BEGIN:VCALENDAR VERSION:2.0 PRODID:Linklings LLC BEGIN:VTIMEZONE TZID:America/New_York X-LIC-LOCATION:America/New_York BEGIN:DAYLIGHT TZOFFSETFROM:-0500 TZOFFSETTO:-0400 TZNAME:EDT DTSTART:19700308T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:-0400 TZOFFSETTO:-0500 TZNAME:EST DTSTART:19701101T020000 RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20260924T172352Z LOCATION:B302-B305 DTSTART;TZID=America/New_York:20241121T100000 DTEND;TZID=America/New_York:20241121T170000 UID:submissions.supercomputing.org_SC24_sess534@linklings.com SUMMARY:Research, ACM SRC, and Doctoral Showcase Posters DESCRIPTION:Prompt Phrase Ordering Using Large Language Models in HPC: Eva luating Prompt Sensitivity\n\nLarge language models (LLMs) often require w ell-designed prompts for effective responses, but optimizing prompts is ch allenging due to prompt sensitivity, where small changes can cause signifi cant performance variations. This study evaluates prompt performance acros s all permutations of independent ...\n\n\nNoah Thomasson (Oak Ridge Natio nal Laboratory (ORNL), PCIP Internship Program) and Hilda Klasky (Oak Ridg e National Laboratory (ORNL))\n---------------------\nTrusted Platform Pro visioning for the OpenCHAMI Cluster Management Stack\n\nHigh performance c omputing (HPC) clusters have traditionally relied on proprietary provision ing and management infrastructure. This can be problematic, especially wit h regard to ongoing security and maintenance for vendored systems.\n\nAs a n alternative to this, the Los Alamos National Laboratory (LAN...\n\n\nLuc as Ritzdorf (Los Alamos National Laboratory (LANL), University of Washingt on) and Nicholas Jones, David Allen, and Travis Cotton (Los Alamos Nationa l Laboratory (LANL))\n---------------------\nLagrangian Particle-Tracking in GPU-Enabled Extreme Scale Turbulence Simulations\n\nMany practical turb ulent flow phenomena are naturally studied using a Lagrangian approach tha t treats the fluid medium as a collection of infinitesimal fluid particles . We present a GPU-accelerated algorithm for tracking particles in direct numerical simulations of isotropic turbulence, scaling up t...\n\n\nRohini Uma-Vaideswaran and P.K. Yeung (Georgia Institute of Technology) and Kira n Ravikumar (Science and Technology Corporation, NASA Ames Research Center )\n---------------------\nMachine Learning Applications for Early-Stage Ov arian Cancer Diagnosis\n\nOvarian cancer (OC) significantly impacts women' s health, and despite its prevalence, remains without a definitive cure. E arly detection is crucial for improving treatment outcomes and reducing mo rtality rates and healthcare system costs. Leveraging advancements in mach ine learning, our study seeks ...\n\n\nDelina Mekonnen and Kazem Cheshmi ( McMaster University, Ontario, Canada)\n---------------------\nExploration of Super-Resolution Techniques for Image Compression\n\nTEZIP is a (de)com pression framework leveraging PredNet, a deep neural network designed for video prediction tasks, to exploit temporal locality in time-evolving data . This study evaluates video super-resolution (VSR) models, which enhance low-resolution images by reconstructing high-resolution ones...\n\n\nAlexi s Amoyo (RIKEN, Florida State University); Amarjit Singh and Kento Sato (R IKEN); and Weikuan Yu (Florida State University)\n---------------------\nP roposal for a Parallel Automatic Tuning Using d-Spline According to the Op erating State of the Computer System\n\nSoftware auto-tuning (AT) is a tec hnology that parameterizes factors that affect performance as "performance parameters" and automatically tunes performance. In AT, the tool searches for better values for performance parameters by repeatedly running the pr ogram. Therefore, if the target is a program...\n\n\nYuga Yajima, Akihiro Fujii, and Teruo Tanaka (Kogakuin University, Japan)\n-------------------- -\nPredicting Dataset Popularity for Improved Distributed Content Caching in High Energy Physics\n\nIn High Energy Physics (HEP), large-scale experi ments generate massive amounts of data that are distributed globally. To r educe redundant data transfers and improve analysis efficiency, a disk cac hing system named XCache is used to manage data accesses. By analyzing 11 months of access logs (4.5 mil...\n\n\nMalavikha Sudarshan (University of California, Berkeley)\n---------------------\nA Comparison Study of Open S ource LLMs for HPC Ticket Answering​\n\nWe are designing an automatic tick et answering service for computing centers such as Texas Advanced Computin g Center (TACC), National Center for Supercomputing Applications (NCSA), a nd San Diego Supercomputer Center (SDSC). In this work, we investigate the capability and feasibility of open source l...\n\n\nFangru Linghu, Mingka i Zheng, and Zhao Zhang (Rutgers University)\n---------------------\nGener ating Coupled Cluster Code for Modern Distributed Memory Tensor Software\n \nScientific groups are struggling to adapt their codes to quickly-develop ing GPU-based HPC platforms. The domain of distributed coupled cluster (CC ) calculations is not an exception. Moreover, our applications to tiny QED effects require higher-order CC which include thousands of tensor contrac tions,...\n\n\nJan Brandejs, Johann Pototschnig, and Trond Saue (Laboratoi re de Chimie et Physique Quantiques)\n---------------------\nQuantum Volum e Benchmarking Simulators on HPC Systems\n\nQuantum computing simulators v ia classical computing are essential for small-scale comprehension and tes ting of quantum computing algorithms. Quantum Volume (QV) is a well-establ ished benchmark for comparing Quantum Processing Units (QPUs) in the NISQ era. However, there is no QV benchmark of a larg...\n\n\nChristian Boehme (GWDG, Germany); Lourens van Niekerk (GWDG, Germany; Georg-August Universi tät Göttingen (GWDG)); Dhiraj Kumar and Aasish Kumar Sharma (University of Göttingen, GWDG, Germany); and Tino Meisel and Martin Leandro Paleico (GW DG, Germany)\n---------------------\nScalable Performance and Accuracy Ana lysis for Distributed and Extreme-Scale Systems\n\nThe Scalable Performanc e and Accuracy analysis for Distributed and Extreme-scale systems (SPADE) project focuses on advancing monitoring, optimization, evaluation, and dec ision-making capabilities for extreme-scale systems. This poster presents efforts targeting several advanced monitoring capabilit...\n\n\nHeike Jago de (University of Tennessee, Innovative Computing Laboratory (ICL); Univer sity of Tennessee); Shirley Moore (University of Texas at El Paso); Vincen t Weaver (University of Maine); Anthony Danalis (University of Tennessee, Innovative Computing Laboratory (ICL)); and Christoph Lauter (University o f Texas at El Paso)\n---------------------\nTrackable Agent-Based Evolutio n Models at Wafer Scale\n\nEmerging ML/AI hardware accelerators, like the 850,000 processor Cerebras Wafer-Scale Engine (WSE), hold great promise to scale up the capabilities of evolutionary computation. However, challenge s remain in maintaining visibility into underlying evolutionary processes while efficiently utilizing the...\n\n\nMatthew Andres Moreno and Connor Y ang (University of Michigan), Emily Dolson (Michigan State University), an d Luis Zaman (University of Michigan)\n---------------------\nMeteorologic Real-Time Extreme Learning Machine for Pressure Prediction\n\nSignificant advances in weather prediction stemmed primarily from combining observati onal data, sophisticated modeling techniques, and analysis of simulated or historical weather data. Additionally, the applicability of machine learn ing-based applications on edge devices has expanded, addressing a v...\n\n \nAnagha Ram (University of Chicago); Kate Keahey (University of Chicago, Argonne National Laboratory (ANL)); and Alicia Esquivel Morel (University of Missouri, Columbia)\n---------------------\nAssessing Matrix Multiplica tion Performance with Fully Homomorphic Encryption\n\nAs data from various domains is increasingly shared and processed in the cloud, Homomorphic En cryption (HE) provides a crucial solution for ensuring privacy in the post -quantum era. In this work, we evaluate the performance and accuracy of th e HE matrix multiplication leveraging SEAL\nlibrary kernels...\n\n\nJusto Molina, Rocío Carratalá-Sáez, Manuel F. Dolz, and Sandra Catalán (Jaume I University, Spain)\n---------------------\nLarge Genomic Language Models: Towards Their Hyperparameter Optimization\n\nThis study explores hyperpara meter optimization for encoder-only genomics large language models, balanc ing machine learning performance, hardware resource usage and power consum ption. Multiple search techniques and objective functions were implemented , and from the results we obtained two types of m...\n\n\nAndrei Politov a nd Neringa Jurenaite (Technical University Dresden, ZIH; Center for Scalab le Data Analytics and Artificial Intelligence (ScaDS.AI) Dresden/Leipzig)\ n---------------------\n5G in Practice: Measuring Emerging Wireless Techno logy in Rural Iowa for Edge Devices in Distributed Computation Workloads\n \nEdge computing, the notion of moving computational tasks from servers to the data-generating network edge, is an increasingly popular model for da ta processing. 5G wireless technologies offer an opportunity to enable com plex distributed edge computing workflows by minimizing the overhead incur red in...\n\n\nZack Murry and Alicia Esquivel Morel (University of Missour i, Columbia; University of Chicago) and Kate Keahey (Argonne National Labo ratory (ANL), University of Chicago)\n---------------------\nBreaking the Barriers to Effective Supercomputing: Web Dashboard for Job Accounting and Performance Metrics\n\nThe NSF-funded Anvil supercomputer, built and main tained by the Purdue University Rosen Center for Advanced Computing, enabl es efficient research computing across a variety of scientific domains nat ionwide. In addition to the traditional terminal interface, Anvil uses the open-source web portal fram...