SEARCH · Engineering Papers

Results for “hardware accelerators”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 685 records · Page 38

All sources NASA NTRS DOE OSTI

NASA NTRSDate not supplied

A Multi-Architecture Approach for Implicit Computational Fluid Dynamics on Unstructured Grids

High-performance computing (HPC) architectures are trending toward manycore paradigms such as graphics processing units (GPUs). Approximately half of the top 100 publicly disclosed supercomputers in the world utilize GPU accelerators for performance. This is in contrast to a decade ago, where there were only a few such machines in the top 100. It is not currently possible to compile and run legacy central processing unit (CPU) software efficiently on GPUs without significant refactoring. Though a number of frameworks offering performance portability exist, none offer a standardized specification that is supported by all major hardware vendors. Additionally, experiences show that obtaining a high percentage of peak performance often requires architecture-specific code. This work details a pragmatic multi-architecture computational fluid dynamics library focused on aerospace problems across the speed range from low subsonic to hypersonic flows involving thermochemical nonequilibrium. A thin abstraction layer above NVIDIA CUDA C++ is utilized, which enables primarily single-source software currently capable of running efficiently on multicore CPUs, NVIDIA GPUs, AMD GPUs, and Intel GPUs. Results on various problems of interest across the speed range are presented and performance is compared between various architectures.

GPU↗

NASA NTRSDate not supplied

A Multi-Architecture Approach for Implicit Computational Fluid Dynamics on Unstructured Grids

GPU↗

DOE OSTI2025

Fast, differentiable, and extensible big bang nucleosynthesis package

Here, we introduce light isotope nucleosynthesis with JAX (LINX), a new differentiable public big bang nucleosynthesis code designed for fast parameter estimation. By leveraging JAX, LINX achieves both speed and differentiability, enabling the use of Bayesian inference, including gradient-based methods. We discuss the formalism used in LINX for rapid primordial elemental abundance predictions and give examples of how LINX can be used. When combined with differentiable cosmic microwave background power spectrum emulators, LINX can be used for joint cosmic microwave background and big bang nucleosynthesis analyses without requiring extensive computational resources, including on personal hardware.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

DOE OSTI2024

New Advances in Optical Stochastic Cooling

Recently, Optical Stochastic Cooling (OSC) became the first demonstrated method for ultra-high-bandwidth stochastic cooling. The initial experiments at Fermilab’s IOTA ring explored the essential physics of the method and demonstrated cooling, heating and manipulation of beams and single particles. Having been validated in practice, with continued development, OSC carries the potential for dramatic advances in the state-of-the-art performance and flexibility for beam cooling and control. The ongoing program at Fermilab is now focused on the development of an OSC system that includes high-gain optical amplification, which promises a two-order-of-magnitude increase in the strength of the OSC force. In this talk, we briefly review the results of the initial experimental campaign, describe the status of the conceptual and hardware designs for the amplified OSC system, report initial experimental results of our high-gain amplifier development, and explore near-term operational plans and use cases.

43 PARTICLE ACCELERATORS↗

DOE OSTI2021

Performance Analysis of PIConGPU: Particle-in-Cell on GPUs using NVIDIA’s NSight Systems and NSight Compute

