SEARCH · Search NASA

Results for “SPACE EXPLORATION”

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 595 records · Page 33

All sources NASA NTRS DOE OSTI

DOE OSTI2018

Coupling Visualization, Simulation, and Deep Learning for Ensemble Steering of Complex Energy Models

We describe a new framework that allows users to explore and steer ensembles of energy systems simulations by coupling multiple energy models and interactive visualization through a dataflow API. Through the visual interface, users can interactively explore complex parameter spaces populated by hundreds, or thousands, of simulation runs and interactively spawn new simulations to 'fill in' regions of interest in the parameter space. The computational and visualization capabilities reside within a general-purpose dataflow architecture for connecting producers of multidimensional timeseries data, such as energy simulations, with consumers of that data, whether they be visualizations, statistical analyses, or datastores. Fast computation and agile dataflow can enhance the engagement with energy simulations, allowing users to populate the parameter space in real time. However, many energy simulations are far too slow to provide an interactive response. To support interactive feedback, we are creating reduced-form simulations developed through machine learning techniques, which provide statistically sound estimates of the results of the full simulations at a fraction of the computational cost. These reduced-form simulations have response times on the order of seconds, suitable for real-time human-in-the-loop design and analysis. The approximation methods apply to a wide range of computational models, including supply-chain models, electric power grid simulations, and building models. Such reduced-form representations do not replace or re-implement existing simulations, but instead supplement them by enabling rapid scenario design and exploration for large ensembles of simulations. The improved understanding, facilitated by the reduced-form models, dataflow API, and visualization tools, allows researchers to better allocate computational resources to capture informative relationships within the system as well as provide a low-cost method for validating and quality-checking large-scale modeling efforts.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

When in-memory computing meets spiking neural networks—A perspective on device-circuit-system-and-algorithm co-design

This review explores the intersection of bio-plausible artificial intelligence in the form of spiking neural networks (SNNs) with the analog in-memory computing (IMC) domain, highlighting their collective potential for low-power edge computing environments. Through detailed investigation at the device, circuit, and system levels, we highlight the pivotal synergies between SNNs and IMC architectures. Additionally, we emphasize the critical need for comprehensive system-level analyses, considering the inter-dependencies among algorithms, devices, circuit, and system parameters, crucial for optimal performance. An in-depth analysis leads to the identification of key system-level bottlenecks arising from device limitations, which can be addressed using SNN-specific algorithm–hardware co-design techniques. This review underscores the imperative for holistic device to system design-space co-exploration, highlighting the critical aspects of hardware and algorithm research endeavors for low-power neuromorphic solutions.

Physics↗

DOE OSTI2022

Moving the needle: Employing deep reinforcement learning to push the boundaries of coarse-grained vaccine models

Highly mutable infectious disease pathogens (hm-IDPs) such as HIV and influenza evolve faster than the human immune system can contain them, allowing them to circumvent traditional vaccination approaches and causing over one million deaths annually. Agent-based models can be used to simulate the complex interactions that occur between immune cells and hm-IDP-like proteins (antigens) during affinity maturation—the process by which antibodies evolve. Compared to existing experimental approaches, agent-based models offer a safe, low-cost, and rapid route to study the immune response to vaccines spanning a wide range of design variables. However, the highly stochastic nature of affinity maturation and vast sequence space of hm-IDPs render brute force searches intractable for exploring all pertinent vaccine design variables and the subset of immunization protocols encompassed therein. To address this challenge, we employed deep reinforcement learning to drive a recently developed agent-based model of affinity maturation to focus sampling on immunization protocols with greater potential to improve the chosen metrics of protection, namely the broadly neutralizing antibody (bnAb) titers or fraction of bnAbs produced. Using this approach, we were able to coarse-grain a wide range of vaccine design variables and explore the relevant design space. Our work offers new testable insights into how vaccines should be formulated to maximize protective immune responses to hm-IDPs and how they can be minimally tailored to account for major sources of heterogeneity in human immune responses and various socioeconomic factors. Our results indicate that the first 3 to 5 immunizations, depending on the metric of protection, should be specially tailored to achieve a robust protective immune response, but that beyond this point further immunizations require only subtle changes in formulation to sustain a durable bnAb response.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2025