\n\n\nRichie Tan (Purdue University), Anjal i Rajesh (Rutgers University), and Guangzhen Jin and Rajesh Kalyanam (Purd ue University)\n---------------------\nPipeInfer: Accelerating LLM Inferen ce Using Asynchronous Pipelined Speculation​\n\nInference of large languag e models (LLMs) across computer clusters has become a focal point of resea rch in recent times, with many acceleration techniques taking inspiration from CPU speculative execution. These techniques reduce memory bandwidth r equirements, but also increase\nlatency per inference...\n\n\nBranden Butl er and Sixing Yu (Iowa State University), Arya Mazaheri (Technical Univers ity Darmstadt), and Ali Jannesari (Iowa State University)\n--------------- ------\nScalable Motif Counting on Large-Scale Dynamic Graphs\n\nMotifs, s mall subgraphs of k vertices, such as triangles and cliques, are well stud ied for static networks. They are used to characterize different biologica l networks and align different networks. Counting motifs reveal insights i nto the topological structure of a network such as MPI event graphs. ...\n \n\nAli Khan (University of North Texas)\n---------------------\nProfiling and Bottleneck Identification for Large Language Model Optimizations\n\nL arge language models (LLMs) have shown they can perform scientific tasks. They are capable of assisting researchers in data interpretation, instrume nt operation, knowledge synthesis, and hypothesis generation. However, LLM s must first be trained on a large dataset of scientific tasks and data. T rai...\n\n\nAlvin Hoang and Brian Chen (University of California, Riversid e; Pacific Northwest National Laboratory (PNNL))\n---------------------\nE stablishing Best Practices for Applying Inline Compressed Arrays to Improv e Performance in HPC\n\nHPC applications require massive amounts of memory to process large datasets. While data compression is used to avoid bottle necks in transmission and storage, it is still necessary to decompress thi s data into memory to use it. Inline compressed arrays (ICA) is a method w hich keeps the data compress...\n\n\nAdam Whitestone (Clemson University)\ n---------------------\nPerformance of LAMMPS-SNAP in Different Runtime En vironments\n\nLAMMPS (Large-scale Atomic/Molecular Massively Parallel Simu lator) is a widely used molecular dynamics simulator. It is used here to s imulate the high-pressure BC8 phase of carbon using the Spectral Neighbor Analysis Potential (SNAP). This simulation employs the Kokkos C++ performa nce portability la...\n\n\nShubh Pachchigar (San Francisco State Universit y, NERSC at LBNL (Lawrence Berkeley National Laboratory)) and Adam Lavely, Rahulkumar Gayatri, and Brandon Cook (NERSC at LBNL (Lawrence Berkeley Na tional Laboratory))\n---------------------\nExploring Software-Defined Net working for Routing in Dragonfly Topology\n\nWhile efficient routing in Dr agonfly networks presents significant challenges, the advent of Software-d efined Networking (SDN) offers new opportunities for routing optimization by providing a global network view. This research proposes an SDN-based ad aptive routing approach for Dragonfly interconnec...\n\n\nRam Sharan Chaul again and Xin Yuan (Florida State University)\n---------------------\nClus ter-Based Methodology for Characterizing the Performance of Portable Appli cations\n\nThis work focuses on performance portability and proposes a met hodological approach to assessing and explaining how different kernels beh ave across various hardware architectures using the RAJA Performance Suite (RAJAPerf). Our methodology leverages metrics from the Intel top-down pip eline and clust...\n\n\nBefikir Bogale (University of Tennessee, Knoxville ); Olga Pearce and Tom Scogland (Lawrence Livermore National Laboratory (L LNL)); and Michela Taufer (University of Tennessee, Knoxville)\n---------- -----------\nHardware-Independent Sampling Library for CPUs and (Multi-)GP Us: hws\n\nTo be energy efficient and fully utilize modern hardware, it is important to gain as much insight as possible into the performance and ef ficiency of an application. Especially in the age of artificial intelligen ce, it gets more and more important to keep, e.g., track of the total ener gy consumption ...\n\n\nMarcel Breyer (University of Stuttgart, Germany; I nstitute for Parallel and Distributed Systems) and Alexander Van Craen and Dirk Pflüger (University of Stuttgart, Germany; University of Stuttgart, Institute for Parallel and Distributed Systems)\n---------------------\niS eeMore: Design of a 256-Node RPi Cluster to Visualize LLM Computation Thro ugh Light and Movement for Mass Audiences\n\niSeeMore is a kinetic cluster of 256-Raspberry Pi (RPi) computers that visually realizes supercomputing concepts (parallelism, data flow, algorithms) through servo and LED drive n movement. In this poster, we describe the design of iSeeMore, the first large-scale cluster to combine movement and light...\n\n\nHayden Estes, Gr egory Bolet, Collin Dunphy, Matt Lorenzo, Mallika Pamula, Kalina Kazmiercz ak, Sahithi Sarva, Margaret Ellis, Sam Blanchard, Godmar Back, and Kirk Ca meron (Virginia Tech)\n---------------------\nFAS-GED: GPU-Accelerated Gra ph Edit Distance Computation\n\nGraph Edit Distance (GED) is a fundamental metric for assessing graph similarity with critical applications across v arious domains, including bioinformatics, classification, and pattern reco gnition. However, the exponential computational complexity of GED has hind ered its adoption for large-scale gr...\n\n\nAdel Dabah and Andreas Herten (Forschungszentrum Jülich)\n---------------------\nPcMINER: Mining Perfor mance-Related Commits at Scale\n\nPerformance inefficiencies in software c an severely impact application quality and resource utilization. Addressin g these issues often requires significant developer effort, yet the lack o f large-scale, open-source performance datasets hinders the development of effective mitigation strategies. To f...\n\n\nMd Abul Kalam Azad, Manoj A lexender, Matthew Alexender, Syed Salauddin Mohammad Tariq, Foyzul Hassan, and Probir Roy (University of Michigan - Dearborn)\n--------------------- \nEdge-Enabled Real-Time Data Processing in Power-Efficient Weather Statio ns Using IBIS\n\nThere is a growing need to acquire a larger quantity of m eteorological data to address climate change. In this paper, we design an improved Automatic Weather Station (AWS) based on a prototype from the Nat ional Center for Atmospheric Research (NCAR). We integrate this weather st ation with IBIS, a pl...\n\n\nWilliam Fowler (Tufts University); Alicia Es quivel Morel (University of Missouri, Columbia); and Kate Keahey (Universi ty of Chicago, Argonne National Laboratory (ANL))\n---------------------\n An Accurate and Scalable Multidimensional Quantum Solver for Partial Diffe rential Equations\n\nQuantum computing is an innovative technology that ca n solve certain problems faster than classical computing. One of its promi sing applications is in solving partial differential equations (PDEs). How ever, current PDE solvers that are based on variational-quantum-eigensolve r (VQE) techniques suffer...\n\n\nManu Chaudhary, Ishraq Islam, Alvir Nobe l, Dylan Kneidel, Vinayak Jha, Ben Phillips, Kareem El-Araby, Manish Singh , and Esam El-Araby (University of Kansas)\n---------------------\nExploit ing Data Compression and Low Precision for Exascale Fusion Turbulence Simu lations\n\nFuture exascale systems will feature unprecedented computing po wer with over 10^18 FLOPS, provided by thousands of heterogeneous computin g nodes. To fully harness such potential, applications must scale effectiv ely and efficiently utilize these computing architectures.\n\nMany perform ance losses on la...\n\n\nJan Laukemann (University of Erlangen-Nuremberg, Germany; Erlangen National High Performance Computing Center); Felix Jung (Technical University of Munich); Carl-Martin Pfeiler (Max Planck Institu te for Plasma Physics); Diego Jimenez (Max Planck Computing and Data Facil ity); Carsten Clauss (ParTec AG, Germany); Tilman Dannert (Max Planck Comp uting and Data Facility); Frank Jenko (Max Planck Institute for Plasma Phy sics); Erwin Laure (Max Planck Computing and Data Facility); Martin Schulz (Technical University of Munich); and Gerhard Wellein (Erlangen National High Performance Computing Center)\n---------------------\nExploring DAOS as a Burst Buffer for a 100 Gbps DAQ Real-Time Streaming System\n\nWe pres ent an experimental evaluation of a burst buffer for a real-time DAQ strea ming system designed to transmit instrument data to remote data centers. T he system is based on EJ-FAT, a load balancing system capable of Nx 100Gbp s streams, distributing data from event sources to processing nodes. We... \n\n\nXinxin Mei, Graham Heyes, Vardan Gyurjyan, Carl Timmer, Michael Good rich, Amitoj Singh, and Ilya Baldin (Thomas Jefferson National Accelerator Facility)\n---------------------\nQDD: Multi-Node Implementation of Decis ion Diagram-Based Quantum Circuit Simulator with Ring Communication and Au to SWAP Insertion\n\nThis poster introduces QDD, a multi-node implementati on of a decision diagram (DD)-based quantum circuit simulator. DD-based si mulators offer faster simulation of algorithms like Shor's compared to sta tevector (SV) simulators by compressing the quantum state using a graph re presentation. However, pa...\n\n\nYusuke Kimura and Junpei Koyama (Fujitsu Ltd) and Shaowen Li, Hiroyuki Sato, and Masahiro Fujita (The University o f Tokyo, Japan)\n---------------------\nTowards Scalable Quantum Simulatio n on Wafer-Scale Engines\n\nQuantum computing poses many benefits to the c omputing world, given its efficiency over classical computers in problem s paces. However, the high noise and low logical qubit count of contemporary noisy-intermediate-scale-quantum (NISQ) devices make the development and execution of quantum algorithms ...\n\n\nBen Phillips, Dylan Kneidel, Alvi r Nobel, and Esam El-Araby (University of Kansas)\n---------------------\n Improving SpGEMM Performance Through Reordering and Cluster-Wise Computati on\n\nSparse Matrix-Matrix Multiplication (SpGEMM) is a key kernel in many scientific applications and graph workloads. SpGEMM is known to suffer fr om poor performance due to irregular memory access patterns. Gustavson's a lgorithm, a traditional approach for SpGEMM, involves row/column-wise oper ations, fa...