PIConGPU, Particle In Cell on GPUs, is an open source simulations framework for plasma and laser-plasma physics used to develop advanced particle accelerators for radiation therapy of cancer, high energy physics and photon science. While PIConGPU has been optimized for at least 5 years to run well on NVIDIA GPU-based clusters, there has been limited exploration by the development team of potential scalability bottlenecks using recently updated and new tools including NVIDIA’s NVProf tool and the brand-new NVIDIA NSight Suite (Systems and Compute) tools. PIConGPU is a highly optimized application that runs production jobs at scale on a system Oak Ridge Leadership Facility’s (OLCF) Summit supercomputer (using the full machine at 4600 nodes; at 98% of GPU utilization on all ~28000 NVIDIA Volta GPUs). PIConGPU has been selected as one of the the eight applications for OLCF’s coveted Center for Accelerated Application Readiness (CAAR) program aimed at the facility’s Frontier supercomputer (OLCF’s first exascale system to launch in 2021), to partner with our vendors (primary vendors: AMD and Cray/HPE) ensuring that Frontier will be able to perform large-scale science when it opens to users in 2022. To this effect, performance engineers on the PIConGPU team wanted to dive deep into the application to understand at the finest granularity, which portions of the code could be further optimized to exploit the hardware on Summit at it’s maximum potential and also to elucidate which key kernels should be tracked and optimized for the CAAR effort to port this code to Frontier. Any bottlenecks that are observed via performance profiling on Summit are likely to also impact scalability on the Frontier-dev system and the Frontier Early Access (EA) system. Additionally, the engineers wanted to take a closer look at the newest NVIDIA profiling tools which allows us to identify the most useful features on these tools and will provide an opportunity to compare it to new AMD and Cray’s performance analysis tool releases and provide feedback to our vendor partners on what features are most important and mission critical for CAAR efforts. The primary goal of this report is to focus on the evaluation of PIConGPU’s most time-intensive kernels using NVProf and NSight Suite. Three kernels, Current Deposition (also known as Compute Current), Particle Push (Move and Mark), and Shift Particles are known to be some of the most time-consuming kernels in PIConGPU. The Current xi Deposition kernel and Particle Push kernel both set up the particle attributes for running any physics simulation with PIConGPU, so it is crucial to improve the performance of these two kernels. In this report, we measure single GPU metrics for the three kernels, offer high level takeaways from the conducted analysis, and compare the profiling data from NSight Compute to that of NVProf. This analysis was performed using a grid size of 240 x 272 x 224, and 10 time steps with the Mid-November Figure of Merit (FOM) run setup. The Traveling Wave Electron Acceleration (TWEAC) science case used in this run is a representative science case for PIConGPU. This execution can also be used for baseline analysis on AMD MI50/ MI60 systems. As of the time of writing, the PIConGPU application has limited use for features of NSight Systems, so this report will mainly focus on insights garnered from NSight Compute. For this analysis, we run the “full” metric set available in NSight Compute version 2020.1.2 and use NSight Systems version 2020.3.1 to generate the application timeline.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

DOE OSTI2023

Revisiting Temporal Blocking Stencil Optimizations

Iterative stencils are used widely across the spectrum of High Performance Computing (HPC) applications. Many efforts have been put into optimizing stencil GPU kernels, given the prevalence of GPU-accelerated supercomputers. To improve the data locality, temporal blocking is an optimization that combines a batch of time steps to process them together. Under the observation that GPUs are evolving to resemble CPUs in some aspects, we revisit temporal blocking optimizations for GPUs. We explore how temporal blocking schemes can be adapted to the new features in the recent Nvidia GPUs, including large scratchpad memory, hardware prefetching, and device-wide synchronization. We propose a novel temporal blocking method, EBISU, which champions low device occupancy to drive aggressive deep temporal blocking on large tiles that are executed tile-by-tile. We compare EBISU with state-of-the-art temporal blocking libraries: STENCILGEN and AN5D. We also compare with state-of-the-art stencil auto-tuning tools that are equipped with temporal blocking optimizations: ARTEMIS and DRSTENCIL. Over a wide range of stencil benchmarks, EBISU achieves speedups up to 2.53x and a geometric mean speedup of 1.49x over the best state-of-the-art performance in each stencil benchmark.

Zhang, Lingqi↗

DOE OSTI2021

Hexagonal Geometries in MPACT

The MPACT code is a high-fidelity light-water reactor analysis code using whole-core pin-resolved neutron transport calculations on modern parallel-computing hardware. MPACT uses the 2D/1D method to solve 3D neutron transport problems by decomposing the problem into a stack of 2D slices, each of which is solved independently using the method of characteristics (MOC). The slices are then coupled axially using the P3 nodal expansion method (NEM-P3) for the 1D axial calculations. MPACT also employs the coarse mesh finite difference (CMFD) method to accelerate calculations. This manuscript details work supporting advanced reactor designs using hexagonal pins and hexagonal assemblies such as the VVER-1000. If performed correctly, MOC is geometry agnostic. However, MPACT previously had optimizations in place for Cartesian geometries, specifically in the modularization and current calculations. Sections 2 and 3 detail the changes made to MPACT to support MOC and CMFD calculations on hexagonal geometries. Section 4 reports results demonstrating solution consistency for problems run with and without CMFD acceleration, results demonstrating solution consistency when run in serial and parallel, and pincell results using the Monte Carlo code, McCard’s benchmark results.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2025