Basin & Range Investigation for Developing Geothermal Energy

Hidden geothermal systems represent a potentially prolific energy resource that could support critical U.S. public and government energy priorities. Basin and Range Investigations for Developing Geothermal Energy (BRIDGE) addressed some the challenges associated with hidden system exploration by prioritizing cost-effective exploration early on through strategic workflow and informed decision-making that mitigates early risk and shifts resources to later exploration stages (e.g., drilling). Sandia National Laboratories partnered with U.S. Navy Geothermal Office, Geologic Geothermal Group, and independent consultants, with additional collaboration with U.S. Geological Survey and private industry. The primary tool of the BRIDGE project was to deploy a regional-scale airborne electromagnetic method to investigate the shallow resistivity structure in areas with high prospectivity. This was followed up at several prospects by a multidisciplinary exploration approach, including additional geologic, geophysical and geochemical studies. A central tenet to the BRIDGE methodology is that zones of low resistivity frequently occur over geothermal systems in the Basin and Range, and when paired with other data constraints, imaging these zones can enable discovery of these systems. In addition to exploring greenfield areas (i.e., Grover Point), the BRIDGE project also flew HTEM resistivity surveys over known geothermal systems including those with established power plants (Don A. Campbell and Salt Wells) and prospects that are known to the literature but remain undeveloped, at least in part, due to a lack of understanding on the location of their producible reservoirs. BRIDGE produced a comprehensive set of data from prospects identified in the Nevada Play Fairway Analysis along with conceptual models for top ranking prospects, wherein all of the observations are used to inform an interpreted model of the system. These models present a range of possible system parameters such as temperature and size, and they are further informed by system analogues in the Basin and Range province and elsewhere. The results of this work leave space for further exploration that may now occur at prospects ‘down the list’ rather than distribution exploration resources evenly across all prospects.

15 GEOTHERMAL ENERGY↗

DOE OSTI2024

Accelerating the design of lattice structures using machine learning

Lattices remain an attractive class of structures due to their design versatility; however, rapidly designing lattice structures with tailored or optimal mechanical properties remains a significant challenge. With each added design variable, the design space quickly becomes intractable. To address this challenge, research efforts have sought to combine computational approaches with machine learning (ML)-based approaches to reduce the computational cost of the design process and accelerate mechanical design. While these efforts have made substantial progress, significant challenges remain in (1) building and interpreting the ML-based surrogate models and (2) iteratively and efficiently curating training datasets for optimization tasks. Here, we address the first challenge by combining ML-based surrogate modeling and Shapley additive explanation (SHAP) analysis to interpret the impact of each design variable. We find that our ML-based surrogate models achieve excellent prediction capabilities (R 2 > 0.95) and SHAP values aid in uncovering design variables influencing performance. We address the second challenge by utilizing active learning-based methods, such as Bayesian optimization, to explore the design space and report a 5 × reduction in simulations relative to grid-based search. Collectively, these results underscore the value of building intelligent design systems that leverage ML-based methods for uncovering key design variables and accelerating design.

