Search NASA⌕ Search

SEARCH · Search NASA

Results for “scalable computing”

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.

At least 1,585 records · Page 88

Horizontally scaling dChache SRM with the Terracotta platform

The dCache disk caching file system has been chosen by a majority of LHC experiments Tier 1 centers for their data storage needs. It is also deployed at many Tier 2 centers. The Storage Resource Manager (SRM) is a standardized grid storage interface and a single point of remote entry into dCache, and hence is a critical component. SRM must scale to increasing transaction rates and remain resilient against changing usage patterns. The initial implementation of the SRM service in dCache suffered from an inability to support clustered deployment, and its performance was limited by the hardware of a single node. Using the Terracotta platform, we added the ability to horizontally scale the dCache SRM service to run on multiple nodes in a cluster configuration, coupled with network load balancing. This gives site administrators the ability to increase the performance and reliability of SRM service to face the ever-increasing requirements of LHC data handling. In this paper we will describe the previous limitations of the architecture SRM server and how the Terracotta platform allowed us to readily convert single node service into a highly scalable clustered application.

99 GENERAL AND MISCELLANEOUS↗

A Log-Scaling Fault Tolerant Agreement Algorithm for a Fault Tolerant MPI

The lack of fault tolerance is becoming a limiting factor for application scalability in HPC systems. The MPI does not provide standardized fault tolerance interfaces and semantics. The MPI Forum's Fault Tolerance Working Group is proposing a collective fault tolerant agreement algorithm for the next MPI standard. Such algorithms play a central role in many fault tolerant applications. This paper combines a log-scaling two-phase commit agreement algorithm with a reduction operation to provide the necessary functionality for the new collective without any additional messages. Error handling mechanisms are described that preserve the fault tolerance properties while maintaining overall scalability.

99 GENERAL AND MISCELLANEOUS↗

RingX: Scalable Parallel Attention for Long-Context Learning on HPC

The attention mechanism has become foundational for remarkable AI breakthroughs since the introduction of the Transformer, driving the demand for increasingly longer context to power frontier models such as large-scale reasoning language models and high-resolution image/video generators. However, its quadratic computational and memory complexities present substantial challenges. Current state-of-the-art parallel attention methods, such as ring attention, are widely adopted for long-context training but utilize a point-to-point communication strategy that fails to fully exploit the capabilities of modern HPC network architectures. In this work, we propose ringX, a scalable family of parallel attention methods optimized explicitly for HPC systems. By enhancing workload partitioning, refining communication patterns, and improving load balancing, ringX achieves up to 3.4 × speedup compared to conventional ring attention on the Frontier supercomputer. Optimized for both bi-directional and causal attention mechanisms, ringX demonstrates its effectiveness through training benchmarks of a Vision Transformer (ViT) on a climate dataset and a Generative Pre-Trained Transformer (GPT) model, Llama3 8B. Our method attains an end-to-end training speedup of approximately 1.5 × in both scenarios. To our knowledge, the achieved 38% model FLOPs utilization (MFU) for training Llama3 8B with a 1M-token sequence length on 4,096 GPUs represents one of the highest training efficiencies reported for long-context learning on HPC systems. Our code implementation is available at https://github.com/jqyin/ringX-attention.

Yin, Junqi [ORNL] (ORCID:0000000338435520)↗

Reversible Parallel Discrete Event Formulation of a TLM-based Radio Signal Propagation Model

Radio signal strength estimation is essential in many applications, including the design of military radio communications and industrial wireless installations. For scenarios with large or richly- featured geographical volumes, parallel processing is required to meet the memory and computa- tion time demands. Here, we present a scalable and efficient parallel execution of the sequential model for radio signal propagation recently developed by Nutaro et al. Starting with that model, we (a) provide a vector-based reformulation that has significantly lower computational overhead for event handling, (b) develop a parallel decomposition approach that is amenable to reversibility with minimal computational overheads, (c) present a framework for transparently mapping the conservative time-stepped model into an optimistic parallel discrete event execution, (d) present a new reversible method, along with its analysis and implementation, for inverting the vector-based event model to be executed in an optimistic parallel style of execution, and (e) present performance results from implementation on Cray XT platforms. We demonstrate scalability, with the largest runs tested on up to 127,500 cores of a Cray XT5, enabling simulation of larger scenarios and with faster execution than reported before on the radio propagation model. This also represents the first successful demonstration of the ability to efficiently map a conservative time-stepped model to an optimistic discrete-event execution.