Digital Twin Framework for PIP-II Linac: AI-Driven Multi-Scale Modeling from Ion Source to 800 MeV

The PIP-II linac will enable >1.2 MW beam power for DUNE, requiring unprecedented operational reliability across its warm front-end (RFQ, MEBT) and five distinct SRF sections operating at 162.5/325/650 MHz. We present a comprehensive digital twin framework uniquely combining a fully differentiable fast beam transport code with neural network surrogates trained on high-fidelity PIC simulations, capturing space charge and nonlinear dynamics beyond traditional envelope codes while achieving 10⁴ speedup at <1% accuracy. End-to-end differentiability enables gradient-based optimization across 500+ parameters simultaneously previously impossible with conventional tools while the model incorporates static/dynamic errors and serves as a virtual commissioning platform for diverse hardware integration. The framework facilitates reinforcement learning for pulsed/CW mode transitions, predictive maintenance through anomaly detection, and autonomous tuning algorithm development with real-time execution capability. Validation against physics simulations shows excellent agreement for the front-end, with initial results demonstrating potential for 30% commissioning time reduction and proactive fault mitigation, providing a scalable blueprint for operating next-generation high-intensity accelerators.

Pathak, Abhishek [Fermilab] (ORCID:000000021704208↗

NASA NTRS2010

Modeling Flight: The Role of Dynamically Scaled Free-Flight Models in Support of NASA's Aerospace Programs

The state of the art in aeronautical engineering has been continually accelerated by the development of advanced analysis and design tools. Used in the early design stages for aircraft and spacecraft, these methods have provided a fundamental understanding of physical phenomena and enabled designers to predict and analyze critical characteristics of new vehicles, including the capability to control or modify unsatisfactory behavior. For example, the relatively recent emergence and routine use of extremely powerful digital computer hardware and software has had a major impact on design capabilities and procedures. Sophisticated new airflow measurement and visualization systems permit the analyst to conduct micro- and macro-studies of properties within flow fields on and off the surfaces of models in advanced wind tunnels. Trade studies of the most efficient geometrical shapes for aircraft can be conducted with blazing speed within a broad scope of integrated technical disciplines, and the use of sophisticated piloted simulators in the vehicle development process permits the most important segment of operations the human pilot to make early assessments of the acceptability of the vehicle for its intended mission. Knowledgeable applications of these tools of the trade dramatically reduce risk and redesign, and increase the marketability and safety of new aerospace vehicles. Arguably, one of the more viable and valuable design tools since the advent of flight has been testing of subscale models. As used herein, the term "model" refers to a physical article used in experimental analyses of a larger full-scale vehicle. The reader is probably aware that many other forms of mathematical and computer-based models are also used in aerospace design; however, such topics are beyond the intended scope of this document. Model aircraft have always been a source of fascination, inspiration, and recreation for humans since the earliest days of flight. Within the scientific community, Leonardo da Vinci, George Cayley, and the Wright brothers are examples of early aviation pioneers who frequently used models during their scientific efforts to understand and develop flying machines. Progress in the technology associated with model testing in worldwide applications has firmly established model aircraft as a key element in new aerospace research and development programs. Models are now routinely used in many applications and roles, including aerodynamic data gathering in wind tunnel investigations for the analysis of full-scale aircraft designs, proof-of-concept demonstrators for radical aeronautical concepts, and problem-solving exercises for vehicles already in production. The most critical contributions of aerospace models are to provide confidence and risk reduction for new designs and to enhance the safety and efficiency of existing configurations.