36 MATERIALS SCIENCE↗

DOE OSTI2024

Neural architecture search via similarity adaptive guidance

Evolutionary neural network architecture search (ENAS) has attracted the attention of many experts due to its global optimization capabilities to automatically search for convolutional neural network architectures based on the target task. The current search space for ENAS is not to design a fully structured network, but to search for smaller cell architectures to reduce search costs. However, blind search strategies do not effectively utilize the potential experience of the population. In order to utilize the potential experience learned by the current population to guide the evolutionary search of the population, we propose a similarity guided neural network architecture search algorithm based on cell architecture, which utilizes the similarity between pairwise architectures in the population as empirical knowledge learned by the population. Our proposed algorithm provides a novel method for calculating architecture similarity, which calculates architecture similarity separately from the cell and macro-structure. Then we decouple the connections and operations in the cell and calculate connection and operation similarity separately. In addition, we propose adaptive similarity selection and binary tournament selection strategies to enhance the algorithm’s global and local search capabilities and effectively explore the search space. Finally, we design an improved single-point crossover operator to enhance the local search ability of the evolutionary operator. The experimental results show that SAGNAS is a competitive algorithm that achieves 97.44% and 81.60% in CIFAR10 and CIFAR100 with only 1.9 GPU-days spent.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2020

Data-Driven Operator Theoretic Methods for Global Phase Space Learning

In this work, we propose to apply the recently developed Koopman operator techniques to explore the global phase space of a nonlinear system from time-series data. In particular, we address the problem of identifying various invariant subsets of a phase space from the spectral properties of the associated Koopman operator constructed from time-series data. Moreover, in the case when the system evolution is known locally in various invariant subspaces, then a phase space stitching result is proposed that can be applied to identify a global Koopman operator. A biological system, bistable toggle switch and a second-order nonlinear system example are considered to illustrate the proposed results. The construction of this global Koopman operator is very helpful in experimental works as multiple experiments can't be run at the same time starting from several initial conditions.

Nandanoori, Sai Pushpak↗

DOE OSTI2014

Risk Estimation Methodology for Launch Accidents.

As compact and light weight power sources with reliable, long lives, Radioisotope Power Systems (RPSs) have made space missions to explore the solar system possible. Due to the hazardous material that can be released during a launch accident, the potential health risk of an accident must be quantified, so that appropriate launch approval decisions can be made. One part of the risk estimation involves modeling the response of the RPS to potential accident environments. Due to the complexity of modeling the full RPS response deterministically on dynamic variables, the evaluation is performed in a stochastic manner with a Monte Carlo simulation. The potential consequences can be determined by modeling the transport of the hazardous material in the environment and in human biological pathways. The consequence analysis results are summed and weighted by appropriate likelihood values to give a collection of probabilistic results for the estimation of the potential health risk. This information is used to guide RPS designs, spacecraft designs, mission architecture, or launch procedures to potentially reduce the risk, as well as to inform decision makers of the potential health risks resulting from the use of RPSs for space missions.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2013

Expert System for Building TRU Waste Payloads - 13554

The process for grouping TRU waste drums into payloads for shipment to the Waste Isolation Pilot Plant (WIPP) for disposal is a very complex process. Transportation and regulatory requirements must be met, along with striving for the goals of shipment efficiency: maximize the number of waste drums in a shipment and minimize the use of empty drums which take up precious underground storage space. The restrictions on payloads range from weight restrictions, to limitations on flammable gas in the headspace, to minimum TRU alpha activity concentration requirements. The Overpack and Payload Assistant Tool (OPAT) has been developed as a mixed-initiative intelligent system within the WIPP Waste Data System (WDS) to guide the construction of multiple acceptable payloads. OPAT saves the user time while at the same time maximizes the efficiency of shipments for the given drum population. The tool provides the user with the flexibility to tune critical factors that guide OPAT's operation based on real-time feedback concerning the results of the execution. This feedback complements the user's external knowledge of the drum population (such as location of drums, known challenges, internal shipment goals). This work demonstrates how software can be utilized to complement the unique domain knowledge of the users. The mixed-initiative approach combines the insight and intuition of the human expert with the proficiency of automated computational algorithms. The result is the ability to thoroughly and efficiently explore the search space of possible solutions and derive the best waste management decision. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2021

