Search NASA⌕ Search

SEARCH · Search NASA

Results for “Heterogeneous memory systems”

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.

33 records · Page 2
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

BULKI-Store v0.3.2

BULKI-Store is a distributed object storage system optimized for high-performance computing environments. Built with a Rust core and Python bindings, it efficiently manages scientific and machine learning datasets across HPC clusters. The system employs a client-server architecture with MPI integration, enabling seamless scaling on supercomputers like Perlmutter. BULKI-Store's object-oriented approach provides intuitive data organization with rich metadata support, contrasting with traditional file-based solutions. Key optimizations include selective checkpoint loading, unified checkpoint files, and object chunking for large data transfers. For machine learning workloads, BULKI-Store offers advantages through fine-grained access patterns, dynamic data sharing between training instances, and reduced memory pressure. Memory management features include strategic Python GC calls, minimized data copies, and batch processing capabilities. The system leverages Rayon's thread pool for asynchronous data prefetching and supports multiple CPU architectures (ARM64, x86, AMD, RISC-V). By combining performance optimizations with developer-friendly APIs, BULKI-Store addresses the complex data management challenges of modern HPC applications while maintaining compatibility across heterogeneous computing environments.

Zhang, Wei [Lawrence Berkeley National Laboratory ↗
NASA NTRS2000

Dynamic Load-Balancing for Distributed Heterogeneous Computing of Parallel CFD Problems

The developed methodology is aimed at improving the efficiency of executing block-structured algorithms on parallel, distributed, heterogeneous computers. The basic approach of these algorithms is to divide the flow domain into many sub- domains called blocks, and solve the governing equations over these blocks. Dynamic load balancing problem is defined as the efficient distribution of the blocks among the available processors over a period of several hours of computations. In environments with computers of different architecture, operating systems, CPU speed, memory size, load, and network speed, balancing the loads and managing the communication between processors becomes crucial. Load balancing software tools for mutually dependent parallel processes have been created to efficiently utilize an advanced computation environment and algorithms. These tools are dynamic in nature because of the chances in the computer environment during execution time. More recently, these tools were extended to a second operating system: NT. In this paper, the problems associated with this application will be discussed. Also, the developed algorithms were combined with the load sharing capability of LSF to efficiently utilize workstation clusters for parallel computing. Finally, results will be presented on running a NASA based code ADPAC to demonstrate the developed tools for dynamic load balancing.

Ecer, A.↗
NASA NTRS2006

Radiation Tolerant Intelligent Memory Stack (RTIMS)

The Radiation Tolerant Intelligent Memory Stack (RTIMS), suitable for both geostationary and low earth orbit missions, has been developed. The memory module is fully functional and undergoing environmental and radiation characterization. A self-contained flight-like module is expected to be completed in 2006. RTIMS provides reconfigurable circuitry and 2 gigabits of error corrected or 1 gigabit of triple redundant digital memory in a small package. RTIMS utilizes circuit stacking of heterogeneous components and radiation shielding technologies. A reprogrammable field programmable gate array (FPGA), six synchronous dynamic random access memories, linear regulator, and the radiation mitigation circuitries are stacked into a module of 42.7mm x 42.7mm x 13.00mm. Triple module redundancy, current limiting, configuration scrubbing, and single event function interrupt detection are employed to mitigate radiation effects. The mitigation techniques significantly simplify system design. RTIMS is well suited for deployment in real-time data processing, reconfigurable computing, and memory intensive applications.

Ng, Tak-kwong↗
NASA NTRS2024

Accurate Assessment of Land-Atmosphere Coupling in Climate Models Requires High Frequency Data Output