99 GENERAL AND MISCELLANEOUS↗

Parallel Multigrid in Time and Space for Extreme-Scale Computational Science: Chaotic and Hyperbolic Problems

The coming massive parallelism of exascale computing presents a pressing challenge for the many DOE simulations of time-dependent partial differential equations (PDEs), which typically use traditional sequential time stepping methods. Since this traditional approach is inherently serial, it presents a sequential bottleneck when moving to exascale computing, because future performance gains will come through greater concurrency, not faster clock speeds. Thus, the goal of this work is to research parallelism in time, i.e., methods that compute multiple time values simultaneously, not sequentially. The focus will be on hyperbolic and chaotic problems of interest to DOE, with the goal of enabling scalable simulations of time-dependent hyperbolic and chaotic problems on future architectures. The chosen methodology for solving these problems parallel-in-time is multigrid, because multigrid (when it works) is a powerful, optimal, and scalable solver for discretized PDEs. Multigrid is already commonly used in many DOE simulations for scalably and optimally solving space-only PDE problems. The areas of hyperbolic and chaotic problems are chosen because of their relevance to problems of programmatic interest to DOE. However, these problems are also well-known to be difficult for parallelin-time methods, with the most common method, parareal, diverging in many cases. The current stateof-the-art for parallel-in-time at LLNL is the multigrid reduction in time (MGRIT) XBraid package, which also struggles for such problems, while still showing some improvement over parareal. In summary, new methods are needed for an efficient parallel-in-time scheme for hyperbolic and chaotic problems, and this work shall research promising new multigrid methods in this area. In particular, we take inspiration from the Least Squares Shadowing (LSS by Wang) approach for solving chaotic problems. Here, an optimization approach is able to find “well-conditioned” shadow trajectories/solutions to the original “ill-conditioned” chaotic problem. Thus, the new multigrid methods researched here also arise in an optimization context.

97 MATHEMATICS AND COMPUTING↗

Scalable edge clustering of dynamic graphs via weighted line graphs

Timestamped relational datasets consisting of records (or connections) between pairs of entities are ubiquitous in network science. For applications like peer-to-peer communication, email, various social network interactions, and computer network security, it is useful to organize these records into groups based on how and when they are occurring. Weighted line graphs offer a natural way to model how records are related in such datasets but for large real-world graph topologies, building and utilizing the line graph is prohibitively expensive. Here, we present the framework to cluster the edges of a dynamic graph via the associated line graph that contains two major contributions. The first is a method to work with the line graph implicitly and the second is a distributed scale implementation of an agglomerative hierarchical graph clustering algorithm. We outline a novel hierarchical dynamic graph edge clustering approach that efficiently breaks massive relational datasets into small sets of edges containing events at various timescales. This is in stark contrast to traditional graph clustering algorithms that prioritize highly connected (clique-like) community structures. Our approach relies on constructing a sufficient subgraph of a weighted line graph and applying a hierarchical agglomerative clustering. This approach is related to scalable techniques from spatial clustering, nonlinear-dimension reduction, topological data analysis, and draws particular inspiration from HDBSCAN. As an edge clustering, this method yields an overlapping node clustering. Our algorithm is parallelizable and we demonstrate efficient clustering of a billion-scale, real-world dynamic graph into small edge sets that correlate in topology and time. The entire clustering process for a graph with tens of billions of edges takes just a few minutes of run time on 256 nodes of a distributed compute environment. We argue how the output of the edge clustering is useful for a multitude of data visualization and powerful machine learning tasks, both involving the original massive dynamic graph data and metadata associated with the nodes and edges. Finally, we describe how this approach can be extended to dynamic hypergraphs and dynamic graphs/hypergraphs with unstructured data living on vertices and edges.

Data Analysis↗

Enforcing End-to-end I/O Policies for Scientific Workflows using Software-Defined Storage Resource Enclaves

Data-intensive knowledge discovery requires scientific applications to run concurrently with analytics and visualization codes executing in situ for timely output inspection and knowledge extraction. Consequently, I/O pipelines of scientific workflows can be long and complex because they comprise many stages of analytics across different layers of the I/O stack of high-performance computing systems. Performance limitations at any I/O layer or stage can cause an I/O bottleneck resulting in greater than expected end-to-end I/O latency. In this paper, we present the design and implementation of a novel data management infrastructure called Software-Defined Storage Resource Enclaves (SIREN) at system level to enforce end-to-end policies that dictate an I/O pipeline's performance. SIREN provides an I/O performance interface for users to specify the desired storage resources in the context of in-situ analytics. If suboptimal performance of analytics is caused by an I/O bottleneck when data are transferred between simulations and analytics, schedulers in different layers of the I/O stack automatically provide the guaranteed lower bounds on I/O throughput. Lastly, our experimental results demonstrate that SIREN provides performance isolation among scientific workflows sharing multiple storage servers across two I/O layers (burst buffer and parallel file systems) while maintaining high system scalability and resource utilization.