\n\n\nAbdullah Al Raqibul Islam (University of North Carolina at Charlotte), Helen Xu (Georgia Institute of Technology), Dong Dai (Univ ersity of Delaware), and Aydin Buluç (Lawrence Berkeley National Laborator y (LBNL))\n---------------------\nDesign of Reliable and Efficient Syscall Hooking Library for a Parallel File System\n\nTo provide file systems in user space, FUSE and syscall interception libraries have been used. Howeve r, FUSE has a performance degradation problem, and the syscall interceptio n library has problems with reliability and portability. \n\nThis study pr oposes a design and implementation of the reliable an...\n\n\nHaruka Miyau chi (University of Tsukuba, Japan)\n---------------------\nInteractive and Tool-Agnostic ML-Driven Workflow for Automated HPC Performance Modeling\n \nThis work presents an automated, reproducible, ML-based performance mode ling workflow for HPC systems. The proposed workflow fully automates data generation, preprocessing, ML model training and validation. Since the pro posed approach is generic and not tailored to a specific application, our workfl...\n\n\nZhaobin Zhu and Sarah Neuwirth (Johannes Gutenberg Universi ty Mainz)\n---------------------\nParallelization of the Finite Element-Ba sed Mesh Warping Algorithm Using Hybrid Parallel Programming\n\nWarping la rge volume meshes is computationally expensive and has applications to bio mechanics, aerodynamics, image processing, and cardiology. Existing parall el implementations of mesh warping algorithms do not take advantage of sha red-memory and one-sided communication features available in MPI-3. ...\n\ n\nAbir Haque (University of Kansas)\n---------------------\nPINE: Efficie nt Yet Effective Piecewise Linear Trees\n\nDecision trees are popularly us ed in statistics and machine learning. Piecewise linear trees, a type of m odel-based decision tree, employ linear models to evaluate splits and pred ict outcomes at the leaf nodes. While they can offer high accuracy, they a re computationally expensive, and currently, no...\n\n\nZecheng Li and Jia jia Li (North Carolina State University)\n---------------------\nPoseidon: A Source-to-Source Translator for Holistic HPC Optimization of Ocean Mode ls on Regular Grids\n\nOcean simulation models often underperform on moder n high-performance computing (HPC) architectures, necessitating costly and time-consuming code rewrites.\n\nWe introduce Poseidon, an HPC-oriented s ource-to-source translator for Fortran-based fluid dynamics solvers used i n ocean and weather models wi...\n\n\nMaurice Brémond (French Institute fo r Research in Computer Science and Automation (INRIA); Université Grenoble Alpes, France); Hugo Brunie (Université Grenoble Alpes, France; French In stitute for Research in Computer Science and Automation (INRIA)); Laurent Debreu (French Institute for Research in Computer Science and Automation ( INRIA); Université Grenoble Alpes, France); Rupert W. Ford (Hartree Centre , Science and Technology Facilities Council (STFC), UK); Florian Lemarié ( French Institute for Research in Computer Science and Automation (INRIA); Université Grenoble Alpes, France); Anna Mittermair (Technical University of Munich); Andrew R. Porter (Hartree Centre, Science and Technology Facil ities Council (STFC), UK); Julien Rémy (French Institute for Research in C omputer Science and Automation (INRIA); Université Grenoble Alpes, France) ; Philippe Rosales and Martin Schreiber (Université Grenoble Alpes, France ; French Institute for Research in Computer Science and Automation (INRIA) ); Martin Schulz (Technical University Munich); Sergi Siso (Hartree Centre , Science and Technology Facilities Council (STFC), UK); and Arthur Vidard (French Institute for Research in Computer Science and Automation (INRIA) ; Université Grenoble Alpes, France)\n---------------------\nSanQus: Stale ness and Quantization-Aware Full-Graph Decentralized Training in GNNs\n\nG raph neural networks (GNNs) have demonstrated significant success in model ing graphs; however, they encounter challenges in efficiently scaling to l arge graphs. To address this, we propose the SanQus system, advancing our previous work, Sancus. SanQus reduces the need for expensive communication am...\n\n\nHongbo Yin (Hong Kong University of Science and Technology, Gu angzhou); Jingshu Peng (Hong Kong University of Science and Technology); Q iyu Liu (Southwest University); Zhao Chen (Hong Kong University of Science and Technology); Yingxia Shao (Beijing University of Posts and Telecommun ications); Yanyan Shen (Shanghai Jiao Tong University); Lei Chen (Hong Kon g University of Science and Technology); and Jiannong Cao (The Hong Kong P olytechnic University)\n---------------------\nLM-Offload: Performance Mod el-Guided Generative Inference of Large Language Models with Parallelism C ontrol\n\nLarge language models (LLMs) have achieved remarkable success in various natural language processing tasks. However, LLM inference is high ly computational and memory-intensive, creating extreme deployment challen ges. Tensor offloading, combined with tensor quantization and asynchronous task executio...\n\n\nJianbo Wu (University of California, Merced); Jie R en (College of William & Mary); Shuangyan Yang (University of California, Merced); Konstantinos Parasyris, Giorgis Georgakoudis, and Ignacio Laguna (Lawrence Livermore National Laboratory (LLNL)); and Dong Li (University o f California, Merced)\n---------------------\nAn Error-Bounded Lossy Compr ession Method with Bit-Adaptive Quantization for Particle Data\n\nWe prese nt error-bounded lossy compression tailored for particle datasets from div erse scientific applications in cosmology, fluid dynamics, and fusion ener gy sciences. As today's high-performance computing capabilities advance, t hese datasets often reach trillions of points, posing significant anal...\ n\n\nCongrong Ren (The Ohio State University), Sheng Di (Argonne National Laboratory (ANL)), Longtao Zhang and Kai Zhao (Florida State University), and Hanqi Guo (The Ohio State University)\n---------------------\nWeb-Base d Simulator of Superscalar RISC-V Processors\n\nUnlock the power of supers calar processor design with our cutting-edge RISC-V simulator! Tailored fo r IT students, researchers, and HPC professionals, this web-based tool bri ngs complex architectures to life with an intuitive, customizable interfac e. Explore processor components, tweak configuration...\n\n\nJiri Jaros (B rno University of Technology, Faculty of Information Technology) and Micha l Majer, Jakub Horky, and Jan Vavra (Brno University of Technology)\n----- ----------------\nA Survey-Based Evaluation of the Efficacy of a Girls Who Code Club at the University of Southern Indiana\n\nPrograms like Girls Wh o Code (GWC) are pivotal in working to inspire and equip young women with the skills and confidence needed to pursue careers in computing. Understan ding the impact of such initiatives is particularly important for addressi ng the decline in interest among girls aged 13 to 17, a ...\n\n\nMaya Sesh an, Cathy Sandoval, Alyson Collins, and Srishti Srivastava (University of Southern Indiana)\n---------------------\nIntegrating HPCToolkit with Tool s for Automated Analysis\n\nHPCToolkit enables users to gather detailed in formation about application performance. Users can capture fine-grained me asurement data, which may include instruction-level samples on CPUs and GP Us. Collected data can be huge, making manual inspection using GUI tools d ifficult and time-consuming. We ...\n\n\nDragana Grbic (Lawrence Livermore National Laboratory (LLNL), Rice University); Matthew LeGendre and Olga P earce (Lawrence Livermore National Laboratory (LLNL)); and John Mellor-Cru mmey (Rice University)\n---------------------\nGPU Compression (for Scient ific Data) Done Right\n\nError-bounded lossy compression is a critical tec hnique for significantly reducing scientific data volumes. Compared to CPU -based compressors, GPU-based compressors exhibit substantially higher thr oughputs, fitting better for today's HPC applications. To overcome the dat a challenge, GPU-based scient...\n\n\nJiannan Tian (Indiana University), X in Liang (University of Kentucky), and Sheng Di and Franck Cappello (Argon ne National Laboratory (ANL))\n---------------------\nAI-Based Scalable An alytics for Improving Performance and Resilience of HPC Systems\n\nAs high -performance computing (HPC) advances to exascale levels, its role in scie ntific fields such as medicine, climate research, finance, and scientific computing becomes increasingly critical. However, these large-scale system s are susceptible to performance variations caused by anomalies, includ... \n\n\nEfe Sencan, Beste Oztop, Ayse Coskun, Brian Kulis, and Manuel Egele (Boston University) and Benjamin Schwaller, Vitus Leung, and Jim Brandt (S andia National Laboratories)\n---------------------\nProfiling the Impact of Hyper-Threading on Pagosa Hydrocodes\n\nPagosa is a hydrodynamics compu ter code designed for massively parallel environments, operating within an Eulerian framework using a fixed Cartesian mesh. This project investigate s the performance of Pagosa on Sapphire Rapids nodes, which feature many-c ore architectures and high bandwidth memory. By...\n\n\nBlake Milstead (Un iversity of Tennessee, Knoxville)\n---------------------\nOptimal Client S election Algorithms for Federated Learning\n\nDue to the heterogeneity of resources and data, client selection plays a paramount role in the efficac y of Federated Learning (FL) systems. The time taken by a training round i s determined by the slowest client. Also, energy consumption and carbon fo otprint are seen as primary concerns. In this cont...\n\n\nAlan Nunes (Flu minense Federal University, Brazil; University of Bordeaux, France); Crist ina Boeres and Lúcia Drummond (Fluminense Federal University, Brazil); and Laércio Pilla (University of Bordeaux, French Institute for Research in C omputer Science and Automation (INRIA))\n---------------------\nMIGnificie nt: Fast, Isolated, and GPU-Enabled Serverless Functions\n\nEmerging appli cations in machine learning and personalized medicine introduce new challe nges and requirements for secure computing. While the exclusive allocation of resources to a single tenant provides necessary isolation, it also com es at the cost of hardware underutilization. While solutions lik...