Land-atmosphere (L-A) interactions are important for understanding convective processes, climate feedbacks, the development and perpetuation of droughts, heatwaves, pluvials, and other land-centred climate anomalies. Local L-A coupling (LoCo) metrics capture relevant L-A processes, highlighting the impact of soil and vegetation states on surface flux partitioning, and the impact of surface fluxes on boundary layer (BL) growth, development, and entrainment of air above the BL. A primary goal of the Climate Process Team on Coupling Land and Atmospheric Subgrid Parameterizations (CLASP) is parameterizing and characterizing the impact of subgrid heterogeneity in global and regional earth system models (ESMs) to improve the connection between land and atmospheric states and processes. A critical step in achieving that aim is the incorporation of L-A metrics, especially LoCo metrics, into climate model diagnostic process streams. However, because land-atmosphere interactions span time scales of minutes (e.g., turbulent fluxes), hours (e.g., BL growth and decay), days (e.g., soil moisture memory), and seasons (e.g., variability of behavioural regimes between soil moisture and latent heat flux), with multiple processes of interest happening in different geographic regions at different times of year, there is not a single metric that captures all the modes, means, and methods of interaction between the land and the atmosphere. And while monthly means of most of the LoCo-relevant variables are routinely saved from ESM simulations, data storage constraints typically preclude routine archival of the hourly data that would enable the calculation of all LoCo metrics. Here we outline a reasonable data request that would allow for adequate characterization of sub-daily coupling processes between the land and the atmosphere, preserving enough sub-daily output to describe, analyse, and better understand L-A coupling in modern climate models. A secondary request involves embedding calculations within the models to determine mean properties in and above the BL to further improve characterization of model behaviour. Higher-frequency model output will (i) allow for more direct comparison with observational field campaigns on process-relevant time scales, (ii) enable demonstration of inter-model spread in L-A coupling processes, and (iii) aid in targeted identification of sources of deficiencies and opportunities for improvement of the models.

Kirsten L. Findell↗
DOE OSTI2025

Benchmarking Operators in Deep Neural Networks for Improving Performance Portability of SYCL

SYCL is a portable programming model for heterogeneous computing, so it is important to obtain reasonable performance portability of SYCL. Towards the goal of better understanding and improving performance portability of SYCL for machine learning workloads, we have been developing benchmarks for basic operators in deep neural networks (DNNs). These operators could be offloaded to heterogeneous computing devices such as graphics processing units (GPUs) to speed up computation. In this paper, we introduce the benchmarks, evaluate the performance of the operators on GPU-based systems, and describe the causes of the performance gap between the SYCL and Compute Unified Device Architecture (CUDA) kernels. We find that the causes are related to the utilization of the texture cache for read-only data, optimization of the memory accesses with strength reduction, use of local memory, and register usage per thread. We hope that the efforts of developing benchmarks for studying performance portability will stimulate discussion and interactions within the community.

Jin, Zheming [ORNL] (ORCID:000000027197780X)↗
NASA NTRSDate not supplied

A System to Provide Deterministic Flight Software Operation and Maximize Multicore Processing Performance: The Safe and Precise Landing – Integrated Capabilities Evolution (SPLICE) Datapath

A method and design are described for a system that processes multiple data streams, utilizing a multicore asymmetric processing architecture, that eliminates data interrupts to the application processors. The design supports a deterministic environment for flight software in NASA’s Safe and Precise Landing – Integrated Capabilities Evolution (SPLICE) project. The SPLICE project develops sensor, algorithm, and compute technologies for Precision Landing and Hazard Avoidance (PL&HA) capabilities. The compute technology for SPLICE is the Descent and Landing Computer (DLC). The DLC hosts several SPLICE algorithms with high computational resource requirements that must be executed in a real-time and deterministic manner. The software runs on a custom Single Board Computer (SBC), with a Xilinx Ultrascale+ Multiprocessor System-on-a-Chip (MPSoC). Input data for the flight software is from a variety of sensors, unique with respect to data rate and packet size. A data path between the SPLICE sensors and algorithms is designed to efficiently deliver this data to the flight software using the MPSoC asymmetric processing cores and Field Programmable Gate Array (FPGA) fabric. This is implemented in a manner that isolates the application processors running the flight software from interrupts associated with the input data. By leveraging real-time processors on the MPSoC, and a structure with the appropriate interfaces in the shared memory on the SBC, the flight software can use the full set of application processors. The available utilization for each processor in this set is also maximized for the SPLICE applications, providing a sufficiently deterministic execution environment without the cost and overhead of a real-time operating system.

heterogeneous processing system↗
NASA NTRS1976

HEP - A semaphore-synchronized multiprocessor with central control

The paper describes the design concept of the Heterogeneous Element Processor (HEP), a system tailored to the special needs of scientific simulation. In order to achieve high-speed computation required by simulation, HEP features a hierarchy of processes executing in parallel on a number of processors, with synchronization being largely accomplished by hardware. A full-empty-reserve scheme of synchronization is realized by zero-one-valued hardware semaphores. A typical system has, besides the control computer and the scheduler, an algebraic module, a memory module, a first-in first-out (FIFO) module, an integrator module, and an I/O module. The architecture of the scheduler and the algebraic module is examined in detail.

