Search NASA⌕ Search

Engineering topics

Allcock, William

Publications and source records attributed to Allcock, William.

2023 Operational Assessment Report: Argonne Leadership Computing Facility

In 2004, the U.S. Department of Energy’s (DOE’s) Advanced Scientific Computing Research (ASCR) program established the Leadership Computing Facility (LCF) with a mission to provide the world’s most advanced computational resources to the open science community. The LCF is a huge investment in the nation’s scientific and technological future, inspired by a growing demand for large-scale computing and its impact on science and engineering. The LCF operates two world-class centers in support of open science at Argonne National Laboratory (Argonne) and at Oak Ridge National Laboratory (Oak Ridge) and deploys diverse machines that are among the most powerful systems in the world today. The LCF ranks among the top U.S. scientific facilities delivering impactful science. The work performed at these centers informs policy decisions and advances innovations in far-reaching areas such as energy assurance, ecological sustainability, and global security. The leadership-class systems at Argonne and Oak Ridge operate around the clock every day of the year. The high level of services these centers provide and the exceptional science they produce justify their existence to the DOE Office of Science and the U.S. Congress

97 MATHEMATICS AND COMPUTING↗

Empowering Scientific Discovery Through Computing at the Advanced Photon Source

This paper explores the challenges and solutions for managing and processing the vast amount of data generated by the Advanced Photon Source (APS), a synchrotron light source facility producing ultra-bright x-rays for diverse scientific domains. With 68 experimental beamlines covering materials research, biology, and more, the APS serves a wide user base across academia, government, and industry. The ongoing upgrade of the APS storage ring and installation of new instruments will amplify data generation and processing demands. This paper discusses the approach to address these demands through automated data processing using standardized workflows that produce faster scientific insights. The APS Data Management System coordinates various data related tasks to manage storage, data transfer, metadata cataloging, data processing, and interfaces with tools provided by Globus. Through integration with the Argonne Leadership Computing Facility (ALCF), APS users can efficiently access high-performance computing resources. Standardized workflows have led to reduced computational burdens on scientists and greater accessibility of high performance computing resources. We demonstrate how standardization and collaboration enable scientists to rapidly convert raw data into meaningful scientific results, establishing a streamlined path from data collection to analysis and ultimately to publication.

Parraga, Hannah↗

DRAS: Deep Reinforcement Learning for Cluster Scheduling in High Performance Computing

Cluster schedulers are crucial in high-performance computing (HPC). They determine when and which user jobs should be allocated to available system resources. Existing cluster scheduling heuristics are developed by human experts based on their experience with specific HPC systems and workloads. However, the increasing complexity of computing systems and the highly dynamic nature of application workloads have placed tremendous burden on manually designed and tuned scheduling heuristics. More aggressive optimization and automation are needed for cluster scheduling in HPC. In this work, we present an automated HPC scheduling agent named DRAS (Deep Reinforcement Agent for Scheduling) by leveraging deep reinforcement learning. DRAS is built on a hierarchical neural network incorporating special HPC scheduling features such as resource reservation and backfilling. An efficient training strategy is presented to enable DRAS to rapidly learn the target environment. Once being provided a specific scheduling objective given by the system manager, DRAS automatically learns to improve its policy through interaction with the scheduling environment and dynamically adjusts its policy as workload changes. We implement DRAS into a HPC scheduling platform called CQGym. CQGym provides a common platform allowing users to flexibly evaluate DRAS and other scheduling methods such as heuristic and optimization methods. Furthermore, the experiments using CQGym with different production workloads demonstrate that DRAS outperforms the existing heuristic and optimization approaches by up to 50%.

97 MATHEMATICS AND COMPUTING↗

RAM as a Network Managed Resource

This paper describes an architecture for and several application examples of using a dynamically allocable RAM pool over the network as part of a deep memory hierarchy. We present four different use cases, including in situ analysis, machine learning, quantum chemistry simulations, and virtualization. In each use case, we modified or implemented software to use a network RAM pool, and then evaluated the performance. The cases of in situ analysis, machine learning, and virtualization demonstrated good performance. While the quantum chemistry experiments exhibited an average slowdown of 53% for all runs, including those where resources were clearly oversubscribed, the scalability results showed promise.

Allcock, William↗

Experience and Practice of Batch Scheduling on Leadership Supercomputers at Argonne

The mission of the DOE Argonne Leadership Computing Facility (ALCF) is to accelerate major scientific discoveries and engineering breakthroughs for humanity by designing and providing world-leading computing facilities in partnership with the computational science community. The ALCF operates supercomputers that are generally amongst the Top 5 fastest machines in the world. Specifically, ALCF is looking for the science that is either too big to run anywhere else, or it would take so long as to be impractical (i.e., “capability jobs”). At ALCF, batch scheduling plays a critical role for achieving a set of site goals within a set of constraints. While system utilization is an important goal at ALCF, its largest mission constraint is to enable extreme scale parallel jobs to take precedence. In this paper, we will describe the specific scheduling goals and constraints, analyze the workload traces collected in 2013–2017 from the 48-rack petascale supercomputer Mira, and discuss the upcoming scheduling challenges at ALCF.

Allcock, William↗

A Job Sizing Strategy for High-Throughput Scientific Workflows

The user of a computing facility must make a critical decision when submitting jobs for execution: how many resources (such as cores, memory, and disk) should be requested for each job? If the request is too small, the job may fail due to resource exhaustion; if the request is too big, the job may succeed, but resources will be wasted. This decision is especially important when running hundreds of thousands of jobs in a high throughput workflow, which may exhibit complex, long tailed distributions of resource consumption. In this paper, we present a strategy for solving the job sizing problem: (1) applications are monitored and measured in user-space as they run; (2) the resource usage is collected into an online archive; and (3) jobs are automatically sized according to historical data in order to maximize throughput or minimize waste. We evaluate the solution analytically, and present case studies of applying the technique to high throughput physics and bioinformatics workflows consisting of hundreds of thousands of jobs, demonstrating an increase in throughput of 10-400 percent compared to naive approaches.

Tovar, Benjamin (ORCID:0000000252942281)↗

DOE High Performance Computing Operational Review (HPCOR): Enabling Data-Driven Scientific Discovery at HPC Facilities

U.S. Department of Energy (DOE) High Performance Computing (HPC) facilities are on the verge of a paradigm shift in the way they deliver systems and services to science and engineering teams. Research projects are producing a wide variety of data at unprecedented scale and level of complexity, with community-specific services that are part of the data collection and analysis workflow. On June 18-19, 2014 representatives from six DOE HPC centers met in Oakland, CA at the DOE High Performance Operational Review (HPCOR) to discuss how they can best provide facilities and services to enable large-scale data-driven scientific discovery at the DOE national laboratories. The report contains findings from that review.

97 MATHEMATICS AND COMPUTING↗