arXiv subjects

Mason Hooten

Publications and source records attributed to Mason Hooten.

3 recordsLinked to original sources

arXiv2025

RHAPSODY: Execution of Hybrid AI-HPC Workflows at Scale

Hybrid AI-HPC workflows combine large-scale simulation, training, high-throughput inference, and tightly coupled, agent-driven control within a single execution campaign. These workflows impose heterogeneous and often conflicting requirements on runtime systems, spanning MPI executables, persistent AI services, fine-grained tasks, and low-latency AI-HPC coupling. Existing systems typically address only subsets of these requirements, limiting their ability to support emerging AI-HPC applications at scale. We present RHAPSODY, a multi-runtime middleware that enables concurrent execution of heterogeneous AI-HPC workloads through uniform abstractions for tasks, services, resources, and execution policies. Rather than replacing existing runtimes, RHAPSODY composes and coordinates them, allowing simulation codes, inference services, and agentic workflows to coexist within a single job allocation on leadership-class HPC platforms. We evaluate RHAPSODY with Dragon and vLLM on multiple HPC systems using representative heterogeneous, inference-at-scale, and tightly coupled AI-HPC workflows. Our results show that RHAPSODY introduces minimal runtime overhead, sustains increasing heterogeneity at scale, achieves near-linear scaling for high-throughput inference workloads, and data- and control-efficient coupling between AI and HPC tasks in agentic workflows.

cs.DC↗

arXiv2025

Applications of Enhanced Sampling Methods to Biomolecular Self-Assembly: A Review

This review article discusses some common enhanced sampling methods in relation to the process of self-assembly of biomolecules. An introduction to self-assembly and its challenges is covered followed by a brief overview of the methods and analysis for replica-exchange molecular dynamics, umbrella sampling, metadynamics, and machine learning based techniques. Applications of select methods towards peptides, proteins, polymers, nucleic acids, and supramolecules are discussed. Finally, a short discussion of the future directions of some of these methods is provided.

cond-mat.soft↗

arXiv2022

Pipeline for Automating Compliance-based Elimination and Extension (PACE2): A Systematic Framework for High-throughput Biomolecular Material Simulation Workflows

The formation of biomolecular materials via dynamical interfacial processes such as self-assembly and fusion, for diverse compositions and external conditions, can be efficiently probed using ensemble Molecular Dynamics. However, this approach requires a large number of simulations when investigating a large composition phase space. In addition, there is difficulty in predicting whether each simulation is yielding biomolecular materials with the desired properties or outcomes and how long each simulation will run for. These difficulties can be overcome by rules-based management systems which include intermittent inspection, variable sampling, premature termination and extension of the individual Molecular Dynamics simulations. The automation of such a management system can significantly reduce the overhead of managing large ensembles of Molecular Dynamics simulations. To this end, a high-throughput workflows-based computational framework, Pipeline for Automating Compliance-based Elimination and Extension (PACE2), for biomolecular materials simulations is proposed. The PACE2 framework encompasses Simulation-Analysis Pipelines. Each Pipeline includes temporally separated simulation and analysis tasks. When a Molecular Dynamics simulation completes, an analysis task is triggered which evaluates the Molecular Dynamics trajectory for compliance. Compliant Molecular Dynamics simulations are extended to the next Molecular Dynamics phase with a suitable sample rate to allow additional, detailed analysis. Non-compliant Molecular Dynamics simulations are eliminated, and their computational resources are either reallocated or released. The framework is designed to run on local desktop computers and high performance computing resources. In the future, the framework will be extended to address generalized workflows and investigate other classes of materials.

physics.comp-ph↗