SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
The objective of RAPIDS2 SciDAC Institute for Computer Science and Data is to assist Office of Science (SC) application teams in overcoming computer science and data challenges in the use of DOE supercomputing resources to achieve science breakthroughs.
Marine cloud systems cover a substantial portion of the world’s oceans. Most of these clouds form relatively close to the ocean surface, typically within one to two kilometers, a region referred to by meteorologists as the marine boundary layer. They are composed predominantly of liquid water, although ice particles can occur in mid- and high-latitude marine clouds during winter. In satellite imagery, these clouds appear bright against the darker ocean surface below, reflecting a large fraction of incoming sunlight back into space that would otherwise warm the ocean. Because marine boundary layer clouds cover such an extensive area of the ocean, they exert a significant influence on Earth’s overall transfer of solar energy absorbed by the surface and thermal energy emitted to space, a balance known as the planetary radiation budget. Marine boundary layer clouds are typically thin, and their formation and dissipation depend on a delicate balance between processes acting at the ocean surface below and the warm, dry air above. They are notoriously difficult to simulate accurately in weather forecast models, which often produce too few marine low clouds in midlatitudes and clouds in tropical regions that are excessively bright, meaning they reflect too much solar radiation. The marine boundary layer is frequently characterized by widespread overcast cloud cover that often transitions from a continuous, single-layer deck to more broken cloud fields toward the tropics. These transitions typically proceed through an intermediate stage in which shallow, broken clouds form beneath the overlying stratiform cloud deck. Once broken clouds develop below the overcast, they frequently self-organize into cloud clusters known as marine boundary layer convective complexes (MBLCCs), although the mechanisms governing the formation and organization of MBLCCs remain poorly understood. Accurately representing these transitions in long-range weather forecast models is essential because they influence the properties of air masses advected over the continental United States and Europe, and they become increasingly important for forecasts on seasonal and longer timescales. We employed two complementary approaches to investigate the processes controlling MBLCCs and their impact on marine cloud cover. Long-term observations from the U.S. Department of Energy’s Eastern North Atlantic (ENA) Observatory provided a unique dataset that allowed us to characterize fundamental properties of MBLCCs, including their typical size and frequency of occurrence. These observations were combined with high-resolution numerical simulations performed on supercomputers to examine the evolution of MBLCCs during cold-air outbreaks over the ENA region.
Phase-change materials play a critical role in industrial energy storage applications to drive efficiency improvements, thermal energy management, and carbon emissions reductions. Recently, it has been shown that rapid solidification of alloys with metastable immiscibility in the liquid phase has the potential to form unique microstructures in which a low-melting phase is uniformly distributed in a high-melting matrix. This feature can be exploited using additive manufacturing to produce components with complex geometries containing such unique phase-change microstructures. Phase-field simulations utilizing high-performance computing were used to provide a detailed description of the evolution of the active phase during service in terms of their morphology and composition in different polycrystalline matrix grain morphologies that are typically produced during additive manufacturing. Phase field simulations were performed using, MEUMAPPS-SL (Microstructure Evolution Using Massively Parallel Phase-field Simulations – Solid Liquid) code that was developed in-house by the Oak Ridge National Laboratory. The simulations utilized the capabilities of the Kestrel supercomputer at the National Renewable Energy Laboratory. The simulation results were compared with experimental results generated at Siemens Energy, Inc. The results indicate that the kinetics of liquid spreading along grain boundaries is largely determined by the mobility of the triple line along the intersection of the grain boundary liquid and the grain boundary plane.
This report summarizes the activities, technical accomplishments, and outcomes of the RAPIDS2 Institute project at the University of Delaware (UD). The RAPIDS2 Institute was a large multi-institution project with the objective of assisting SciDAC and Office of Science application teams in the use of DOE supercomputing resources to achieve scientific breakthroughs. The UD team contributed to this effort through work on formal software verification. This thrust aims to reduce software developer time and effort, especially regarding debugging and testing, and to increase confidence in the correctness of the results computed by the software.
Livermore’s Computing Grand Challenge Program enters its 20th year with more unclassified high-performance computing (HPC) power than ever before. This unique, peer-reviewed competition awards HPC allocations on top supercomputers to multidisciplinary teams with high-impact projects. The Grand Challenge encourages researchers to innovate, pushes scientific discovery to new heights, improves the Laboratory’s HPC capabilities, and extends HPC accessibility to collaborators. Awardees must adapt to successive generations of HPC hardware and learn to run simulations at scale. The feature article spotlights three Grand Challenge teams whose research broke new ground in key scientific pursuits—the essence of dark matter, explosion-generated seismic waves, and protein interactions linked to cancer—while underscoring the importance of academic partnerships and considering the program’s future.