Gilliland, M. C.↗
DOE OSTI2025

CRCNS22 Learning Rules in the Hippocampus and their Mapping to Neuromorphic Systems (Final Technical Report)

Large scale biologically-realistic computational models are key to investigating the interplay between structure and function in nervous systems, thus paving the way to new clinical methods and neuro-inspired computing solutions. This project focuses on the hippocampus, in particular the CA3-CA1 regions, due to their role in associative learning and memory, pattern separation and completion, and spatial navigation. Investigations into the neuronal organization and learning rule(s) of this circuit can shed light into how declarative memories are formed, stored, recalled and forgotten and inform computational, experimental and clinical neuroscience work. Our project aims at developing a novel data-driven methodology supported by a broad heterogeneous base of neuroscience experimental knowledge and inspired from advances in computer science and engineering. Specifically, this work will benchmark existing and new learning rules within a full-scale spiking neural network simulation of the CA3-CA1 region. The model will be based on an open-source repository, called the Hippocampome, which contains neuronal morphologies, firing patterns, synapse probabilities, and most other required parameters for all known neuron types in the rodent hippocampal formation. The model will be first trained in a supervised fashion for associative memory tasks using backpropagation through time traditionally used in computer science, enhanced with a new technique called the surrogate gradient method. This optimization method will be used to obtain a global loss minimization, but it is not biologically inspired as it assumes the use of data not locally available to the synapses. However, we propose its use as a benchmarking tool, to compare the training performance of local biologically plausible and hardware-mappable learning rules at scale. New rules or combinations will be proposed and tested as needed, based on the obtained results. Progress in this area will also drive the development of novel hardware-mappable algorithms for continual lifelong learning and categorization of new events from few presented examples. This project goes beyond the existing state-of-the-art by looking at large scale realistic neuronal circuits as networks trainable via global optimization methods such as surrogate gradient descent. The objective function of the brain that supports learning is largely unknown, but it is likely that it operates through local learning rules. Studying network trajectories around local minima as proposed in this work represents a useful strategy for understanding whether a network is training by using a specific (set of) learning rule(s). Starting from a completely untrained network is a challenging test since it is difficult to determine how the learning rule affects the trajectory of the network. This interdisciplinary project will help understand what rule governs learning in these regions or if multiple learning rules are involved. The work will develop a robust methodology to measure if the network is converging to the target solution, oscillating around it, or diverging away.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Block encoding of the three-dimensional heterogeneous Poisson equation with application to fracture flow

Quantum linear system (QLS) algorithms offer the potential to solve large-scale linear systems exponentially faster than classical methods. However, applying QLS algorithms to real-world problems remains challenging due to issues such as state preparation, data loading, and efficient information extraction. In this work, we study the feasibility of applying QLS algorithms to solve discretized three-dimensional (3D) heterogeneous Poisson equations, with specific examples relating to groundwater flow through geologic fracture networks. We explicitly construct a block encoding for the 3D heterogeneous Poisson matrix by leveraging the sparse local structure of the discretized operator. While classical solvers benefit from preconditioning, we show that block encoding the system matrix and preconditioner separately does not improve the effective condition number that dominates the QLS run-time. This differs from classical approaches where the preconditioner and the system matrix can often be implemented independently. Nevertheless, due to the structure of the problem in three dimensions, the quantum algorithm achieves a run-time of 𝑂⁡(𝑁 2/3 polylog 𝑁 ⋅log (1/𝜖)), outperforming the best classical methods (with run times of 𝑂⁡(𝑁⁢log 𝑁 ⋅log (1/𝜖))) and offering exponential memory savings. These results highlight both the promise and limitations of QLS algorithms for practical scientific computing, and point to effective condition-number reduction as a key barrier in achieving quantum advantages.

58 GEOSCIENCES↗
DOE OSTI2026

Thin-film lithium niobate on diamond as a platform for efficient spin-phonon coupling

Negatively charged silicon vacancy (SiV − ) centers in diamond are leading candidates for solid-state quantum memory that can be controlled using electromagnetic or acoustic waves. The latter are particularly promising due to the strong strain response of SiV − , enabling large spin-phonon interaction strengths. Indeed, coherent spin control via surface acoustic waves (SAWs) has been demonstrated and is essential for developing on-chip phononic quantum networks. However, the absence of piezoelectricity in diamond requires an interface with a piezoelectric material for efficient transduction and delivery of acoustic waves to spins in diamonds. Here, we demonstrate a heterogeneously integrated phononic platform that combines thin-film lithium niobate with diamond to enable acoustic control of single SiV − spins. Additionally, leveraging large SAW-induced strain at the location of SiV − , we achieve coherent acoustic control of an electron spin with more than twofold improvement in Rabi frequency compared to state-of-the-art devices based on aluminum nitride on diamond. Furthermore, this work represents a crucial step towards realizing phonon-based quantum information processing systems.