Chambers, Joseph↗

NASA NTRS2019

Trends in Human Spaceflight: Failure Tolerance, High Reliability and Correlated Failure History

In a half century of human spaceflight, NASA has continuously refined agency safety and reliability requirements in response to mission demands, critical failures, and technology development. Early spacecraft, including Mercury, Gemini and Apollo vehicles, were highly reliant on dissimilar redundancy and demonstrated test margins. Later programs, such as the reusable Space Transportation System (STS) and International Space Station (ISS), introduced probabilistic studies and isolated two-failure tolerance to improve robustness at the expense of added complexity. More recently, the Orion Multi-Program Crew Vehicle (MPCV) program adopted universal single-failure tolerance with two categorical exceptions; Zero-Failure Tolerant (0FT) and Design for Minimum Risk (DFMR) hardware. Failure tolerance variances are defined and managed in accordance with agency human-rating requirements, and require concurrence from program Technical Authorities (TA) as well as the MPCV Safety and Mission Assurance Safety and Engineering Review Panel (MSERP). To understand and reaffirm standards applied to Apollo, Space Shuttle and Orion vehicles, Orion and Deep Space Gateway Safety and Mission Assurance (S&MA) representatives conducted accelerated research to compare unique safety and reliability criteria against ground and flight anomalies, based on information contained in post-mission reports and the Problem Reporting and Corrective Action (PRACA) database. In some cases, high-profile failures and narrow escapes have reinforced decisions to maintain or adapt safety requirements. In others, empirical trends have highlighted the need for vigilance and innovative safety guidelines. Given the inability to achieve absolute compliance with evolving safety and reliability requirements, the team conducted a targeted review of DFMR and 0FT propulsion elements within the framework of changing system design, inspection, materials and process developments to formulate conclusions on technological maturity, failure density, and net changes in safety risk. Based on the aggregate performance of high-reliability and failure-tolerant systems, the authors have attempted to establish best practices and guidelines to inform future program decisions. On a somewhat cautionary note, this study is not intended to direct a universal set of requirements for future missions based on prior lessons learned. Spacecraft safety is a multi-variable problem, and attempts to mitigate past failures will not guarantee future success. However, this assessment offers a retrospective review of policy changes, implementation and effectiveness. In the future, NASA, European Space Agency (ESA) and industry partners may benefit from a more robust correlation between requirements and performance, as space-faring nations work toward more challenging, complex and long-duration commercial and deep-space ventures.

Green, Carrie↗

DOE OSTI2025

Digital Twin Framework for PIP-II Linac: AI-Driven Multi-Scale Modeling from Ion Source to 800 MeV

The PIP-II linac will enable >1.2 MW beam power for DUNE, requiring unprecedented operational reliability across its warm front-end (RFQ, MEBT) and five distinct SRF sections operating at 162.5/325/650 MHz. We present a comprehensive digital twin framework uniquely combining a fully differentiable fast beam transport code with neural network surrogates trained on high-fidelity PIC simulations, capturing space charge and nonlinear dynamics beyond traditional envelope codes while achieving 10⁴× speedup at <1% accuracy. End-to-end differentiability enables gradient-based optimization across 500+ parameters simultaneously—previously impossible with conventional tools—while the model incorporates static/dynamic errors and serves as a virtual commissioning platform for diverse hardware integration. The framework facilitates reinforcement learning for pulsed/CW mode transitions, predictive maintenance through anomaly detection, and autonomous tuning algorithm development with real-time execution capability. Validation against physics simulations shows excellent agreement for the front-end, with initial results demonstrating potential for 30% commissioning time reduction and proactive fault mitigation, providing a scalable blueprint for operating next-generation high-intensity accelerators.