\n\n\nM arcin Copik and Alexandru Calotoiu (ETH Zürich); Pengyu Zhou (University o f Toronto, Canada); Lukas Tobler (AYES); and Torsten Hoefler (ETH Zürich)\ n---------------------\nPerfFlowAspect: A User-Friendly Performance Tool f or Scientific Workflows\n\nIn scientific computing, artificial intelligenc e and/or machine learning (AI/ML) are appearing increasingly often in scie ntific workflows on supercomputers, due to their ability to solve more com plex problems. With respect to performance analysis tools, the nature of t hese workflows creates new requ...\n\n\nAliza Lisan (University of Oregon) , Tapasya Patki and Stephanie Brink (Lawrence Livermore National Laborator y (LLNL)), Zhiwei Yang and Spencer Greene (Worcester Polytechnic Institute (WPI)), Konstantinos Parasyris (Lawrence Livermore National Laboratory (L LNL)), Shubbhi Taneja (Worcester Polytechnic Institute (WPI)), and Hank Ch ilds (University of Oregon)\n---------------------\nIncreasing the Efficie ncy of Neutral Atoms by Reducing Qubit Waste from Measurement-Related Ejec tions\n\nQuantum computing is an emerging field that has had an impact on various domains. This poster focuses on the advantageous neutral atom tech nology for quantum computing. In neutral atom technology, a significant ch allenge arises in the measurement of quantum output, where the need to eje ct atoms phys...\n\n\nJason Ludmir and Tirthak Patel (Rice University)\n-- -------------------\nAlgorithmic Patterns from Computational Biology for P roxy Application Development and Co-Design\n\nHigh-performance computing h ardware is co-developed with U.S. DOE codes, and proxy applications (apps) based on these codes are critical technologies for iterative innovation. Numerical modeling/simulation proxies have been the most impactful in co-d esign. To broaden the types of computation availab...\n\n\nLogan Williams and Gavin Conant (North Carolina State University) and Jan Ciesko and Amy Powell (Sandia National Laboratories)\n---------------------\nDART-X: Soft ware Infrastructure for Prototyping In-Memory Data Transfer Between Ensemb le Data Assimilation and Coupled Earth Systems Models\n\nData assimilation (DA) integrates real-world data into coupled climate models, enhancing pr ediction accuracy and capturing Earth system complexity. NCAR’s Data Assim ilation Research Testbed (DART) is an ensemble DA tool for climate predict ions with NCAR’s Community Earth System Model (CE...\n\n\nAnh Pham (Mount Holyoke College, National Center for Atmospheric Research (NCAR)); Suman S hekhar (Rutgers University, National Center for Atmospheric Research (NCAR )); and Helen Kershaw, Dan Amrhein, and Ufuk Turuncoglu (National Center f or Atmospheric Research (NCAR))\n---------------------\nScored Non-Determi nistic Finite Automata Processor for Sequence Alignment\n\nWith the surge in symbolic data across various fields, efficient pattern matching and reg ular expression processing have become crucial. Non-deterministic Finite A utomata (NFA), commonly used for pattern matching, face memory bottlenecks on general-purpose platforms. This has driven interest in Doma...\n\n\nRy an Karbowniczak and Rasha Karakchi (University of South Carolina)\n------- --------------\nParallel Verification of Neural Networks Applied to Medica l Imaging\n\nNeural network verification provides model robustness guarant ees in the presence of noise. We generate verification specifications for medical imaging models based on the U-Net architecture and solve pixel-by- pixel verification problems on a massive scale. Efficiency of this NP-comp lete problem is s...\n\n\nJonathan Andreasen (Georgia Tech Research Instit ute), Diego Manzanas Lopez and Taylor Johnson (Vanderbilt University), and Yatis Dodia (Georgia Tech Research Institute)\n---------------------\nBen chmarking and Modeling of Producer-Consumer Data Movement Performance in S cientific Workflows\n\nIn this poster, we present the Analytics4X (A4X) fr amework, a workflow framework that enables systematic studies, evaluation, and modeling of HPC/HTC workflow data movement in a flexible and controll ed environment. We apply A4X to an in situ molecular dynamics workflow to assess the performance trad...\n\n\nIan Lumsden (University of Tennessee); Olga Pearce, Jae-Seung Yeom, and Tom Scogland (Lawrence Livermore Nationa l Laboratory (LLNL)); and Michela Taufer (University of Tennessee, Knoxvil le)\n---------------------\nEfficient Approaches to Analyzing Large Dynami c Networks\n\nDynamic graphs, characterized by their evolving topologies o ver time, necessitate continuous updates to their associated graph propert ies, including shortest paths, vertex coloring, and strongly connected com ponents. Traditional static graph algorithms, which re-compute properties following each mod...\n\n\nArindam Khanda and S.M. Shovan (Missouri Univer sity of Science and Technology), Aashish Pandey and Farahnaz Hosseini (Uni versity of North Texas), Sajal K. Das (Missouri University of Science and Technology), Boyana Norris (University of Oregon), and Sanjukta Bhowmick ( University of North Texas)\n---------------------\nCluster Management with Containerization on Switches\n\nThis research explores the utilization of underused computational resources in network switches from Arista and Mel lanox by deploying containers for auxiliary tasks in computational cluster networks. We tested five scenarios: 1) running cloud-init services for ef ficient boot processes, 2) using Tele...\n\n\nRobin Lee Simpson (Los Alamo s National Laboratory (LANL); California Polytechnic State University, San Luis Obispo); Anvitha Ramachandran (Los Alamos National Laboratory (LANL) , University of Southern California (USC)); and Dohyun Lee (Los Alamos Nat ional Laboratory (LANL), University of Illinois Urbana-Champaign)\n------- --------------\nJUmPER: Performance Data Monitoring, Instrumentation and V isualization for Jupyter Notebooks\n\nComputational performance, e.g. CPU or GPU utilization, is crucial for analyzing machine learning (ML) applica tions and their resource-efficient deployment. However, the ML community o ften lacks accessible tools for holistic performance engineering, especial ly during exploratory programming such as ...\n\n\nElias Werner and Anton Rygin (Center for Scalable Data Analytics and Artificial Intelligence (Sca DS.AI) Dresden/Leipzig; Technical University Dresden, Center for Interdisc iplinary Digital Sciences (CIDS)); Andreas Gocht-Zech (German Environment Agency, AI Lab); and Sebastian Döbel and Matthias Lieber (Technical Univer sity Dresden, ZIH; Technical University Dresden, Center for Interdisciplin ary Digital Sciences (CIDS))\n---------------------\nQFw: A Quantum Framew ork for Large-Scale HPC Ecosystems\n\nThis work extends Quantum Framework (QFw) by integrating it with Northwest Quantum Simulator (NWQ-Sim) and by introducing a lightweight python library (qfw\_backend) that allows multip le frontends (e.g., Qiskit) to interact with QFw. This extension enables Q Fw to flexibly decouple frontends from bac...\n\n\nSrikar Chundury (North Carolina State University); Amir Shehata, Thomas Naughton, and Seongmin Ki m (Oak Ridge National Laboratory (ORNL)); Frank Mueller (North Carolina St ate University); and In-Saeng Suh (Oak Ridge National Laboratory (ORNL))\n ---------------------\nSeesaw: Elastic Scaling for Task-Based Distributed Programs\n\nModern batch schedulers in HPC environments enable the shared use of available computational resources via provisioning discrete sets of resources matching user requirements. The lack of elasticity in such scen arios is often addressed using a Pilot job model where multiple separate r equests are pool...\n\n\nMatthew Chung (University of California, Riversid e)\n---------------------\nImproving Polyhedral-Based Optimizations with D ynamic Coordinate Descent\n\nPolyhedral optimizations have been a cornerst one of kernel optimization for many years. These techniques use a geometri c model of loop iterations to enable transformations like tiling, fusion, and fission. The elegance of this approach lies in its ability to produce highly efficient code through ful...\n\n\nGaurav Verma (Stony Brook Univer sity; Stony Brook University, Institute for Advanced Computational Science (IACS)); Michael Canesche (Universidade Federal de Minas Gerais, Brazil); Barbara Chapman (Stony Brook University; Stony Brook University, Institut e for Advanced Computational Science (IACS)); and Fernando Magno Quintão P ereira (Universidade Federal de Minas Gerais, Brazil)\n------------------- --\nMatRIS: Performance Portable Math Library of IRIS Runtime for Multi-De vice Heterogeneity\n\nWe present the recent efforts for MatRIS, the perfor mance portable math library of IRIS runtime for multi-device heterogeneity . MatRIS provides dense linear algebra — BLAS (Basic Linear Algebra Subpro grams) and LAPACK (Linear Algebra PACKage) — capabilities across different back ends ava...\n\n\nMohammad Alaul Haque Monil, Narasinga Rao Miniskar, Pedro Valero-Lara, Keita Teranishi, and Jeffrey S. Vetter (Oak Ridge Nati onal Laboratory (ORNL))\n---------------------\nCoVA: Compiler for Versati le Architectures\n\nThe rapid advancement in computing demands and the inc reasing complexity of modern applications (e.g., image processing, numeric al computation, and machine learning) necessitate efficient heterogeneous computing solutions. Project CoVA proposes an MLIR-based compilation flow designed to bridge the g...\n\n\nHongzheng Tian (University of California, Irvine; Hewlett Packard Labs); Alok Mishra (Hewlett Packard Labs); and Si tao Huang (University of California, Irvine)\n---------------------\nI/O C haracterization of Heterogeneous Workflows\n\nWorkflows consist of individ ual applications such as scientific simulations and data analytics. These applications constitute different stages of the workflow, each comprising heterogeneous characteristics such as run-times and system requirements. T he heterogeneity in these workflow stages dictates...