Data-Driven Strategies for Accelerated Materials Design

The ongoing revolution of the natural sciences by the advent of machine learning and artificial intelligence sparked significant interest in the material science community in recent years. The intrinsically high dimensionality of the space of realizable materials makes traditional approaches ineffective for large-scale explorations. Modern data science and machine learning tools developed for increasingly complicated problems are an attractive alternative. An imminent climate catastrophe calls for a clean energy transformation by overhauling current technologies within only several years of possible action available. Tackling this crisis requires the development of new materials at an unprecedented pace and scale. For example, organic photovoltaics have the potential to replace existing silicon-based materials to a large extent and open up new fields of application. In recent years, organic light-emitting diodes have emerged as state-of-the-art technology for digital screens and portable devices and are enabling new applications with flexible displays. Reticular frameworks allow the atom-precise synthesis of nanomaterials and promise to revolutionize the field by the potential to realize multifunctional nanoparticles with applications from gas storage, gas separation, and electrochemical energy storage to nanomedicine. In the recent decade, significant advances in all these fields have been facilitated by the comprehensive application of simulation and machine learning for property prediction, property optimization, and chemical space exploration enabled by considerable advances in computing power and algorithmic efficiency. In this Account, we review the most recent contributions of our group in this thriving field of machine learning for material science. We start with a summary of the most important material classes our group has been involved in, focusing on small molecules as organic electronic materials and crystalline materials. Specifically, we highlight the data-driven approaches we employed to speed up discovery and derive material design strategies. Subsequently, our focus lies on the data-driven methodologies our group has developed and employed, elaborating on high-throughput virtual screening, inverse molecular design, Bayesian optimization, and supervised learning. We discuss the general ideas, their working principles, and their use cases with examples of successful implementations in data-driven material discovery and design efforts. Furthermore, we elaborate on potential pitfalls and remaining challenges of these methods. Finally, we provide a brief outlook for the field as we foresee increasing adaptation and implementation of large scale data-driven approaches in material discovery and design campaigns.

36 MATERIALS SCIENCE↗

DOE OSTI2018

High-Pressure Synthesis: A New Frontier in the Search for Next-Generation Intermetallic Compounds

The application of high pressure adds an additional dimension to chemical phase space, opening up an unexplored expanse bearing tremendous potential for discovery. Our continuing mission is to explore this new frontier, to seek out new intermetallic compounds and new solid-state bonding. Simple binary elemental systems, specifically those composed of pairs of elements that do not form compounds under ambient pressures, can yield novel crystalline phases under compression. Thus, high-pressure synthesis can provide access to solid-state compounds that cannot be formed with traditional thermodynamic methods. An emerging approach for the rapid exploration of composition–pressure–temperature phase space is the use of hand-held high-pressure devices known as diamond anvil cells (DACs). These devices were originally developed by geologists as a way to study minerals under conditions relevant to the earth’s interior, but they possess a host of capabilities that make them ideal for high-pressure solid-state synthesis. Of particular importance, they offer the capability for in situ spectroscopic and diffraction measurements, thereby enabling continuous reaction monitoring—a powerful capability for solid-state synthesis. Here, we offer an overview of this approach in the context of research we have performed in the pursuit of new intermetallic compounds. We start with a discussion of pressure as a fundamental experimental variable that enables the formation of intermetallic compounds that cannot be isolated under ambient conditions. We then introduce the DAC apparatus and explain how it can be repurposed for use as a synthetic vessel with which to explore this phase space, going to extremes of pressure where no chemist has gone before. The remainder of the Account is devoted to discussions of recent experiments we have performed with this approach that have led to the discovery of novel intermetallic compounds in the Fe–Bi, Cu–Bi, and Ni–Bi systems, with a focus on the cutting-edge methods that made these experiments possible. We review the use of in situ laser heating at high pressure, which led to the discovery of FeBi2, the first binary intermetallic compound in the Fe–Bi system. Our work in the Cu–Bi system is described in the context of in situ experiments carried out in the DAC to map its high-pressure phase space, which revealed two intermetallic phases (Cu11Bi7 and CuBi). Finally, we review the discovery of β-NiBi, a novel high-pressure phase in the Ni–Bi system. We hope that this Account will inspire the next generation of solid-state chemists to boldly explore high-pressure phase space.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2024