As Lawrence Livermore National Laboratory’s (LLNL’s) Laboratory Directed Research and Development (LDRD) program enters its fifth decade of leading-edge research and development, its impact and importance have never been stronger. The program continues to advance strategic investments in pioneering science, technology, and engineering, ensuring LLNL will be ready to deliver on our mission as it evolves over the coming decades. Investing in LDRD research, and the people who perform this critical work, gives LLNL the ability to sustain our role as a leader in the Department of Energy and National Nuclear Security Administration enterprise. The LDRD program enables high-risk, high-payoff research that anticipates emerging threats and future mission needs. By nurturing the ingenuity of the Lab’s greatest asset, its people, LDRD funding advances not only our research but also grows and nurtures our workforce: engaging future innovators with student mentoring, challenging postdoctoral researchers to apply their skills to support national security, and strengthening the leadership skills of early career staff. This annual report documents how LDRD investments advance LLNL’s science, technology, and engineering across our mission space. To assess LDRD’s impact we track both short and long-term metrics such as peer-reviewed publications, number of students, or professional fellows. In addition to reviewing these metrics, I encourage you to delve deeper into the breadth of science and technology that illustrate the strategic value of this research portfolio. For instance, a recent exploratory research project used advanced manufacturing to construct miniaturized three-dimensional ion traps for a quantum computer with reduced quantum error rates to enable applications that address national security missions and support basic science. Another project has delved into studying detonation by examining deflagration to enhance the safety and security of the nuclear weapons stockpile. LDRD researchers are also deploying AI agents on two of the world’s most powerful supercomputers to automate and accelerate inertial confinement fusion experiments. Other teams are delivering more accurate optical constants to enable improved validation for aluminum to advance atomic and molecular physics models. LDRD-driven discoveries of how metals deform under extreme conditions strengthen our ability to model and design materials for demanding national security environments. National security challenges are increasingly complex and continuously evolving. LDRD focuses our most innovative science and technology on these challenges, ensuring the Laboratory is developing creative, forward-leaning solutions for our nation and the world. The following pages feature highlights of published scientific advances, patents, and honors that stem from LDRD investments. As you read this report, I hope you will understand how these investments position the Laboratory, and our partners, to meet the demands of the decades ahead.
Omega version 0, the first step toward a new ocean model, has been developed to take full advantage of next-generation supercomputers, including those powered by GPUs. Omega is rebuilt from the ground up, redesigning computational aspects of ocean models to run faster and more efficiently across diverse high-performance computing systems.
This document serves as the user manual for HydraGNN v5.0, a scalable graph neural network (GNN) architecture for simultaneous prediction of multiple target properties using multi-task learning (MTL). This version of HydraGNN has been developed primarily to support the development, training, and deployment of predictive graph-based deep learning (DL) models for atomistic materials modeling. HydraGNN is templated over 13 message-passing policies, including invariant models (GIN, PNA, PNAPlus, GAT, MFC, CGCNN, SAGE, SchNet, DimeNet) and equivariant models (EGNN, PNAEq, PAINN, MACE), and supports distributed training via distributed data parallelism (DDP), DeepSpeed, and Fully Sharded Data Parallelism (FSDP) on leadership-class supercomputers. Although HydraGNN can be applied to problems beyond atomistic materials modeling, its current use is confined to homogeneous graphs. Additional capabilities include machine-learned interatomic potentials with energy-conserving forces, General, Powerful, and Scalable Graph Transformer (GraphGPS) global attention, periodic boundary conditions, hyperparameter optimization, mixed-precision training, and uncertainty quantification.
We benchmark the DAOS filesystem on Argonne's Aurora supercomputer (127 nodes, 4,064 targets) using fio, IOR, mdtest, and IO500 to characterize I/O and metadata performance across the DFS API and DFuse+POSIX. Single-client fio shows POSIX bandwidth saturating at 1–2 MiB I/O sizes, with write-heavy workloads outperforming reads. Multi-node IOR shows DFS bandwidth scaling well up to ~32 tasks/node, with write latency growing faster than read latency. An 8-node IO500 evaluation shows DFS achieving ~5x higher bandwidth and ~190x higher IOPS than POSIX. Results indicate DAOS is well-suited to read-heavy workloads like AI training data loading, given appropriately sized transfers and concurrency.