\n\n\nOlga Kogiou (Fl orida State University), Hariharan Devarajan and Chen Wang (Lawrence Liver more National Laboratory (LLNL)), Weikuan Yu (Florida State University), a nd Kathryn Mohror (Lawrence Livermore National Laboratory (LLNL))\n------- --------------\nThe P3 Explorer: Exploring the Performance, Portability, a nd Productivity Wilderness\n\nThis paper documents the development of a we b-based tool designed to organise and present visual representations of pe rformance, portability, and productivity (P3) data from previously publish ed scientific studies. The P3 Explorer operates as both an open repository of scientific data and a data das...\n\n\nMatthew A. Smith and Steven A. Wright (University of York, England) and Zaman Lantra and Gihan Mudalige ( University of Warwick, England)\n---------------------\nKVSort: Drasticall y Improving LLM Inference Performance via KV Cache Compression\n\nLarge la nguage model (LLM) deployment necessitates high inference throughput due t o the increasing demand for text generation. To accelerate inference, the prefill mechanism avoids repeated computations via introducing KV Cache (i n HBM). However, the KV cache size increases with the input and genera...\ n\n\nBaixi Sun (Indiana University); Dingwen Tao (Institute of Computing T echnology, Chinese Academy of Sciences); Xiaodong Yu (Stevens Institute of Technology); and Fengguang Song (Indiana University)\n------------------- --\nNew Semi-Implicit Electrostatic Particle-In-Cell Method to Extend Scop e of the Exascale WarpX Code\n\nTime-explicit fully-kinetic simulation of magnetically confined fusion plasmas is out of reach of existing supercomp uters due to the multi-scale nature of the system. Physics approximations and time-implicit methods are typically used to simulate such plasmas. A n ew energy-conserving, semi-implicit ...\n\n\nRoelof E. Groenewald and Dani el C. Barnes (TAE Technologies); Mikhail Tyushev (University of Saskatchew an, Canada); Weiqun Zhang and Axel Huebl (Lawrence Berkeley National Labor atory (LBNL)); Ales Necas (TAE Technologies); Andrei Smolyakov (University of Saskatchewan, Canada); Jean-Luc Vay (Lawrence Berkeley National Labora tory (LBNL)); Toshiki Tajima (TAE Technologies; University of California, Irvine); and Sean Dettrick (TAE Technologies)\n---------------------\nTurb ocharging Dask Apps: Accelerating Data Flow with ProxyStore\n\nDespite adv ancements in distributed computing libraries, performance challenges, such as data serialization and transfer, still persist. We focus on understand ing data limitations within Dask, a versatile and popular Python library d esigned for distributed and parallel computing, and then investigat...\n\n \nKlaudiusz Rydzy (Loyola University, Chicago)\n---------------------\nEnh ancing Performance Reproducibility on HPC Workflows\n\nReproducibility is related to achieving consistent performance across multiple runs of the sa me application in an identical computing environment. From a computational perspective, jobs should run for equal time, with equal performance repea tedly. Without this, we see significant variations in perfo...\n\n\nAinara Garcia (Clemson University, Argonne National Laboratory (ANL))\n--------- ------------\nHPC Fastpass: Visualizing Descriptive and Predictive HPC Que ue Time Data\n\nWith large HPC systems, users will often jockey for better queue times to get quicker results. Unfortunately, getting accurate estim ations of queue times requires understanding complex and abundant data col lected from myriad HPC system loggers. To aid with this, researchers are e xploring machine lea...\n\n\nConnor Scully-Allison (University of Utah, Sc ientific Computing and Imaging Institute (SCI); National Renewable Energy Laboratory (NREL))\n---------------------\nUncover the Overhead and Resour ce Usage for Handling KV Cache Overflow in LLM Inference\n\nLLM inference includes two phases: prefill phase and decode phase. Prefill phase process es input tokens simultaneously to generate the first token. Decode phases generate the subsequent tokens one after another until they either meet a termination or reach the max length. To avoid recomputation, the...\n\n\nJ ie Ye (Illinois Institute of Technology), Bogdan Nicolae (Argonne National Laboratory (ANL)), and Anthony Kougkas and Xian-He Sun (Illinois Institut e of Technology)\n---------------------\nScalable Low-Latency Hardware Fun ction Chaining with Chain Control Circuit\n\nTo deliver advanced services with high performance and flexibility, we are developing a computing platf orm that integrates various hardware accelerators (HWAs). Our goal is to b uild customized services for each user by combining application functions. To achieve this, we propose Hardware Function Ch...\n\n\nYuta Ukon, Takah iro Kawahara, Yuki Arikawa, Naoki Miura, and Teruaki Ishizaki (NTT Corpora tion); Wataru Kanemori, Ryoko Tamura, and Kenjiro Mori (Fujitsu Ltd); and Takeshi Sakamoto (NTT Corporation)\n---------------------\nSWARM: Scientif ic Workflow Applications on Resilient Metasystem\n\nCurrent (centralized) resource management strategies typically require a global view of distribu ted HPC systems, relying on a cluster-wide resource manager for scheduling , with static, expert-tuned rules. This centralized decision-making approa ch suffers from resilience, efficiency and scalability i...\n\n\nPawel Zuk (University of Southern California, Information Sciences Institute); Hong wei Jin (Argonne National Laboratory (ANL)); Imtiaz Mahmud (Lawrence Berke ley National Laboratory (LBNL)); Krishnan Raghavan (Argonne National Labor atory (ANL)); Komal Thareja (Renaissance Computing Institute (RENCI)); Shi xun Wu (University of California, Riverside); Prasanna Balaprakash (Oak Ri dge National Laboratory (ORNL)); Franck Cappello (Argonne National Laborat ory (ANL)); Zizhong Chen (University of California, Riverside); Ewa Deelma n (University of Southern California, Information Sciences Institute); She ng Di (Argonne National Laboratory (ANL)); Aiden Hamade and Mariam Kiran ( Oak Ridge National Laboratory (ORNL)); Anirban Mandal, Erik Scott, and Con g Wang (Renaissance Computing Institute (RENCI)); and John Wu (Lawrence Be rkeley National Laboratory (LBNL))\n---------------------\nRAPIDS: Reduced API Data-Transfer Specifications\n\nPerformance of all-purpose communicat ion libraries like MPI is fundamentally limited by the all-in-one approach these libraries use. RAPIDS (Reduced API Data-transfer Specifications) di vides the functionality of these libraries into separate, more focused API s, enabling library and application devel...\n\n\nRiley Shipley and Anthon y Skjellum (Tennessee Technological University), Purushotham Bangalore (Un iversity of Alabama), and Patrick Bridges (University of New Mexico)\n---- -----------------\nJACC: HPC Meta-Programming and Performance Portability Ecosystem for Julia Language\n\nWe present JACC (Julia for ACCelerators), the first high-level, meta-programming, and performance-portable model for the just-in-time and LLVM-based Julia language. JACC provides a unified a nd lightweight front end across different back ends available in Julia, en abling the same Julia code to run ef...\n\n\nPedro Valero-Lara, William Go doy, Het Mankad, Keita Teranishi, and Jeffrey Vetter (Oak Ridge National L aboratory (ORNL))\n---------------------\nGNN-RL: An Intelligent HPC Resou rce Scheduler\n\nEfficient resource allocation in high-performance computi ng (HPC) environments is crucial for optimizing utilization, minimizing ma ke-span, and enhancing throughput. We propose GNN-RL, a novel intelligent scheduler that leverages a hybrid Graph Neural Network and Reinforcement L earning model, learni...\n\n\nKyrian Chinemeze Adimora and Hongyang Sun (U niversity of Kansas)\n---------------------\nA Novel Gradient Compression Design with Ultra-High Compression Ratio for Communication-Efficient Feder ated Learning\n\nFederated learning is a privacy-preserving machine learni ng approach. It allows numerous geographically distributed clients to coll aboratively train a large model while maintaining local data privacy. In h eterogeneous device settings, limited network bandwidth is a major bottlen eck that constrains s...\n\n\nZhijing Ye and Xiaodong Yu (Stevens Institut e of Technology)\n---------------------\nMind Your Manners: Detoxifying La nguage Models via Attention Head Intervention\n\nTransformer-based Large L anguage Models have advanced natural language processing with their abilit y to generate fluent text. However, these models exhibit and amplify toxic ity and bias learned from training data, posing new ethical challenges. We build upon the \attnlens{} framework to allow for sc...\n\n\nJordan Petty john (Colorado School of Mines)\n---------------------\nExploring Fine-Gra ined Memory Analysis for PIM Offloading\n\nIn modern computing, a challeng e is the data bottleneck between the CPU and RAM. This issue arises becaus e the CPU can process data faster than it can be accessed from RAM; this i s worsened by the fact that large amounts of RAM are less accessible than a powerful CPU. Furthermore, RAM’s high c...\n\n\nThomas Papka (Loyola Uni versity, Chicago) and Nanda Velugoti and Kyle Hale (Illinois Institute of Technology)\n---------------------\nCharacterizing the Performance of the GENE-X Code for Gyrokinetic Turbulence Simulations\n\nSimulating plasma tu rbulence in the edge region of a magnetic confinement fusion (MCF) device is crucial for identifying optimal operational scenarios for future fusion energy commercialization. GENE-X, an Eulerian electromagnetic gyrokinetic code, can simulate plasma turbulence throughout an MCF de...\n\n\nJordy T rilaksono (Max Planck Institute for Plasma Physics, Tokamak Theory Divisio n); Jeremy Johnathan Williams (KTH Royal Institute of Technology); Philipp Ulbl (Max Planck Institute for Plasma Physics, Tokamak Theory Division); Tilman Dannert and Erwin Laure (Max Planck Computing and Data Facility); S tefano Markidis (KTH Royal Institute of Technology); and Frank Jenko (Max Planck Institute for Plasma Physics, Tokamak Theory Division)\n----------- ----------\nNetCDFaster: A Geospatial Cyberinfrastructure Enhancing Multi- Dimensional Scientific Dataset Access and Visualization Through Machine Le arning Optimization\n\nThis project introduces an enhanced solution for ac cessing and processing NetCDF data, a widely used standard in geosciences for storing multidimensional data. Existing tools often compromise on perf ormance or lack full workflow support. The proposed system integrates mach ine learning, specifically ...