The effect of gamma ray irradiation on few layered MoSe2: A material for nuclear and space applications

In recent years, emerging two-dimensional (2D) materials, such as molybdenum diselenide (MoSe2), have been at the center of attention for many researchers. This is due to their unique and fascinating physicochemical properties that make them attractive in space and defense applications that include shielding harsh irradiation environments. In this study, we examined the effects of gamma (γ) rays at various doses on the structural, chemical, and optical properties of MoSe2 layers. After the samples were exposed to intense gamma radiation (from a 60Co source) with various exposure times to vary the total accumulated dosage (up to 100 kGy), Raman and photoluminescence spectroscopies were used to study and probe radiation-induced changes to the samples. When compared to pristine materials, very few changes in optical properties were typically observed, indicating good robustness with little sensitivity, even at relatively high doses of gamma radiation. The imaging using scanning electron microscopy revealed a number of nano-hillocks that were connected to substrate alterations. X-ray photoelectron spectroscopies revealed that Mo’s binding energies remained the same, but Se’s binding energies blueshifted. We associated this shift with the decrease in Se vacancies that occurred after irradiation as a result of Mo atoms creating adatoms next to Se atoms. When compared to pristine materials, very few changes in optical, chemical, and structural properties were typically observed. These findings highlight the inherent resilience of MoSe2 in hostile radioactive conditions, which spurs additional research into their optical, electrical, and structural characteristics as well as exploration for potential space, energy, and defense applications.

Materials Science↗

DOE OSTI2016

Fermi/LAT observations of dwarf galaxies highly constrain a dark matter interpretation of excess positrons seen in AMS-02, HEAT, and PAMELA

It is shown that a Weakly Interacting Massive dark matter Particle (WIMP) interpretation for the positron excess observed in a variety of experiments, HEAT, PAMELA, and AMS-02, is highly constrained by the Fermi/LAT observations of dwarf galaxies. In particular, this paper examines the annihilation channels that best fit the current AMS-02 data (Boudaud et al., 2014), specifically focusing on channels and parameter space not previously explored by the Fermi/LAT collaboration. The Fermi satellite has surveyed the γ-ray sky, and its observations of dwarf satellites are used to place strong bounds on the annihilation of WIMPs into a variety of channels. For the single channel case, we find that dark matter annihilation into (b b-bar ,e{sup +}e{sup -}, μ{sup +}μ{sup -}, τ{sup +}τ{sup -},4-e or 4-τ ) is ruled out as an explanation of the AMS positron excess (here b quarks are a proxy for all quarks, gauge and Higgs bosons). In addition, we find that the Fermi/LAT 2σ upper limits, assuming the best-fit AMS-02 branching ratios, exclude multichannel combinations into b b-bar and leptons. The tension between the results might relax if the branching ratios are allowed to deviate from their best-fit values, though a substantial change would be required. Of all the channels we considered, the only viable channel that survives the Fermi/LAT constraint and produces a good fit to the AMS-02 data is annihilation (via a mediator) to 4-μ, or mainly to 4-μ in the case of multichannel combinations.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2019

High entropy alloys: A focused review of mechanical properties and deformation mechanisms