FullWave is a computer code that simulates how radio-frequency (RF) waves travel and deposit energy in the hot plasma inside a fusion reactor. RF waves are used to heat the plasma and drive electrical current, which is essential for sustaining fusion reactions. The code uses a new algorithm that can handle much finer spatial detail than previous codes — more than 100 times finer — while running efficiently on national supercomputers. It incorporates a detailed physics model that captures subtle kinetic effects important for accurate prediction of wave behavior. Under this project, FullWave was extended to cover multiple RF frequency ranges relevant to present and future tokamaks, and validated against experimental parameters from the DIII-D tokamak at General Atomics. Results were published in peer-reviewed journal articles.
The SIMBER project is centered around advancing the state-of-the-art in the science of modelling of the human heart and leveraging the collaborations between leading research groups at the University of California Berkeley (UC Berkeley), the Lawrence Berkeley National Laboratory (Berkeley Lab), and Simula Research Laboratory (Simula). In particular, the following areas of expertise are shared and expanded through this project: “heart-on-chip” experimental systems from UC Berkeley, mathematical modelling of the heart from Simula, and supercomputing software from Simula and Berkeley Lab. This intersection of expertise is already enabling the development of novel tools and knowledge that produce more accurate models of the human heart in both health and disease, which in turn are leading to drug screening technologies that make cardiac drug development faster, cheaper and more humane.
This final technical report summarizes research conducted under DOE Award DE-SC0021653 to develop unitary Quantum Lattice Algorithms for modeling electromagnetic wave propagation and scattering in complex media, including plasmas. The project developed and validated quantum-inspired formulations of Maxwell's equations that preserve unitary evolution and can be evaluated on classical high-performance computing systems while providing a foundation for future quantum-computing implementations. Major accomplishments include the development of two- and three-dimensional algorithms for electromagnetic scattering; scalable, distributed-memory implementations demonstrated on the Perlmutter supercomputer; formulations for nonlinear lossless fluid dynamics and cold, lossless, inhomogeneous magnetized plasmas; and an explicit quantum algorithm for a time-discretized Lorenz model. Simulations reproduced a range of characteristic wave phenomena, including transient effects that are not readily apparent in conventional frequency-domain studies, demonstrating the effectiveness of the proposed approach for modeling complex electromagnetic and plasma systems. The work establishes a unified theoretical and computational framework for quantum and quantum-inspired simulation and provides a foundation for future implementation on fault-tolerant quantum systems.
Viskores is a scientific visualization library that is the primary deployment of such algorithms to the parallel accelerated processors of modern DOE supercomputers. In this paper, we review the capabilities provided by Viskores and how these capabilities are leveraged by other software in the high-performance computing ecosystem. We discuss the Viskores data representation and pay particular attention to array management. Through this array management we describe how data is adapted between Viskores and other software along with strategies for converting dynamic, polymorphic objects to static representations better suited to GPU processing. We conclude with several examples of Viskores integrating with high-performance software that is used in production today.
Recent advancements in AI models, like ChatGPT, are structured to retain user interactions, which could inadvertently include sensitive healthcare data. In the healthcare field, particularly when radiologists use AI-driven diagnostic tools hosted on online platforms, there is a risk that medical imaging data may be repurposed for future AI training without explicit consent, spotlighting critical privacy and intellectual property concerns around healthcare data usage. Addressing these privacy challenges, a novel approach known as Unlearnable Examples (UEs) has been introduced, aiming to make data unlearnable to deep learning models. A prominent method within this area, called Unlearnable Clustering (UC), has shown improved UE performance with larger batch sizes but was previously limited by computational resources (e.g., a single workstation). To push the boundaries of UE performance with theoretically unlimited resources, we scaled up UC learning across various datasets using Distributed Data Parallel (DDP) training on the Summit supercomputer. Our goal was to examine UE efficacy at high-performance computing (HPC) levels to prevent unauthorized learning and enhance data security, particularly exploring the impact of batch size on UE’s unlearnability. Utilizing the robust computational capabilities of the Summit, extensive experiments were conducted on diverse datasets such as Pets, MedMNist, Flowers, and Flowers102. Our findings reveal that both overly large and overly small batch sizes can lead to performance instability and affect accuracy. However, the relationship between batch size and unlearnability varied across datasets, highlighting the necessity for tailored batch size strategies to achieve optimal data protection. The use of Summit’s high-performance GPUs, along with the efficiency of the DDP framework, facilitated rapid updates of model parameters and consistent training across nodes. Our results underscore the critical role of selecting appropriate batch sizes based on the specific characteristics of each dataset to prevent learning and ensure data security in deep learning applications. The source code is publicly available at https: // github. com/ hrlblab/ UE_ HPC .