Pathak, Abhishek [Fermilab] (ORCID:000000021704208↗

DOE OSTI2020

CSB-RNN: A Faster-Than-Realtime RNN Acceleration Framework with Compressed Structured Blocks

Recurrent Neural Networks (RNN) is widely applied to temporal sequence analysis, where real-time performance is usually in demand. However, RNN suffers a heavy computational workload as the model comes with a large weight matrix. To alleviate the pain, model compression (pruning) schemes have been proposed for RNN that pruning the redundant (near-zero) weight-values. On the one hand, the non-structured pruning methods achieve a considerable pruning rate while bringing the computational irregularity, which is un-friendly to parallel-hardware. On the other hand, the existing structured pruning methods consider the hardware parallelism; However, they suffer a poor pruning rate due to the restrict constraints on pruning structure. This paper presents CSB-RNN, an optimized full-stack RNN framework with the novel compressed structured block (CSB) technique. The CSB-pruned RNN model comes with both fine-granularity that benefits the pruning rate and regular structure that facilitates the hardware-parallelism. Further, we propose a novel hardware architecture for inferencing the CSB-pruned model. Different from conventional parallel hardware, this architecture solves the block-workload imbalance issue and achieves an over 95% hardware utilization. With the experiments on 10 RNN models in 5 application domains, the CSB-RNN realizes 7×-20× lossless compression and up to 50× acceptable lossy-compression, which is 2×-7× to the prior art. With the addition of the novel hardware, the compressed-RNN inference reaches a super real-time latency of 10-400µs with FPGA implementation.

Shi, Runbin↗

DOE OSTI2022

Converter-Interfaced CHP Plant for Improved Grid-Integration, Flexibility and Resiliency

GE Research and its partner GE Renewables have proposed the use of an interface converter solution to increase the penetration of small to medium-sized CHP (1MWe to 20MWe) into distribution grids and improve their flexibility and grid support capability. Indeed, the proposed interface converter solution thanks to presence of the grid-ready inverter, allows to streamline the compliance to grid codes requirements, reduce the interconnection delays and costs and ultimately one of the main barriers for CHP adoption by commercial and industrial facilities. An additional benefit provided by the interface converter is the use of the grid-ready inverter for reactive power which eliminates the need of sizing the generator for that capability. These two benefits highly favor the economic feasibility of converter-interfaced CHP. Five user cases, each in one of the leading U.S states for CHP potential reported by the DOE in its estimation of the U.S Technical Potential of CHP, were selected to compare the economic performances of converter-interfaced CHP as compared with directly-coupled. They include a college campus in California, a hospital in New York, a water reclamation plant in Texas, a hotel in Minnesota, and a large office building in Pennsylvania. Results showed that, the presence of the interface converter allows to increase the return on investment (ROI) by 0.5 to 2 percentage points in most of the cases (4 of 5). Indeed, the interface converter by shortening the interconnection process allows to accelerate revenues while reducing interconnection costs. Added to the reduced cost of the required generator these savings trade favorably the capital cost of the converter. The analysis also showed that the profitability of the converter-interfaced CHP is highly sensitive to the energy price, interconnection delay, and converter cost. However, it appears that if the interface converter can shorten the interconnection process by at least 6 months, adopting this solution will be more economically viable than directly-coupled configuration in almost all the +23,000 sites of the U.S Technical Potential CHP. The evaluation of the benefits of a converter-interfaced CHP also showed that it enables higher ROI when coupled with other distributed energy resources (DER) such as battery energy systems (BESS) or solar photovoltaic (PV). Indeed, in those scenario, the grid-ready inverter included in the interface converter eliminates the need of separate inverters if DC-coupling is used. On the technical performance, it has been verified that the presence of the interface converter allows to reduce by 70% to 80% the CHP short-circuit contribution to grid faults. This not only reduces the mechanical and thermal stresses exposed to the CHP electrical components but also increases the grid hosting capacity which ultimately enables higher penetrations CHP. Another key benefit of the interface converter validated with hardware-in-the-loop simulations and testing is its superior capability for reactive power support. Indeed, using a power hardware testbed with two +700kW inverters configured in back-to-back, a microgrid controller and actual facilities loads it was demonstrated that the presence of the interface converter can help maintain a power factor near ~1 or regulate the voltage to ~1.0pu at the point of common coupling. This benefit can be highly valuable if in the future, due to higher penetration of renewable distributed energy resources (DER), utilities start billing demand charge based on kVA instead of kW as currently. It was also validated that converter-interfaced CHP can dispatch heat and power commands and seamlessly switch between the two modes while consistently controlling the power factor or voltage at PCC. Indeed, the power hardware testing showed that grid-connected converter-interfaced CHP can follow either the power or heat demand while maintaining a unity power factor at converter output. This research proved that the adoption of an interface converter as the solution for interconnection of CHP system into the distribution grid can greatly improve the economic feasibility of small to medium-sized CHP as well as the plant power quality, flexibility and resiliency. Additionally, it allows increased penetrations of CHP into the distribution grid, extends their grid support capability, and facilitates the integration of BESS and PV DER by streamlining their collocation within the same facilities. This ultimately provides an opportunity for commercial and small industrial facilities in the U.S to accelerate their energy transition thanks to the high energy efficiency of CHP systems and its reliable, flexible, and resilient microgrid operation when interconnected with an interface converter.