\n\n\nZhenlei Song and Zhe Zhang (Texas A&M University) and Alan Sussman (University of Maryland)\n------------------- --\nBenchmarking Quantum-Inspired Optimization Platforms and Tools on an H PC Cluster\n\nThis work compares performance of various QUBO software tool s on different platforms available on the Sol supercomputer at Arizona Sta te University. CPU, GPU and the NEC vector engine (VE) card provide variou s means to implement these computations on simulated qubits while employin g various solvers,...\n\n\nJencey Dole, Gil Speyer, and Torey Battelle (Ar izona State University)\n---------------------\nImproving the Performance of Proof-of-Space in Blockchain Systems\n\nBlockchain technologies enable the success of digital currencies by providing security, decentralization, and trustless operation. Two dominant consensus algorithms, Bitcoin’s Pro of-of-Work and Ethereum’s Proof-of-Stake, balance security, scalability, a nd energy efficiency, though PoW is...\n\n\nVarvara Bondarenko (Illinois I nstitute of Technology), Renato Diaz (University of Central Florida), and Lan Nguyen (Illinois Institute of Technology)\n---------------------\nAsse ssing the Impact of Real-Time Traffic Updates on Traffic Flow: A High-Perf ormance Computing Perspective on Scalability and Demand\n\nIn the dynamic landscape of smart cities, it is necessary to improve the potential of rea l-time traffic data and high-performance computing to optimize traffic flo w through dynamic re-routing strategies. Our research contributes to the a ssessment of how real-time traffic optimization and alternative...\n\n\nPa ulo Silva, Pavlína Smolková, Sofia Michailidu, Jakub Beránek, Roman Macháč ek, Kateřina Slaninová, and Jan Martinovič (IT4Innovations, VSB - Technica l University of Ostrava) and Radim Cmar (Sygic a.s.)\n-------------------- -\nLarge-Scale Randomized Program Generation with Large Language Models​\n \nLarge, diverse datasets of executable programs are required for training and running machine learning models to find insights in program performan ce. While many open-source code repositories exist freely on popular softw are development websites such as GitHub, the safety and executability of s uch pr...\n\n\nTyler Lam (University of California, Berkeley) and Lingda L i (Brookhaven National Laboratory)\n---------------------\nHARVEST-2.0: Hi gh-Performance Vision Framework for End-to-End Preprocessing, Training, In ference, and Visualization\n\nDeep learning (DL) thrives on data; however, it inherits a major limitation: training and testing datasets must be ful ly annotated for supervised deep neural networks (DNNs) training. To addre ss this challenge, we introduce HARVEST-2.0, a high-performance computer-v ision framework for end-to-end dat...\n\n\nNawras Alnaasan, Anirudh Potlap ally, Tian Chen, Matthew Lieber, Aamir Shafi, Hari Subramoni, Scott Sheare r, and Dhabaleswar K. Panda (The Ohio State University)\n----------------- ----\nPrototype Development and Testing of a Smart Buoy System for Coastal and Marine Ecosystems Using IBIS\n\nEffective environmental monitoring tr aditionally required physical presence at research sites. IoT technologies now enable remote data collection, revolutionizing this field. This work presents a prototype of smart buoys designed for coastal and marine ecosys tems. Located in the IBIS testbed, these ...\n\n\nJulia Harper (Loyola Uni versity, Chicago)\n---------------------\nFormal Approaches to Characteriz e Emerging Arithmetic Realizations\n\nAs HPC models grow increasingly comp lex, disparities in floating point implementations across hardware platfor ms begin to pose significant challenges to reproducibility and reliability . This is especially so, given that HPC employs hardware optimized for per formance, which quite often deviates from ...\n\n\nPaul Jiang (Purdue Univ ersity) and Iana Lin (University of California, Berkeley)\n--------------- ------\nPerformance of Inline Compression with Software Caching for Reduci ng the Memory Footprint in pySDC\n\nThe volume of data required for high p erformance computing (HPC) jobs is growing faster than the memory storage available to store the required data, leading to performance bottlenecks. Hence the need for inline data compression, which reduces the amount of al located memory needed by storing all dat...\n\n\nEmily Lattanzio (Clemson University)\n---------------------\nMemory Disaggregation in Serverless Co mputing\n\nIn recent years, the slowing of advancements in memory technolo gy and applications’ increasing demand for memory have resulted in high pe rformance computation becoming bottlenecked by availability of memory. One existing solution, far memory, involves swapping pages to a remote machin e rather ...\n\n\nJoshua Lowe (Illinois Institute of Technology, Rose-Hulm an Institute of Technology) and Nanda Velugoti (Illinois Institute of Tech nology)\n---------------------\nBringing It HOME: Analyzing Contention Hot spots Across the Memory Hierarchy with Low Overhead\n\nThe increasing dema nd for computing in scientific research has given rise to memory contentio n and performance bottlenecks. Existing solutions often carry high overhea ds or lack the necessary detail for effective contention mitigation. To ta ckle these challenges, we are developing a powerful tool, H...\n\n\nDhruv Gajaria and Andrés Márquez (Pacific Northwest National Laboratory (PNNL))\ n---------------------\nEve: Less Memory, Same Might\n\nAdaptive optimizer s, which adjust the learning rate for individual parameters, have become t he standard for training deep neural networks. AdamW is a popular adaptive method that maintains two optimizer state values (momentum and variance) per parameter, doubling the model’s memory usage durin...\n\n\nAditya Toma r (University of California, Berkeley) and Siddharth Singh, Tom Goldstein, and Abhinav Bhatele (University of Maryland)\n---------------------\nPerf ormance of N10 Benchmarks with Different BLAS Implementations\n\nThe NERSC -10 Benchmark Suite is a collection of tests which are designed to evaluat e various aspects of system architecture and performance. This study utili zes the suite to evaluate the CPU compute nodes of two leadership-class su percomputers, Perlmutter at NERSC and Summit at OLCF, using two N10 b...\n \n\nShubh Pachchigar (NERSC at LBNL (Lawrence Berkeley National Laboratory ), San Francisco State University) and Adam Lavely, Rahulkumar Gayatri, an d Brandon Cook (NERSC at LBNL (Lawrence Berkeley National Laboratory))\n-- -------------------\nOn the Accuracy and Efficiency of Approximate Triangl e Counting via Randomized Numerical Linear Algebra\n\nWe study two algorit hmic approaches to approximate triangle counting and compare their accurac y and efficiency. The first one is based on randomized matrix-matrix multi plication, which can be faster, simpler, and more parallelizable on modern processors. The second is based on trace estimation, whic...\n\n\nIrene S imó Muñoz, Koby Hayashi, and Richard Vuduc (Georgia Institute of Technolog y)\n---------------------\nEnhancing HPC Resource Management to Integrate Quantum Workflows\n\nIn recent years, quantum computing has demonstrated t he potential to revolutionize specific algorithms and applications by solv ing problems exponentially faster than classical computers. However, its w idespread adoption for general computing remains a future prospect. In thi s work we demonstrate the...\n\n\nAmir Shehata, Thomas Naughton, and In-Sa eng Suh (Oak Ridge National Laboratory (ORNL))\n---------------------\nSim plifying HPC Resource Selection: A Tool for Optimizing Execution Time and Cost on Azure\n\nAzure Cloud offers a wide range of resources for running HPC workloads, requiring users to configure their deployment by selecting VM types, number of VMs, and processes per VM. Suboptimal decisions may le ad to longer execution times or additional costs for the user. We are deve loping an open-source...\n\n\nMarco A. S. Netto, Wolfgang De Salvador, and Davide Vanzo (Microsoft Corporation)\n---------------------\nStalls and M emory Analysis on Fujitsu A64FX and NVIDIA Grace\n\nARM-based multicore CP Us, such as NVIDIA Grace and Fujitsu A64FX, dominate contemporary HPC, fea turing 32-256 cores with cache hierarchies and up to 1 TB/s memory bandwid th. While benchmarks like STREAM show similar performance across these sys tems, diverse applications, particularly graph and neare...\n\n\nYan Kang (Pennsylvania State University), Sayan Ghosh (Pacific Northwest National L aboratory (PNNL)), Mahmut Kandemir (Pennsylvania State University), and An dres Marquez (Pacific Northwest National Laboratory (PNNL))\n------------- --------\nIdentifying Regions of Non-Determinism in HPC Simulations Throug h Event Graph Alignment\n\nHigh performance computing (HPC) applications u sing MPI (Message Passing Interface) often face non-determinism (ND) due t o asynchronous MPI calls, making ND source identification challenging. Mod eling execution as an event graph, where MPI calls are nodes and communica tion is edges, can be useful. F...\n\n\nDhroov Pandey (University of North Texas)\n---------------------\nPersistent and Partitioned MPI for Stencil Communication\n\nMany parallel applications rely on iterative stencil ope rations, whose performance is dominated by communication costs at large sc ales. Several MPI optimizations, such as persistent and partitioned commun ication, reduce overheads and improve communication efficiency through amo rtized setup costs and...\n\n\nGerald Collom and Amanda Bienz (University of New Mexico)\n---------------------\nComputational Radiation Hydrodynami cs with FleCSI\n\nPhotons are essential in many matter-light interaction s ystems, with applications in physics (e.g., core-collapse supernovae) and engineering (e.g., neutron transport). Although the theory of radiating fl uids has been established since the 1980s, developing robust and efficient numerical simulations...