In HPC I/O middleware like the Adaptable I/O System (ADIOS) often mediates data transfers between applications. The metadata I/O generated by such systems often presents significant scaling and performance limitations. This work seeks improvement opportunities for metadata I/O by leveraging the DAOS storage systems, a recent storage system solution deployed on high-end systems such as the Aurora supercomputer. We investigate the tradeoffs and the design space for integrating I/O engines for the ADIOS middleware based on the different storage mechanisms supported by DAOS. We present a new DAOS-Array-ChunkSize-aligned engine which provides up to 2.3× improved performance than when using the existing DAOS-POSIX interface, without requiring any application modifications.
Large language models (LLMs) have demonstrated remarkable success as foundational models, benefiting various downstream applications through fine-tuning. Loss scaling studies have demonstrated the superior performance of larger LLMs compared to their smaller counterparts. Nevertheless, training LLMs with billions of parameters poses significant challenges and requires considerable computational resources. For example, training a one trillion parameter GPT-style model on 20 trillion tokens requires a staggering 120 million exaflops. This research explores efficient distributed training strategies to extract this computation from Frontier, the world's first exascale supercomputer. We enable and investigate various model and data parallel training techniques, such as tensor parallelism, pipeline parallelism, and sharded data parallelism, to facilitate training a trillion-parameter model on Frontier. We empirically assess these techniques and their associated parameters to determine their impact on memory footprint, communication latency, and GPU's computational efficiency. We analyze the complex interplay among these techniques and find a strategy to combine them to achieve high throughput through hyperparameter tuning. We have identified efficient strategies for training large LLMs of varying sizes through empirical analysis and hyperparameter tuning. For 22 Billion, 175 Billion, and 1 Trillion parameters, we achieved GPU throughputs of 38.38%, 36.14%, and 31.96%, respectively. For the training of the 175 Billion parameter model and the 1 Trillion parameter model, we achieved 100% weak scaling efficiency on 1024 and 3072 Mi250X GPUs, respectively. We also achieved strong scaling efficiencies of 89% and 87% for these two models. We trained these models only tens of iterations instead of training till completion.
High-performance computing (HPC) applications have traditionally relied on parallel file systems and file transfer services to manage data movement and storage. Alternative approaches have been proposed that use direct communications between application components, trading persistence and fault tolerance for speed. Event-driven architectures, as popularized in enterprise contexts, present a compelling middle ground, avoiding the performance cost and API constraints of parallel file systems while retaining persistence and offering impedance matching between application components. However, adapting streaming frameworks to HPC workloads requires addressing challenges unique to HPC systems. This paper investigates the potential for a streaming framework designed for HPC infrastructures and use cases. We introduce Mofka, a persistent event-streaming framework designed specifically for HPC environments. Mofka combines the capabilities of a traditional streaming service with optimizations tailored to the HPC context, such as support for massively multicore nodes, efficient scaling for large producer-consumer workflows, RDMA-enabled high-performance network communications, specialized network fabrics with multiple links per node, and efficient handling of large scientific data payloads. Built using the Mochi suite of HPC data service components, Mofka provides a lightweight, modular, and high-performance solution for persistent streaming in HPC systems. We present the architecture of Mofka and evaluate its performance against Kafka and Redpanda using benchmarks on diverse platforms, including Argonne's Polaris and Oak Ridge's Frontier supercomputers, showing up to 8× improvement in throughput in some scenarios. We then demonstrate its utility in several real-world applications: a tomographic reconstruction pipeline, a workflow for the discovery of metal-organic frameworks for carbon capture, and the instrumentation of Dask workflows for provenance tracking and performance analysis.
The Texas Gulf Coast region contains significant centers of population, infrastructure, and economy and is threatened by intensifying tropical storms. The flooding from these tropical storms often has multiple compounding drivers. This characteristic presents a complex numerical problem where a simulation must consider multiple hydrologic forcings. While several procedures exist for addressing this problem numerically, they tend to be resource-intensive and cannot be conducted in near real-time. We extend GeoFlood, a reduced physics approach for fluvial flood forecasting, to rapidly predict coastal and compound fluvial-coastal inundation. This method is validated against a numerical ocean circulation model (ADCIRC) simulation of Hurricane Ike, a major coastal flooding event that happened on the Texas Gulf Coast in 2008. We show that the inundation map generated by coastal HAND (c-HAND) has reasonable agreement with the ADCIRC simulation while taking about 1.7% of the time currently needed to run ADCIRC on a supercomputer. While our model correctly predicts 99% of ADCIRC-inundated DEM cells, it also overpredicts inundated area by a factor of approximately 27%. We combine c-HAND with the GeoFlood framework for fluvial flood forecasting to create a compound fluvial-coastal inundation mapping workflow that can be run in near real-time. c-HAND's fast wall-clock time and low CPU requirements can support decision making by first response personnel. The method provides timely and convenient access to crucial information, such as the locations of flooded roads and inundated coastal areas.