01
Algorithms
Algorithms for computational and data science in parallel and distributed computing environmentsStructured and unstructured mesh and meshless methodsDense and sparse linear algebra computationsSpectral methodsN-body computationsClusteringData miningCompressionCombinatorial algorithms (graph and string algorithms)Algorithms for tightly or loosely coupled systemsCommunication, synchronization, power management, distributed resource management, distributed data and transactions, mobilityNovel algorithm designs and implementations for emerging architectures (ML/AI accelerators, quantum computing systems)
02
Applications
Real-world applications (combinatorial, scientific, engineering, data analysis, visualization) using parallel and distributed computingInnovations originating in target application areas for scalable implementationDesign, implementation, and evaluation of parallel and distributed applicationsImplementations targeting emerging architectures (ML/AI accelerators, quantum computing systems)Application domain advances enabled by ML/AI
03
Architecture
Existing and emerging architectures for high performance computingInstruction-level and thread-level parallelismManycore, multicore, accelerator, domain-specific, and special-purpose architectures (including ML/AI accelerators)Reconfigurable architecturesMemory technologies and hierarchiesVolatile and non-volatile emerging memory technologiesCo-design paradigms for processing-in-memory architecturesSolid-state devicesExascale system designsData center and warehouse-scale architecturesNovel big data architecturesNetwork and interconnect architecturesEmerging technologies for interconnectsParallel I/O and storage systemsPower-efficient and green computing systemsResilience, security, and dependable architecturesEmerging architectural principles for machine learning, approximate computing, quantum computing, neuromorphic, analog, and bio-inspired computing
04
Machine Learning and Artificial Intelligence (ML/AI)
ML/AI training on resource-limited platforms (e.g., edge platforms)Computational optimization methods for AI (pruning, quantization, knowledge distillation)Parallel and distributed learning algorithmsEnergy-efficient methods for ML/AIFederated learningAgentic AIDesign and implementation of ML/AI algorithms on parallel architectures (distributed memory, GPUs, tensor cores, emerging ML/AI accelerators)New ML/AI methods benefiting HPC applications or HPC system managementDesign and development of ML/AI software pipelines (frameworks for distributed training, integration of compression, compiler techniques, DSLs)ML/AI innovations best reviewed by ML/AI experts
05
Measurements, Modeling, and Experiments
Experiments and performance-oriented studies in parallel and distributed computingMetrics related to time, energy, power, accuracy, and resilienceMethods, experiments, and tools for measuring, evaluating, and analyzing performanceDesign and experimental evaluation of applications in simulation and analysisExperiments on novel commercial or research accelerators and architectures (quantum, neuromorphic, non-Von Neumann systems)Innovations in support of large-scale infrastructures and facilitiesExperiences and methods for allocating and managing system and facility resources
06
Programming Models, Compilers, and Runtime Systems
Design of parallel programming models and paradigmsLanguages and compilers supporting parallel modelsRuntime and middleware solutionsFrameworks targeting cloud and distributed systemsApplication frameworks for fault tolerance and resilienceSoftware supporting data management and scalable data analyticsRuntime systems for future novel computing platforms (quantum, neuromorphic, bio-inspired computing)Novel compiler techniques and frameworks leveraging machine learning methods
07
System Software
Storage and I/O systemsSystem software for resource management, job scheduling, and energy efficiencySystem software support for accelerators and heterogeneous HPC computing systemsInteractions between operating system, hardware, and other software layersSystem software for data managementSystem software solutions for ML/AI workloads (e.g., energy-efficient software methods)System software support for fault tolerance and resilienceContainers and virtual machinesSpecialized operating systems and related support for high-performance computingSystem software for future novel computing platforms (quantum, neuromorphic, bio-inspired computing)System software advances enabled by ML/AI