The high-entropy alloy (HEA) concept was based on the idea that high mixing entropy can promote formation of stable single-phase microstructures. During the past 15 years, various alloy systems have been explored to identify HEA systems with improved property combinations, leading to an extraordinary growth of this field. In the large pool of alloys with varying characteristics, the first single-phase HEA with good tensile properties, the equiatomic CrMnFeCoNi alloy has become the benchmark material, and it forms the basis of much of our current fundamental understanding of HEA mechanical behavior. As the field is evolving to the more broadly defined complex concentrated alloys (CCAs) and the available data in the literature increase exponentially, a fundamental question remains unchanged: how special are these new materials? In the first part of this review, select mechanical properties of HEAs and CCAs are compared with those of conventional engineering alloys. This task is difficult because of the limited tensile data available for HEAs and CCAs. Additionally, the wider suite of mechanical properties needed to assess structural materials is woefully lacking. Nonetheless, our evaluations have not revealed many HEAs or CCAs with properties far exceeding those of conventional engineering alloys, although specific alloys can show notable enhancements in specific properties. Consequently, it is reasonable to first approach the understanding of HEAs and CCAs through the assessment of how the well-established deformation mechanisms in conventional alloys operate or are modified in the presence of the high local complexity of the HEAs and CCAs. The second part of the paper provides a detailed review of the deformation mechanisms of HEAs with the FCC and BCC structures. For the former, we chose the CrMnFeCoNi (Cantor) alloy because it is the alloy on which the most rigorous and thorough investigations have been performed and, for the latter, we chose the TiZrHfNbTa (Senkov) alloy because this is one of the few refractory HEAs that exhibits any tensile ductility at room temperature. As expected, our review shows that the fundamental deformation mechanisms in these systems, and their dependence on basic physical properties, are broadly similar to those of conventional FCC and BCC metals. The third part of this review examines the theoretical and modeling efforts to date that seek to provide either qualitative or quantitative understanding of the mechanical performance of FCC and BCC HEAs. Since experiments reveal no fundamentally new mechanisms of deformation, this section starts with an overview of modeling perspectives and fundamental considerations. Next, the review turns to the evolution of modeling and predictions as compared to recent experiments, highlighting both successes and limitations. Finally, in spite of some significant successes, important directions for further theory development are discussed. Overall, while the individual deformation mechanisms or properties of the HEAs and CCAs are not, by and large, “special” relative to conventional alloys, the present HEA rush remains valuable because the compositional freedom that comes from the multi-element space will allow exploration of whether multiple mechanisms can operate sequentially or simultaneously, which may yet lead to the creation of new alloys with a spectrum of mechanical properties that are significantly superior to those of current engineering alloys.

36 MATERIALS SCIENCE↗

DOE OSTI2024

In Silico Chemical Experiments in the Age of AI: From Quantum Chemistry to Machine Learning and Back

Computational chemistry is an indispensable tool for understanding molecules and predicting chemical properties. However, traditional computational methods face significant challenges due to the difficulty of solving the Schrödinger equations and the increasing computational cost with the size of the molecular system. In response, there has been a surge of interest in leveraging artificial intelligence (AI) and machine learning (ML) techniques to in silico experiments. Integrating AI and ML into computational chemistry increases the scalability and speed of the exploration of chemical space. However, challenges remain, particularly regarding the reproducibility and transferability of ML models. This review highlights the evolution of ML in learning from, complementing, or replacing traditional computational chemistry for energy and property predictions. Starting from models trained entirely on numerical data, a journey set forth toward the ideal model incorporating or learning the physical laws of quantum mechanics. This paper also reviews existing computational methods and ML models and their intertwining, outlines a roadmap for future research, and identifies areas for improvement and innovation. Ultimately, the goal is to develop AI architectures capable of predicting accurate and transferable solutions to the Schrödinger equation, thereby revolutionizing in silico experiments within chemistry and materials science.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2023

Effect of the Nature of Both Cation and Anion Substitution on the Structural Symmetry of Li‐Rich 3 d ‐Metal Chalcogenide Electrodes