\n\n\nMammadbaghir Baghirzade (The University of Texas at Austin; CCS-7, Los Alamos National Laboratory); Md Shihab Shahria r Khan (Michigan State University; CCS-7, Los Alamos National Laboratory); Yoonsoo Kim (California Institute of Technology; CCS-7, Los Alamos Nation al Laboratory); Sudarshan Neopane (University of Tennessee, Knoxville; CCS -7, Los Alamos National Laboratory); Alexander Strack (University of Stutt gart, Germany; CCS-7, Los Alamos National Laboratory); Farhana Taiyebah (F lorida State University; CCS-7, Los Alamos National Laboratory); and Julie n Loiseau and Hyun Lim (Los Alamos National Laboratory (LANL))\n---------- -----------\nImprovement of Bridges-2 Resource Utilization Through User Op timization\n\nThis poster presents our two-phase solution for improving GP U utilization in NSF-funded ACCESS high-performance computing (HPC) cluste rs, with a pilot implementation on Pittsburgh Supercomputing Center’s Brid ges-2. Our approach addresses the limitations of Open XdMoD, which lacks p er-job GPU u...\n\n\nWalter Ashworth (University of Tennessee, Knoxville); Julian Uran (Pittsburgh Supercomputing Center (PSC), Carnegie Mellon Univ ersity); Michela Taufer (University of Tennessee, Knoxville); and Paola Bu itrago (Pittsburgh Supercomputing Center (PSC), Carnegie Mellon University )\n---------------------\nEnhancing the Traditional Benchmarks for Paralle l Computing Education\n\nThis work introduces enhanced benchmark suites — NeoRodinia, RedOptBench and CUDAMicroBench — specifically designed to enri ch the educational landscape of parallel programming. By integrating pract ical examples and detailed optimization processes into traditional benchma rks, these suites...\n\n\nXinyao Yi (University of North Carolina at Charl otte), Anjia Wang (Intel Corporation), and Yonghong Yan (University of Nor th Carolina at Charlotte)\n---------------------\nActive Learning for Meta material Optimization on HPC and QC Integrated Systems\n\nActive learning algorithms, integrating machine learning, quantum computing and optics sim ulation in an iterative loop, offer a promising approach to optimizing met amaterials. However, these algorithms can face difficulties in optimizing highly complex structures due to computational limitations. Hi...\n\n\nSeo ngmin Kim and In-Saeng Suh (Oak Ridge National Laboratory (ORNL))\n------- --------------\nEmpowering Scientific Datasets with Large Language Models\ n\nThe growing volume and complexity of scientific data pose significant c hallenges in data management, organization, and analysis. Our objective is to enhance the utilization of historical scientific datasets across vario us disciplines. To address this, we propose integrating large language mod els (LL...\n\n\nBrian Chen and Alvin Hoang (University of California, Rive rside; Pacific Northwest National Laboratory (PNNL))\n-------------------- -\nGeneralizing ExaDigiT Datacenter Digital Twin Framework for Multiple Ar chitectures\n\nExaDigiT is an open-source framework for developing compreh ensive digital twins (DTs) of liquid-cooled supercomputers. DTs merge tele metry, simulations, and AI/ML/RL to create a complete virtual representati on of a system. This provides an effective tool for testing a variety of s ystem optimizations...\n\n\nSedrick Bouknight and Matthias Maiterth (Oak R idge National Laboratory (ORNL)); Tim Dykes (Hewlett Packard Enterprise (H PE)); Rashadul Kabir (Colorado State University); and Vineet Kumar, Scott Greenwood, Jesse Hines, Wes Brewer, and Feiyi Wang (Oak Ridge National Lab oratory (ORNL))\n---------------------\nProfiling Communication Overhead i n 3D Parallel Pretrain of Large Language Models\n\nTraining large language models (LLMs) efficiently requires addressing the communication overhead introduced by parallelism strategies like Tensor, Pipeline, and Data Paral lelism. This work profiles the communication patterns in LLM pretraining u sing the Polaris supercomputer, highlighting the impact...\n\n\nWeijin Liu and Xiaodong Yu (Stevens Institute of Technology)\n---------------------\ nFault-Tolerant Numerical Iterative Algorithms at Scale\n\nNumerical itera tive algorithms are struck by multiple error types when deployed on large- scale HPC platforms: fail-stop errors (failures) and silent errors, striki ng both as computation errors and memory bit-flips. Our novel approach pro vides efficient fault-tolerant algorithms that are capable of d...\n\n\nAl ix Tremodeux (ENS Lyon)\n---------------------\nEvolving a Multi-Populatio n Evolutionary-QAOA on Distributed QPUs\n\nOur research combines an Evolut ionary Algorithm with a Quantum Approximate Optimization Algorithm (QAOA) to update the ansatz parameters, in place of traditional gradient-based me thods, and benchmarks on the Max-Cut problem. We demonstrate that our Evol utionary-QAOA pairing performs on par or better...\n\n\nFrancesca Schiavel lo, Edoardo Altamura, and Stefano Mensa (Hartree Centre, Science and Techn ology Facilities Council (STFC), UK)\n---------------------\nDevelopment o f TEZip in PyTorch: Integrating New Prediction Models into an Existing Com pression Framework\n\nIn high performance computing, researchers often wor k with extremely large time-series datasets. Compression techniques allow them to shrink their data down, allowing for quicker and less storage-inte nsive transfers. TEZip is a compression model that utilizes PredNet (a vid eo prediction model) to pr...\n\n\nAkshay Nambudiripad (Columbia Universit y), Amarjit Singh and Kento Sato (RIKEN), and Weikuan Yu (Florida State Un iversity)\n---------------------\nPower Patterns: Understanding the Energy Dynamics of I/O for Parallel Storage Configurations\n\nAs HPC application s become more I/O intensive, understanding their power consumption pattern s is necessary to develop energy-saving solutions. Here, we evaluate the e nergy consumption of I/O operations on two popular HPC parallel file syste ms: Lustre and DAOS. We develop models to predict the energy...\n\n\nMaya Purohit (University of Chicago, Colby College) and Valerie Hayot-Sasson an d Kyle Chard (University of Chicago)\n---------------------\nA Sparse Appr oach for Translation-Based Training of Knowledge Graph Embeddings\n\nKnowl edge graph (KG) learning offers a powerful framework for generating new kn owledge and making inferences. Training KG embedding can take a significan tly long time, especially for larger datasets. Our analysis shows that the gradient computation of embedding and vector normalization are the domin. ..\n\n\nMd Saidul Hoque Anik and Ariful Azad (Indiana University)\n------- --------------\nComparing Cache Utilization Trends for Regional Scientific Caches with Transfer Learning Models\n\nTo enhance data sharing and reduc e access latency in scientific collaborations, high energy physics (LHC CM S experiment) employs regional in-network storage caches. Accurate predict ions of cache utilization trends help design new caching policies and impr ove capacity planning. This study leverages t...\n\n\nErica Wang (Californ ia Institute of Technology, Lawrence Berkeley National Laboratory (LBNL))\ n---------------------\nAn Adaptive Kernel Execution for Dynamic Applicati ons on GPUs Using CUDA Graphs\n\nWe propose a novel approach for executing dynamic applications on GPUs. Different from traditional approaches that use a single kernel, our method allows the GPU to autonomously allocate co mputational resources at runtime. We decompose a kernel into multiple frag ment kernels and dynamically launch a...\n\n\nKento Kitamura, Kenji Tanaka , and Kazunori Seno (NTT Corporation)\n---------------------\nAnalyzing Al ltoall Algorithms with SST\n\nAlltoall collective operations in MPI are cr itical in several types of computation, including matrix multiplication an d transposition, and machine learning applications. As a result, it is cri tical that these operations are performant for large amounts of data. Mean while, dragonfly networks are beco...\n\n\nShannon Kinkead (Sandia Nationa l Laboratories) and Amanda Bienz (University of New Mexico)\n------------- --------\nFault Tolerance in Krylov Subspace Methods\n\nToday’s complex HP C systems are incredibly powerful yet equally likely to experience failure s. The scientific applications on these HPC systems are mostly iterative i n nature. Iterative solvers have some inherent fault tolerance, but they a re still susceptible to errors. One subset of these it...\n\n\nSandesh Pan dit (University of Alabama in Huntsville)\n---------------------\nCommunic ation Hiding for Matrix-Free Finite Element Operators of a Complex PDE: No nlinear Stokes Flow of Earth’s Mantle\n\nFor large-scale matrix-free finit e-element PDE solvers, parallel matrix-vector products typically comprise the dominant computational cost. Synchronization steps for input and outpu t, when degrees of freedom (DOFs) lying on inter-process boundaries are co mmunicated, become the dominant serial portio...\n\n\nMax Heldman and Joha nn Rudi (Virginia Tech)\n---------------------\nA Zero-Copy Storage with M etadata-Driven File Management Using Persistent Memory\n\nPersistent Memor y (PM) is a promising next-generation storage device, combining features o f both volatile memory (like DRAM) and non-volatile memory (like SSDs). Ma ny studies use PM to optimize training to advance deep learning technology . However, these studies have not addressed the issue of multi...\n\n\nGua ngxing Hu (North Carolina State University), Awais Khan (Oak Ridge Nationa l Laboratory (ORNL)), and Frank Mueller (North Carolina State University)\ n---------------------\nNeural Network Optimization and Performance Analys is for Real-Time Object Detection at the Edge\n\nReal-time object detectio n is an important and computationally intensive task that is gaining more attention in the field of autonomous systems. Recently, a novel object det ection algorithm called RT-DETR has emerged, demonstrating superior speed compared to the popular YOLO series. In recent years,...\n\n\nAnimesh Ghos e (Brookhaven National Laboratory, Carnegie Mellon University)\n---------- -----------\nPerformance Engineering and Mesoscale-Microscale Coupling for Wind Energy Simulations\n\nWind farm simulations require data from mesosc ale atmospheric simulations as initial and boundary conditions for the mic roscale turbine environments. The Energy Research and Forecasting (ERF) co de bridges this scale gap and provides an efficient GPU-enabled parallel i mplementation with adaptive mesh...