97 MATHEMATICS AND COMPUTING↗

Toward quantum networking with frequency-bin qudits

Quantum networking holds tremendous promise in transforming computation and communication. Entangled-photon sources are critical for quantum repeaters and networking, while photonic integrated circuits are vital for miniaturization and scalability. In this talk, we focus on generating and manipulating frequency-bin entangled states within integrated platforms. We encode quantum information as a coherent superposition of multiple optical frequencies; this approach is favorable due to its amenability to high-dimensional entanglement and compatibility with fiber transmission. We successfully generate and measure the density matrix of biphoton frequency combs from integrated silicon nitride microrings, fully reconstructing the state in an 8 × 8 two-qudit Hilbert space, the highest so far for frequency bins. Moreover, we employ Vernier electro-optic phase modulation methods to perform time-resolved measurements of biphoton correlation functions. Currently, we are exploring bidirectional pumping of microrings to generate indistinguishable entangled pairs in both directions, aiming to demonstrate key networking operations such as entanglement swapping and Greenberger–Horne–Zeilinger state generation in the frequency domain.

Myilswamy, Karthik V.↗

Performance Analysis and Optimal Node-aware Communication for Enlarged Conjugate Gradient Methods

Krylov methods are a key way of solving large sparse linear systems of equations but suffer from poor strong scalability on distributed memory machines. Furthermore, this is due to high synchronization costs from large numbers of collective communication calls alongside a low computational workload. Enlarged Krylov methods address this issue by decreasing the total iterations to convergence, an artifact of splitting the initial residual and resulting in operations on block vectors. In this article, we present a performance study of an enlarged Krylov method, Enlarged Conjugate Gradients (ECG), noting the impact of block vectors on parallel performance at scale. Most notably, we observe the increased overhead of point-to-point communication as a result of denser messages in the sparse matrix-block vector multiplication kernel. Additionally, we present models to analyze expected performance of ECG, as well as motivate design decisions. Most importantly, we introduce a new point-to-point communication approach based on node-aware communication techniques that increases efficiency of the method at scale.

97 MATHEMATICS AND COMPUTING↗

Develop feedback system for intelligent dynamic resource allocation to improve application performance.

This report provides documentation for the completion of the Sandia Level II milestone 'Develop feedback system for intelligent dynamic resource allocation to improve application performance'. This milestone demonstrates the use of a scalable data collection analysis and feedback system that enables insight into how an application is utilizing the hardware resources of a high performance computing (HPC) platform in a lightweight fashion. Further we demonstrate utilizing the same mechanisms used for transporting data for remote analysis and visualization to provide low latency run-time feedback to applications. The ultimate goal of this body of work is performance optimization in the face of the ever increasing size and complexity of HPC systems.

99 GENERAL AND MISCELLANEOUS↗

A New Route to Quantum-Scale Structures through a Novel Enhanced Germanium Diffusion Mechanism

This project sought to develop a fundamental understanding of the mechanisms underlying a newly observed enhanced germanium (Ge) diffusion process in silicon germanium (SiGe) semiconductor nanostructures during thermal oxidation. Using a combination of oxidationdiffusion experiments, high resolution imaging, and theoretical modeling, a model for the enhanced Ge diffusion mechanism was proposed. Additionally, a nanofabrication approach utilizing this enhanced Ge diffusion mechanism was shown to be applicable to arbitrary 3D shapes, leading to the fabrication of stacked silicon quantum dots embedded in SiGe nanopillars. A new wet etch-based method for preparing 3D nanostructures for highresolution imaging free of obscuring material or damage was also developed. These results enable a new method for the controlled and scalable fabrication of on-chip silicon nanostructures with sub-10 nm dimensions needed for next generation microelectronics, including low energy electronics, quantum computing, sensors, and integrated photonics.