Optically detected magnetic resonance↗
DOE OSTI2026

FitCache: A Transparent Drop-In Framework for Multi-Tier Caching to Accelerate Distributed Deep Learning Workloads

Training in Deep learning (DL) remains highly compute- and data-intensive, with I/O becoming a critical bottleneck as models and datasets scale. Recent studies report that data loading can dominate training time, especially on large-scale HPC systems with shared parallel file systems (PFS). Existing caching approaches either rely on single-tier designs or require intrusive modifications to training pipelines, limiting their portability and effectiveness. In this work, we present FitCache, a transparent drop-in framework for multi-tier caching to accelerate distributed DL training by coordinating fast local memory (e.g., DRAM, Persistent Memory (PMem)) and NVMe as hierarchical caches atop PFS. Our design adapts to hardware diversity, i.e., if NVMe is missing, memory transparently acts as a caching tier, ensuring stable performance. FitCache transparently intercepts I/O requests and issues concurrent fetches across all tiers, returning data from the fastest responder without centralized metadata or static redirection paths. FitCache adapts to dynamic workloads and heterogeneous clusters while maintaining POSIX compatibility. Experiments on Frontier (2048 GPUs) and smaller research clusters show that FitCache reduces training time by up to 40% and per-batch I/O latency by up to 71.6% compared to Lustre Orion PFS, offering a drop-in solution for scalable DL training.

Hu, Guangxing [ORNL] (ORCID:0009000283203614)↗
NASA NTRS2017

On the Development and Application of High Data Rate Architecture (HiDRA) in Future Space Networks

Historically, space missions have been severely constrained by their ability to downlink the data they have collected. These constraints are a result of relatively low link rates on the spacecraft as well as limitations on the time during which data can be sent. As part of a coherent strategy to address existing limitations and get more data to the ground more quickly, the Space Communications and Navigation (SCaN) program has been developing an architecture for a future solar system Internet. The High Data Rate Architecture (HiDRA) project is designed to fit into such a future SCaN network. HiDRA's goal is to describe a general packet-based networking capability which can be used to provide assets with efficient networking capabilities while simultaneously reducing the capital costs and operational costs of developing and flying future space systems.Along these lines, this paper begins by reviewing various characteristics of modern satellite design as well as relevant characteristics of emerging technologies (such as free-space optical links capable of working at 100+ Gbps). Next, the paper describes HiDRA's design, and how the system is able to both integrate and support the operation of not only today's high-rate systems, but also the high-rate systems likely to be found in the future. This section also explores both existing and future networking technologies, such as Delay Tolerant Networking (DTN) protocol (RFC4838 citeRFC:1, RFC5050citeRFC:2), and explains how HiDRA supports them. Additionally, this section explores how HiDRA is used for scheduling data movement through both proactive and reactive link management. After this, the paper moves on to explore a reference implementation of HiDRA. This implementation is currently being realized based on a Field Programmable Gate Array (FPGA) memory and interface controller that is itself controlled by a local computer running DTN software. Next, this paper explores HiDRA's natural evolution, which includes an integration path for software-defined networking (SDN) switches. This section also describes considerations for both near-Earth and deep-space instantiations of HiDRA, describing how differences in latencies between the environments will necessarily influence how the system is configured and the networks operate. Finally, this paper describes future work. This section includes a description of a potential ISS implementation which will allow rapid advancement through the technology readiness levels (TRL). This section also explores work being done to support HiDRA's successful implementation and operation in a heterogeneous network: such a network could include communications equipment spanning many vintages and capabilities, and one significant aspect of HiDRA's future development involves balancing compatibility with capability.