\n\n\nMukul Dave, Ann Almgren, Donald W illcox, Weiqun Zhang, and Aaron Lattanzi (Lawrence Berkeley National Labor atory (LBNL))\n---------------------\nFortranX: Harnessing Code Generation , Portability, and Heterogeneity in Fortran\n\nDue to its historical popul arity, Fortran was used to implement many important scientific application s. The complexity of these applications, along with the transition to mode rn high performance languages like C++, has made modernization and optimiz ation challenging for these applications. Significa...\n\n\nSanil Rao (Car negie Mellon University); Mike Franusich (SpiralGen, Inc.); Mohammad Alaul Haque Monil, Het Mankad, and Jeffrey Vetter (Oak Ridge National Laborator y (ORNL)); and Franz Franchetti (Carnegie Mellon University)\n------------ ---------\nCreating Code LLMs for HPC: It’s LLMs All the Way Down\n\nLarge language models (LLMs) are being used increasingly by software developers , researchers, and students to assist them in coding tasks. While newer LL Ms have been improving their coding abilities with regards to serial codin g tasks, they consistently perform worse when it comes to parallelism and. ..\n\n\nAman Chaturvedi, Daniel Nichols, Siddharth Singh, and Abhinav Bhat ele (University of Maryland)\n---------------------\nORCHA: A Performance Portability System for Flash-X — A Multiphysics Application Software\n\nHe terogeneous platforms demand tailored data structures, algorithm mappings, and efficient execution, often leading to numerous, hard-to-maintain vari ants of source codes. ORCHA, our performance portability orchestration sys tem, addresses these challenges through abstractions and code generation, st...\n\n\nYoungjun Lee and Wesley Kwiecinski (Argonne National Laboratory (ANL)); Klaus Weide (Argonne National Laboratory (ANL), University of Chi cago); Johann Rudi (Virginia Tech); Jared O’Neal (Argonne National Laborat ory (ANL)); Emil Vatai and Mohamed Wahib (RIKEN); and Tom Klosterman and A nshu Dubey (Argonne National Laboratory (ANL))\n---------------------\nDis CostiC: Simulating MPI Applications Without Executing Code\n\nWe present t he cross-architecture parallel simulation framework DisCostiC (Distributed Cost in Clusters). It predicts the performance of real or hypothetical, m assively parallel MPI(+X) programs on current and future supercomputer sys tems. The novelty of DisCostiC is that it employs analytical, firs...\n\n\ nAyesha Afzal, Georg Hager, and Gerhard Wellein (Friedrich-Alexander Unive rsity, Erlangen-Nuremberg; Erlangen National High Performance Computing Ce nter)\n---------------------\nHigh-Performance Computing Resilience Analys is Using Large Language Models\n\nThis doctoral showcase highlights three pivotal works conducted during my PhD that collectively advance the field of high-performance computing (HPC) resilience analysis using large langua ge models (LLMs).\n\nThe first work introduces HAPPA, a modular platform f or HPC Application Resilience Analysis. ...\n\n\nHailong Jiang (Kent State University)\n---------------------\nEffects of Lossy Compression Data on Machine Learning Models\n\nMachine learning is a fundamental tool that is incorporated in fields across academia and industry. Due to the large amou nts of data needed for training machine learning models, compression is ut ilized because it reduces the data footprint playing a critical role in st orage. Machine learning involve...\n\n\nMax Faykus (Clemson University)\n- --------------------\nData Layout Optimizations for Tensor Applications\n\ nThe performance of tensor applications is often bottlenecked by data move ment across the memory subsystem. This dissertation contributes domain-spe cific programming frameworks (compilers and runtime systems) that optimize data movement in tensor applications. We develop novel execution reorderi ng an...\n\n\nMahesh Lakshminarasimhan (University of Utah)\n------------- --------\nFFT-Based Spherical Harmonics and Radial Transforms on GPU\n\nMo dern high-performance computing clusters switch to the GPUs, as opposed to CPUs, as the source of their computational power. GPUs are tailored for d ata-parallel algorithms where multiple cores perform the same operations o n different memory locations. However, making CPU code run within GPU cons tr...\n\n\nDmitrii Tolmachev (ETH Zürich)\n---------------------\nGoing Be yond the Chicken and Egg Situation with Modern MPI Features\n\nMPI has bec ome the de facto standard for distributed memory computing since its incep tion in 1994. While the MPI standard has evolved to include new technologi es like RDMA, many applications still rely on the original set of MPI oper ations.\n\nThis thesis investigates the current usage of MPI. We note...\n \n\nTim Jammer (Technical University Darmstadt)\n---------------------\nAc celerating Communications in High-Performance Scientific Workflows\n\nAdva nces in networks, accelerators, and cloud services encourage programmers t o reconsider where to compute — such as when fast networks make it cost-ef fective to compute on remote accelerators despite added latency. Workflow and cloud-hosted serverless computing frameworks can manage multi-st...\n\ n\nJ. Gregory Pauloski (University of Chicago, Argonne National Laboratory (ANL))\n---------------------\nSupporting End Users in Implementing Quant um Computing Applications\n\nQuantum computing applications require expert knowledge to perform complex steps: (1) selecting a suitable quantum algo rithm, (2) generating the quantum circuit, (3) compiling/executing the qua ntum circuit, and (4) decoding the results. This creates high entry barrie rs for end users with limited exp...\n\n\nNils Quetschlich (Technical Univ ersity of Munich)\n---------------------\nAccelerating HPC Workflow Result s and Performance Reproducibility Analytics\n\nModern high-performance com puting (HPC) workflows produce massive datasets, often exceeding 100+ TB p er day, driven by instruments collecting data at gigabytes per second. The se workflows, executed on advanced HPC systems with heterogeneous storage devices, high-performance microprocessors, accelera...\n\n\nKevin Assogba (Rochester Institute of Technology)\n---------------------\nScalable Plann ing Platform for Orchestration of Autonomous Systems Across Edge-Cloud Con tinuum\n\nEdge accelerators, such as NVIDIA Jetson, are enabling rapid inf erence of deep neural network (DNN) models and computer vision algorithms through low-end graphics processing unit (GPU) modules integrated with ARM -based processors. Their compact form factor allows integration with mobil e platforms, s...\n\n\nSuman Raj (Indian Institute of Science, Bangalore)\ n---------------------\nToward Performance & Portability & Productivity in Parallel Programming\n\nAchieving *performance*, *portability*, and *prod uctivity* for data-parallel computations (e.g., MatMul and convolutions) h as emerged as a major research challenge. The complex hardware design of c ontemporary parallel architectures, including GPUs and CPUs, requires adva nced program optimizations to...\n\n\nAri Rasch (University of Münster, Ge rmany)\n---------------------\nEfficient Large Dynamic Graph Analysis on E merging Storage Technology\n\nGraph-structured data analysis is extensivel y used in various real-world applications, including biology, social media , and recommendation systems. With the increasing prevalence of real-time data, many graphs become dynamic and evolve over time. Thus, dynamic graph processing systems become a neces...\n\n\nAbdullah Al Raqibul Islam (Univ ersity of North Carolina at Charlotte)\n---------------------\nEfficient, Scalable, Robust Neuromorphic High Performance Computing\n\nThe rapid adva ncement in Artificial Neural Networks (ANNs) has paved the way for Spiking Neural Networks (SNNs), which offer significant advantages in energy effi ciency and computational speed, especially on neuromorphic hardware. My re search focuses on the development of Efficient, Robust, and Scal...\n\n\nB iswadeep Chakraborty (Georgia Institute of Technology)\n------------------ ---\nEnhancing HPC I/O Performance: Leveraging Runtime and Offline I/O Opt imization Frameworks\n\nThe existing parallel I/O stack is complex and dif ficult to tune due to the interdependencies among multiple factors that im pact the performance of data movement between storage and compute systems. When performance is slower than expected, end-users, developers, and syst em administrators rely on I/...\n\n\nHammad Bin Ather (University of Orego n)\n---------------------\nQ-NFSO: Exploring Quantum Applications, Noise M anagement, Fault Injection, Resource Scheduling and Optimization in the NI SQ Era\n\nQuantum computing has achieved significant milestones in recent years, underscoring its potential benefits for NP-hard applications both c urrently and in the future. Despite these advancements, contemporary quant um computers are hindered by noise and a limited number of qubits. These l imitations pos...\n\n\nBetis Baheri (Kent State University)\n------------- --------\nAlgorithmic and Optimization Techniques for Graph Applications i n Heterogeneous Systems at Scale\n\nAs heterogeneity becomes commonplace i n HPC systems, algorithmic and optimization techniques are needed to addre ss the challenges that come with it, especially for irregular applications . This includes workload balancing, scheduling, latency tolerance, and mem ory utilization and contention, among ot...\n\n\nReece Neff (North Carolin a State University, Pacific Northwest National Laboratory (PNNL))\n------- --------------\nDesigning Efficient Data Reduction Approaches for Multi-Re solution Simulations on HPC Systems\n\nAs supercomputers advance towards e xascale capabilities, computational intensity increases significantly, and the volume of data requiring storage and transmission experiences exponen tial growth. Multi-resolution methods, such as Adaptive Mesh Refinement (A MR), have emerged as an effective solution ...\n\n\nDaoce Wang (Indiana Un iversity)\n\nRegistration Category: Tech Program Reg Pass, Exhibits Reg Pa ss\n\nSession Chairs: Ayesha Afzal (FAU Erlangen-Nürnberg, Erlangen Nation al High Performance Computing Center); Sally Ellingson (University of Kent ucky); and Alan Sussman (University of Maryland) END:VEVENT END:VCALENDAR