGlohon (MeteoSwiss, ETH Zurich); David Leutwyler a
 nd Xavier Lapillonne (MeteoSwiss); Fabian Bösch, Lukas Drescher, and Henri
 que Mendonça (ETH Zurich / CSCS); Sebastian Schemm (University of Cambridg
 e); Siddhartha Mishra (ETH Zurich); and Oliver Fuhrer (MeteoSwiss)\n------
 ---------------\nACMP01 - Adaptive Multidimensional Quadrature on Multi-GP
 U Systems\n\nThe present work introduces a distributed adaptive determinis
 tic quadrature framework for high-dimensional integration on multi-GPU sys
 tems, enabling accurate evaluation of integrals arising in applications su
 ch as radiative transfer and probabilistic design. The method is formulate
 d as a hierarchic...\n\n\nMelanie Tonarelli (Università della Svizzera Ita
 liana)\n---------------------\nACMP02 - Data Augmentation to Improve the P
 erformance of Deep Learning-Based Seismic Inversion\n\nDeep Learning-based
  Seismic Inversion (DLI) is a promising alternative to Full Waveform Inver
 sion (FWI), but struggles with data scarcity. This work evaluates data aug
 mentation techniques adapted from computer vision to address this bottlene
 ck, and demonstrates that generating synthetic earth models...\n\n\nCarlos
  Gomes de Carvalho Junior (Universidade Federal de São Carlos)\n----------
 -----------\nACMP03 - Deep Reinforcement Learning for Algorithm Selection 
 in Molecular Dynamics Simulations\n\nShort-range particle simulations are 
 crucial in physics and chemistry, requiring efficient neighbor search and 
 interaction algorithms. The C++ library AutoPas supports more than 100 alg
 orithms, but no single algorithm is universally optimal. Traditional manua
 l tuning is impractical due to dynamic si...\n\n\nPatrick Metscher (Techni
 sche Universität München)\n---------------------\nACMP04 - Fine-Tuning Lar
 ge Language Models for HPO-Term Recognition\n\nLarge language models (LLMs
 ) offer a promising approach for extracting structured medical information
  from free-text clinical notes. This work investigates fine-tuning LLMs fo
 r Human Phenotype Ontology (HPO) term recognition, a core task in clinical
  genetics that requires accurate identification and...\n\n\nSrinithi Krish
 namoorthy (Cornell University)\n---------------------\nACMP05 - Hypergraph
  Partitioning for Sparse Matrix Reordering\n\nFill-in during sparse matrix
  factorization remains a critical bottleneck in scientific computing. We p
 resent an efficient hypergraph partitioning approach for sparse matrix reo
 rdering based on the Clique-Node Hypergraph (CNH) representation, building
  on prior work by Çatalyurek et al. and Selvitopi ...\n\n\nRitvik Ranjan (
 ETH Zurich)\n---------------------\nACMP06 - Improving Deep Learning Based
  Seismic Inversion with Online Augmentation\n\nThis study investigates the
  application of data augmentation techniques to optimize Deep Learning-bas
 ed Seismic Inversion (DLI), aiming to overcome the scarcity of labeled dat
 asets in the industry. Using the OpenFWI benchmark, the study evaluates fo
 ur incremental strategies: horizontal flipping, syn...\n\n\nLucas Souza (F
 ederal University of São Carlos)\n---------------------\nACMP07 - Mapping 
 the Productivity-To-Energy Trade-Off in Memory-Bound HPC via DVFS, Core Sc
 aling, and C-State Control on Repurposed Hardware\n\nHigh-performance comp
 uting (HPC) systems in resource-constrained environments, such as Africa, 
 often rely on repurposed hardware, shifting the primary financial burden f
 rom capital expenditure to operational energy costs. This work aims to ide
 ntify realistic and reproducible performance–energy...\n\n\nSuné Toerien (
 University of the Witwatersrand), Vele Nefale (UniverUniversity of the Wit
 watersrand), and Ntandoyenkosi Memela and Mubeen Dewan (University of the 
 Witwatersrand)\n---------------------\nACMP08 - Matrix-Free vs. Matrix-Bas
 ed Finite Element Solvers for 3D Advection–Diffusion–Reaction Equations\n\
 nHigh-order finite element methods are attractive for three-dimensional ad
 vection–diffusion–reaction (ADR) problems, but their efficiency on distrib
 uted-memory systems is often limited by memory traffic and communication i
 n large linear solves. This work compares matrix-based (MAT) and ma...\n\n
 \nZhaohui Song (Politecnico di Milano)\n---------------------\nACMP09 - Me
 dulla: Cluster- and Application-level I/O Performance Diagnosis with LLMs\
 n\nModern High-Performance Computing (HPC) I/O systems offer immense bandw
 idth but are notoriously difficult to utilize effectively.\nIdentifying cr
 itical bottlenecks is a challenge that persists for both\ndomain scientist
 s and cluster administrators. Existing analysis tools\ncan identify common
  bottlenec...\n\n\nAnish Sathyanarayanan (BITS Pilani K K Birla Goa Campus
 )\n---------------------\nACMP10 - Mixed Precision Acceleration of Light M
 atter Dynamics: Enabling HPC Co-Design for Quantum Material Discovery\n\nL
 ight-matter dynamics in topological quantum materials holds the promise fo
 r a sustainable society with ubiquitous and power-hungry artificial intell
 igence (AI) by enabling ultralow-power (attojoule) and ultrafast (petahert
 z) computing and sensing devices. A challenge is simulating multiple field
  an...\n\n\nTaufeq Mohammed Razakh (University of Southern California)\n--
 -------------------\nACMP11 - Mosaic: Automatic Categorization of I/O Patt
 erns from Scientific Applications\n\nWhile newly deployed High-Performance
  Computing (HPC) systems embeds more compute capabilities than the generat
 ions before, parallel file systems (PFS) are struggling to keep up with th
 is same trend, increasing the gap between computing power and I/O performa
 nce. If new paradigms and technologies ar...\n\n\nThéo Jolivel (INRIA, Uni
 versite de Rennes)\n---------------------\nACMP12 - Semantic-Aware Implici
 t Neural Compression for Physics Simulations\n\nMachine learning surrogate
 s and data-driven scientific discovery require efficient access to simulat
 ion data, yet physics simulations generate terabyte-scale datasets. Tradit
 ional compression either achieves insufficient ratios or corrupts physics-
 critical features like conservation laws. Implicit n...\n\n\nJessica Ezemb
 a (Carnegie Mellon University)\n---------------------\nACMP13 - Solver-Int
 egrated Lossy and Lossless Compression for Scalable Flow Simulations\n\nLa
 rge-scale computational fluid dynamics (CFD) simulations on modern GPU-acc
 elerated supercomputers generate terabytes of data per run, making checkpo
 inting, storage, and post-hoc analysis increasingly I/O-bound. This bottle
 neck limits data retention and hinders downstream workflows such as visual
 iz...\n\n\nViral Sudip Shah (University of Illinois Urbana-Champaign)\n---
 ------------------\nACMP14 - Stability and Accuracy of the r²SCAN Function
 al for Group-IV Elemental Solids in a Cost-Aware Workflow Perspective\n\nD
 ensity-functional theory (DFT) is a central tool in computational material
 s science, and its predictive power depends critically on the choice of ex
 change-correlation functional. Here we benchmark the meta-GGA r²SCAN in a 
 study of the group-IV elemental solids C, Si, Ge, and Sn, focusing on stab
 ili...\n\n\nAdonis Haxhijaj (EPFL, ETH Zurich)
END:VEVENT
END:VCALENDAR