DTN↗
NASA NTRS2000

Testing New Programming Paradigms with NAS Parallel Benchmarks

Over the past decade, high performance computing has evolved rapidly, not only in hardware architectures but also with increasing complexity of real applications. Technologies have been developing to aim at scaling up to thousands of processors on both distributed and shared memory systems. Development of parallel programs on these computers is always a challenging task. Today, writing parallel programs with message passing (e.g. MPI) is the most popular way of achieving scalability and high performance. However, writing message passing programs is difficult and error prone. Recent years new effort has been made in defining new parallel programming paradigms. The best examples are: HPF (based on data parallelism) and OpenMP (based on shared memory parallelism). Both provide simple and clear extensions to sequential programs, thus greatly simplify the tedious tasks encountered in writing message passing programs. HPF is independent of memory hierarchy, however, due to the immaturity of compiler technology its performance is still questionable. Although use of parallel compiler directives is not new, OpenMP offers a portable solution in the shared-memory domain. Another important development involves the tremendous progress in the internet and its associated technology. Although still in its infancy, Java promisses portability in a heterogeneous environment and offers possibility to "compile once and run anywhere." In light of testing these new technologies, we implemented new parallel versions of the NAS Parallel Benchmarks (NPBs) with HPF and OpenMP directives, and extended the work with Java and Java-threads. The purpose of this study is to examine the effectiveness of alternative programming paradigms. NPBs consist of five kernels and three simulated applications that mimic the computation and data movement of large scale computational fluid dynamics (CFD) applications. We started with the serial version included in NPB2.3. Optimization of memory and cache usage was applied to several benchmarks, noticeably BT and SP, resulting in better sequential performance. In order to overcome the lack of an HPF performance model and guide the development of the HPF codes, we employed an empirical performance model for several primitives found in the benchmarks. We encountered a few limitations of HPF, such as lack of supporting the "REDISTRIBUTION" directive and no easy way to handle irregular computation. The parallelization with OpenMP directives was done at the outer-most loop level to achieve the largest granularity. The performance of six HPF and OpenMP benchmarks is compared with their MPI counterparts for the Class-A problem size in the figure in next page. These results were obtained on an SGI Origin2000 (195MHz) with MIPSpro-f77 compiler 7.2.1 for OpenMP and MPI codes and PGI pghpf-2.4.3 compiler with MPI interface for HPF programs.

Jin, H.↗
DOE OSTI2025

The Artificial Scientist: in-Transit Machine Learning of Plasma Simulations

Large-scale simulations or scientific experiments produce petabytes of data per run. This poses massive challenges for I/O and storage when scientific analysis workflows are run manually offline. Unsupervised deep learning-based techniques to extract patterns and non-linear relations from these large amounts of data provide a way to build scientific understanding from raw data, reducing the need for manual pre-selection of analysis steps, but require exascale compute and memory to process the full dataset available. In this paper, we demonstrate a heterogeneous streaming workflow in which plasma simulation data is streamed directly to a Machine Learning (ML) application training a model on the simulation data in-transit, completely circumventing the capacity-constrained filesystem bottleneck. This workflow employs openPMD to provide a high level interface to describe scientific data and also uses ADIOS2, to transfer volumes of data that exceed the capabilities of the filesystem. We employ experience replay to avoid catastrophic forgetting in learning from this non-steady state process in a continual manner and adapt it to improve model convergence while learning in-transit. As a proof-of-concept, we approach the ill-posed inverse problem of predicting particle dynamics from radiation in a particle-incell (PIConGPU) simulation of the Kelvin-Helmholtz instability (KHI). We detail hardware-software co-design challenges as we scale PIConGPU to full Frontier, the Top-1 system as of June 2024 Top500 list.

Kelling, Jeffrey [Helmholtz-Zentrum Dresden Rossen↗
NASA NTRS2012

Autonomous Information Unit for Fine-Grain Data Access Control and Information Protection in a Net-Centric System

As communication and networking technologies advance, networks will become highly complex and heterogeneous, interconnecting different network domains. There is a need to provide user authentication and data protection in order to further facilitate critical mission operations, especially in the tactical and mission-critical net-centric networking environment. The Autonomous Information Unit (AIU) technology was designed to provide the fine-grain data access and user control in a net-centric system-testing environment to meet these objectives. The AIU is a fundamental capability designed to enable fine-grain data access and user control in the cross-domain networking environments, where an AIU is composed of the mission data, metadata, and policy. An AIU provides a mechanism to establish trust among deployed AIUs based on recombining shared secrets, authentication and verify users with a username, X.509 certificate, enclave information, and classification level. AIU achieves data protection through (1) splitting data into multiple information pieces using the Shamir's secret sharing algorithm, (2) encrypting each individual information piece using military-grade AES-256 encryption, and (3) randomizing the position of the encrypted data based on the unbiased and memory efficient in-place Fisher-Yates shuffle method. Therefore, it becomes virtually impossible for attackers to compromise data since attackers need to obtain all distributed information as well as the encryption key and the random seeds to properly arrange the data. In addition, since policy can be associated with data in the AIU, different user access and data control strategies can be included. The AIU technology can greatly enhance information assurance and security management in the bandwidth-limited and ad hoc net-centric environments. In addition, AIU technology can be applicable to general complex network domains and applications where distributed user authentication and data protection are necessary. AIU achieves fine-grain data access and user control, reducing the security risk significantly, simplifying the complexity of various security operations, and providing the high information assurance across different network domains.

Chow, Edward T.↗