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2025

Beam optics ramping in underconstrained lattice design: Application to Electron-Ion Collider hadron storage ring cooling section

This paper presents the lattice design and optics ramping strategy for the cooling section of the hadron storage ring at the electron-ion collider. The main challenge is that available tuning knobs exceed beam-optics constraints. Independently optimized injection and top-energy optics often yield disconnected solutions, making interpolation impossible. To address this, we propose two new methods. The first is an intermediate-penalty scheme that ensures ramping path continuity by penalizing constraint violations at intermediate states. The second is a continuation-based approach that adapts high-energy optics to low energy, guided by an adaptive weighting scheme to balance injection and ramping constraints. The solutions meet all beam dynamics and hardware limits. The two methods offer a general strategy for ramping in systems where the solution space is under-constrained and the starting and target configurations are far apart.

43 PARTICLE ACCELERATORS↗

DOE OSTI2021

Materials Challenges and Opportunities for Energy Generation, Conversion, Delivery, and Storage (Applied Energy Tri-Laboratory Consortium Workshop Report)

This report documents the outcomes of the Tri-Laboratory Materials Workshop that was held July 31 and August 1, 2019 to begin addressing the needs, opportunities, and challenges associated with the development, fabrication, and testing of the needed materials and components for integrated hybrid energy systems (i.e., incorporating nuclear, fossil, and renewables for electric and thermal applications). This was accomplished by assembling the research program leads and principal investigators at Idaho National Laboratory (INL), National Energy Technology Laboratory (NETL), and National Renewable Energy Laboratory (NREL), who support the research and development of new technology and system integration. The team then identified and prioritized key materials development needs. This effort was intended to enhance communications and synergy among the Tri-Lab partners. Advanced functional and structural materials are central to transformative energy technologies for energy generation, conversion, delivery, and storage. With that in mind, the workshop focused on identifying and assessing the foundational materials research needs at both the basic and applied levels. Materials challenges include the ability to withstand harsh environments, such as high temperatures and pressures, corrosion, oxidation, or irradiation while maintaining flexible mission profiles and long service lifespans. Advanced energy system material challenges and needs range from materials for the capture, upgrading/concentration, storage, and delivery of low-grade heat to materials for high temperature environments that involve liquid metals, molten salt, and very high temperature gas heat delivery and storage systems. Material improvements are needed for hybrid energy systems due to accelerated corrosion and stress-fatigue failure of materials and equipment, which results from increased frequency and amplitude of thermal, mechanical, and electrical cycling of systems components. Multifunctional materials are needed for high temperature solid-oxide fuel cells, advanced electrochemical reactors, and in-process separation. Relative to materials manufacturing, application of advanced additive and subtractive methods need to be understood to develop both thin-layer homogenous materials and materials of graded composition. Materials modeling and machine learning will be critical to accelerate the design and production of power electronics, and nuclear reactor materials and fuel, as well as to gain an understanding of beneficial materials phenomena or deleterious microstructure evolution. There is also a need for standardized models, computational structures, data reporting protocols and modeling tools across the three laboratories. This would allow consistent results, analysis, and data sharing. Combining computational capabilities between the three laboratories (e.g., hardware, software) would greatly increase computational capabilities and throughput. The workshop identified the need for laboratories to anticipate and address problems that will occur during scale-up. Laboratory work must connect with industry to ensure that research focuses on processes that are scalable and marketable. Industry input and perspective are essential to guide laboratory research to meet these requirements and deploy new technology in industrial demonstrations. Another aspect of scale-up is the integration of multiple systems since new challenges often arise at the subsystem interfaces. Establishing a scale-up manufacturing demonstration/pilot plant, potentially as an industrial user facility, would be beneficial to the laboratories and industry. That modular scale-up manufacturing demonstration/pilot plant would allow researchers to find and resolve interface problems that cannot be identified by focusing only on individual parts. Communication exchanges among the organizers, attendees, and workshop survey responses indicate that the workshop was successful in achieving its goal to identify key technology gaps and research needs. Strong positive feedback was received on the sharing of ideas, capabilities, talent, and passion to move forward on the materials-related action items.