Abstract Li‐rich layered chalcogenides have recently led to better understanding of the anionic redox process and its associated high capacity while providing ways to overcome its practical limitations of voltage fade and irreversibility. This study reports on the feasibility of triggering anionic activity in Li 2 TiS 3 , through anionic substitution (Se for S) or cationic substitution (Fe for Ti). Herein, the chalcogenide chemical space is further explored to prepare mono‐substituted Li 1.7 Ti 0.85 Mn 0.45 Ch 3 (Ch = S/Se) and doubly substituted cationic and anionic phases (Li 1.7 Ti 0.85 Fe 0.45 S 3‐z Se z ) which crystallize either in the O3‐ or O1‐type structures depending upon substituents. All series show a bell‐shape capacity variation as function of the transition metal (TM) substitution degree with values up to 240 mAh g −1 . For specific compositions, a structural O3 to O1 phase transition is observed upon Li removal, which is not reversible upon Li re‐insertion due to kinetic limitations and negatively affects long‐term cycling performance. Density functional theory (DFT) calculations confirm the O3/O1 relative stability along the different series and point subtle electronic differences in the TM‐doping, rationalizing the structural and electrochemical behaviors of these phases upon cycling. These findings provide further insights into the link between structural and electronic stability, which is of key importance for designing chalcogenide‐based anionic redox compounds.

Chemistry↗

DOE OSTI2021

Autotuning PolyBench benchmarks with LLVM Clang/Polly loop optimization pragmas using Bayesian optimization

Here, we develop a ytopt autotuning framework that leverages Bayesian optimization to explore the parameter space search and compare four different supervised learning methods within Bayesian optimization and evaluate their effectiveness. We select six of the most complex PolyBench benchmarks and apply the newly developed LLVM Clang/Polly loop optimization pragmas to the benchmarks to optimize them. We then use the autotuning framework to optimize the pragma parameters to improve their performance. The experimental results show that our autotuning approach outperforms the other compiling methods to provide the smallest execution time for the benchmarks syr2k, 3mm, heat-3d, lu, and covariance with two large datasets in 200 code evaluations for effectively searching the parameter spaces with up to 170,368 different configurations. We find that the Floyd-Warshall benchmark did not benefit from autotuning. To cope with this issue, we provide some compiler option solutions to improve the performance. Then we present loop autotuning without a user's knowledge using a simple mctree autotuning framework to further improve the performance of the Floyd-Warshall benchmark. We also extend the ytopt autotuning framework to tune a deep learning application.

79 ASTRONOMY AND ASTROPHYSICS↗

DOE OSTI2024

Evaluating the potential of disaggregated memory systems for HPC applications

Summary Disaggregated memory is a promising approach that addresses the limitations of traditional memory architectures by enabling memory to be decoupled from compute nodes and shared across a data center. Cloud platforms have deployed such systems to improve overall system memory utilization, but performance can vary across workloads. High‐performance computing (HPC) is crucial in scientific and engineering applications, where HPC machines also face the issue of underutilized memory. As a result, improving system memory utilization while understanding workload performance is essential for HPC operators. Therefore, learning the potential of a disaggregated memory system before deployment is a critical step. This paper proposes a methodology for exploring the design space of a disaggregated memory system. It incorporates key metrics that affect performance on disaggregated memory systems: memory capacity, local and remote memory access ratio, injection bandwidth, and bisection bandwidth, providing an intuitive approach to guide machine configurations based on technology trends and workload characteristics. We apply our methodology to analyze thirteen diverse workloads, including AI training, data analysis, genomics, protein, fusion, atomic nuclei, and traditional HPC bookends. Our methodology demonstrates the ability to comprehend the potential and pitfalls of a disaggregated memory system and provides motivation for machine configurations. Our results show that eleven of our thirteen applications can leverage injection bandwidth disaggregated memory without affecting performance, while one pays a rack bisection bandwidth penalty and two pay the system‐wide bisection bandwidth penalty. In addition, we also show that intra‐rack memory disaggregation would meet the application's memory requirement and provide enough remote memory bandwidth.

Ding, Nan↗