36 MATERIALS SCIENCE↗

Overview of NETL’s Low Temperature CO2 Electrolysis Research

This keynote lecture will briefly overview diverse research areas of National Energy Technology Laboratory (NETL) to advance energy and environmental sustainability along with carbon management. Our electrochemistry efforts on carbon conversion directly support the US goal of achieving carbon-free power sector by 2035 and net zero emissions by 2050. Since CO2 electroreduction is highly structure-sensitive, NETL ongoing research has been focused on the rational design and engineering of electrocatalysts to facilitate the CO2 conversion to desirable products with good selectivity, activity, and durability. Different classes and types of electrocatalytic materials will be covered in this talk, from well-defined atomic-scale model catalysts to heterogenous, scalable powder systems at nano- and micro-scale for “real world” performance evaluation. Several spectroscopic, microscopic, and electrochemical characterization techniques along with computational findings will be additionally discussed to gain more insights into the structure-activity relation. The last part of this seminar will provide more detail on how NETL has transitioned from the most common aqueous H-type reactor for lab-scale validation to more realistic full electrolyzer cell in bench-scale prototype. The knowledge, electrocatalytic materials, and device validation achieved from NETL in-house research will be translated to industrial sector for large scale deployment and the anticipated outcome will help advance the development of low temperature CO2 electrolysis technologies.

Nguyen Phan, Thuy Duong↗

Lightweight storage and overlay networks for fault tolerance.

The next generation of capability-class, massively parallel processing (MPP) systems is expected to have hundreds of thousands to millions of processors, In such environments, it is critical to have fault-tolerance mechanisms, including checkpoint/restart, that scale with the size of applications and the percentage of the system on which the applications execute. For application-driven, periodic checkpoint operations, the state-of-the-art does not provide a scalable solution. For example, on today's massive-scale systems that execute applications which consume most of the memory of the employed compute nodes, checkpoint operations generate I/O that consumes nearly 80% of the total I/O usage. Motivated by this observation, this project aims to improve I/O performance for application-directed checkpoints through the use of lightweight storage architectures and overlay networks. Lightweight storage provide direct access to underlying storage devices. Overlay networks provide caching and processing capabilities in the compute-node fabric. The combination has potential to signifcantly reduce I/O overhead for large-scale applications. This report describes our combined efforts to model and understand overheads for application-directed checkpoints, as well as implementation and performance analysis of a checkpoint service that uses available compute nodes as a network cache for checkpoint operations.

99 GENERAL AND MISCELLANEOUS↗

Adagio 4.14 user s guide.

This document is a user's guide for the code Adagio. Adagio is a three-dimensional, implicit solid mechanics code with a versatile element library, nonlinear material models, and capabilities for modeling large deformation and contact. Adagio is a parallel code, and its nonlinear solver and contact capabilities enable scalable solutions of large problems. It is built on the SIERRA Framework [1, 2]. SIERRA provides a data management framework in a parallel computing environment that allows the addition of capabilities in a modular fashion. The Adagio 4.14 User's Guide provides information about the functionality in Adagio and the command structure required to access this functionality in a user input file. This document is divided into chapters based primarily on functionality. For example, the command structure related to the use of various element types is grouped in one chapter; descriptions of material models are grouped in another chapter. The input and usage of Adagio is similar to that of the code Presto [3]. Presto, like Adagio, is a solid mechanics code built on the SIERRA Framework. The primary difference between the two codes is that Presto uses explicit time integration for transient dynamics analysis, whereas Adagio is an implicit code. Because of the similarities in input and usage between Adagio and Presto, the user's guides for the two codes are structured in the same manner and share common material. (Once you have mastered the input structure for one code, it will be easy to master the syntax structure for the other code.) To maintain the commonality between the two user's guides, we have used a variety of techniques. For example, references to Presto may be found in the Adagio user's guide and vice versa, and the chapter order across the two guides is the same. On the other hand, each of the two user's guides is expressly tailored to the features of the specific code and documents the particular functionality for that code. For example, though both Presto and Adagio have contact functionality, the content of the chapter on contact in the two guides differs. Important references for both Adagio and Presto are given in the references section at the end of this chapter. Adagio was preceded by the codes JAC and JAS3D; JAC is described in Reference 4; JAS3D is described in Reference 5. Presto was preceded by the code Pronto3D. Pronto3D is described in References 6 and 7. Some of the fundamental nonlinear technology used by both Presto and Adagio are described in References 8, 9, and 10. Currently, both Presto and Adagio use the Exodus II database and the XDMF database; Exodus II is more commonly used than XDMF. (Other options may be added in the future.) The Exodus II database format is described in Reference 11, and the XDMF database format is described in Reference 12. Important information about contact is provided in the reference document for ACME [13]. ACME is a third-party library for contact. One of the key concepts for the command structure in the input file is a concept referred to as scope. A detailed explanation of scope is provided in Section 1.2. Most of the command lines in Chapter 2 are related to a certain scope rather than to some particular functionality.

97 MATHEMATICS AND COMPUTING↗