36 MATERIALS SCIENCE↗

NASA NTRS1998

On Convergence Acceleration Techniques for Unstructured Meshes

A discussion of convergence acceleration techniques as they relate to computational fluid dynamics problems on unstructured meshes is given. Rather than providing a detailed description of particular methods, the various different building blocks of current solution techniques are discussed and examples of solution strategies using one or several of these ideas are given. Issues relating to unstructured grid CFD problems are given additional consideration, including suitability of algorithms to current hardware trends, memory and cpu tradeoffs, treatment of non-linearities, and the development of efficient strategies for handling anisotropy-induced stiffness. The outlook for future potential improvements is also discussed.

Mavriplis, Dimitri J.↗

DOE OSTI2025

Improving Thermal Management Strategies for Data Centers: A Physical Testbed Incorporating Small Modular Reactor and Microreactor Technology

This study aims to accelerate the demonstration of various thermal management systems for data centers using nuclear-generated heat to enhance energy and grid reliability. Utilizing mobile containerized and stationary test beds at INL's High Performance Computing (HPC) facility, this project integrates with various nuclear-related energy systems testing facilities. Key components include immersion cooling apparatus, absorption chillers, and adjustable thermal management simulators. Tasks involve acquiring necessary hardware, sensors, and cooling apparatus, engaging with data center industry stakeholders, and providing a testing platform for algorithms, models, tools, and software. The objective is to expedite the deployment of nuclear-powered data centers, thereby improving energy reliability and affordability.

21 - SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLAN↗

DOE OSTI2024

Dual Channel Dual Staging: Hierarchical and Portable Staging for GPU-Based In-Situ Workflow

In-situ workflows have emerged as an attractive approach for addressing data movement challenges at very large scales. Since GPU-based architectures dominate the HPC landscapes, porting these in-situ workflows, and, specifically, the inter-application data exchange, to GPU-based systems can be challenging. Technologies such as GPUDirect RDMA (GDR), which is typically used for I/O in GPU applications as an optimization that circumvents the CPU overhead, can be leveraged to support bulk data exchanges between GPU applications. However, current GDR design often lacks performance portability across HPC clusters built with different hardware configurations. Furthermore, the local CPU may also be effectively used as an auxiliary communication mechanism to offload data exchanges. In this paper, we present a dual channel dual staging approach for efficient, scalable, and performance-portable inter-application data exchange for in-situ workflows. This approach exploits the data access pattern within in-situ workflows along with the inherent execution asynchrony to accelerate data exchanges and, at the same time, improve performance portability. Specifically, the dual channel dual staging method leverages both the local CPU and the remote data staging server to build a hierarchical joint staging area and uses this staging area to transform blocking inter-application bulk data exchanges into best-effort local data movements between GPU and CPU. The dual channel dual staging is implemented as a portability extension of the Dataspaces-GPU staging framework. We present an experimental evaluation of its performance, portability, and scalability using this implementation on three leadership GPU clusters. The evaluation results demonstrate that the dual channel dual staging method saves up to 75% in data-exchange time compared to host-based, GDR, and alternate portable designs, while maintaining scalability (up to 512 GPUs) and performance portability across the three platforms.

Zhang, Bo [University of Utah]↗