Search NASA⌕ Search

SEARCH · Search NASA

Results for “Heterogeneous 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.

At least 343 records · Page 19

ORCHA: A performance portability system for extreme heterogeneity

Heterogeneity is the prevalent trend in the rapidly evolving high-performance computing (HPC) landscape in both hardware and application software. The diversity in hardware platforms, currently comprising various accelerators and a future possibility of specializable chiplets, poses a significant challenge for scientific software developers aiming to harness optimal performance across different computing platforms while maintaining the quality of solutions when their applications are simultaneously growing more complex. Code synthesis and code generation can provide mechanisms to mitigate this challenge. We have developed a divide and conquer approach where different aspects of performance are handled by different stand-alone tools that are interfaced with the application through generated code. This portability system, ORCHA, enables users to configure and orchestrate their computations among available resources on a platform by specifying a high-level recipe, thereby permitting a many-to-many paradigm where each recipe results in a different variant of the application. The core design goal is to let users decide the application’s hardware mapping and orchestration by editing only the high-level recipe—without modifying the maintained source code or binding the application to a particular runtime system. Tools in ORCHA distribution are: CG-Kit for translating the recipe into an execution graph; Milhoja to execute the graph by orchestrating data and task movement among hardware resources; and Macroprocessor that enables users to define their own code-shorthand for higher composability and easier management of code variants. Additionally, the design of ORCHA permits tools to work in a plug-and-play mode where the application can build and run without CG-Kit and Milhoja, and either tool can be swapped out for other tools with similar capabilities by modifying the code generation portion of ORCHA. In this paper, we describe the design of ORCHA and the role that code-generation plays in isolating applications from tools. We demonstrate the breadth of configurations ORCHA enables with a case study in which an application configuration is realized on three distinct hardware mappings—a GPU-centric, a CPU/GPU balanced, and a CPU/GPU concurrent layouts by using different recipes.

Lee, Youngjun↗

Runtime extension for neural network training with heterogeneous memory

Systems, apparatuses, and methods for managing buffers in a neural network implementation with heterogeneous memory are disclosed. A system includes a neural network coupled to a first memory and a second memory. The first memory is a relatively low-capacity, high-bandwidth memory while the second memory is a relatively high-capacity, low-bandwidth memory. During a forward propagation pass of the neural network, a run-time manager monitors the usage of the buffers for the various layers of the neural network. During a backward propagation pass of the neural network, the run-time manager determines how to move the buffers between the first and second memories based on the monitored buffer usage during the forward propagation pass. As a result, the run-time manager is able to reduce memory access latency for the layers of the neural network during the backward propagation pass.

Mappouras, Georgios↗

Novel trends in pair distribution function approaches on bulk systems with nanoscale heterogeneities

Novel materials for high performance applications increasingly exhibit structural order on the nanometer length scale; a domain where crystallography, the basis of Rietveld refinement, fails [1]. In such instances the total scattering approach, which treats Bragg and diffuse scattering on an equal basis, is a powerful approach. In recent years, the analysis of the total scattering data became an invaluable tool and the gold standard for studying nanocrystalline, nanoporous, and disordered crystalline materials. The data may be analyzed in reciprocal space directly, or Fourier transformed to the real-space atomic pair distribution function (PDF) and this intuitive function examined for local structural information. Here we give a number of illustrative examples, for convenience picked from our own work, of recent developments and applications of total scattering and PDF analysis to novel complex materials. There are many other wonderful examples from the work of others.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Addressing Global Data Dependencies in Heterogeneous Asynchronous Runtime Systems on GPUs

Large-scale parallel applications with complex global data dependencies beyond those of reductions pose significant scalability challenges in an asynchronous runtime system. Internodal challenges include identifying the all-to-all communication of data dependencies among the nodes. Intranodal challenges include gathering together these data dependencies into usable data objects while avoiding data duplication. This paper addresses these challenges within the context of a large-scale, industrial coal boiler simulation using the Uintah asynchronous many-task runtime system on GPU architectures. We show significant reduction in time spent analyzing data dependencies through refinements in our dependency search algorithm. Multiple task graphs are used to eliminate subsequent analysis when task graphs change in predictable and repeatable ways. Using a combined data store and task scheduler redesign reduces data dependency duplication ensuring that problems fit within host and GPU memory. Furthermore, these modifications did not require any changes to application code or sweeping changes to the Uintah runtime system. We report results running on the DOE Titan system on 119K CPU cores and 7.5K GPUs simultaneously. Our solutions can be generalized to other task dependency problems with global dependencies among thousands of nodes which must be processed efficiently at large scale.

97 MATHEMATICS AND COMPUTING↗

Development of a heterogeneous laminating resin system

The factors which effect the impact resistance of laminating resin systems and yet retain equivalent performance with the conventional 450 K curing epoxy matrix systems in other areas were studied. Formulation work was conducted on two systems, an all-epoxy and an epoxy/bismaleimide, to gain fundamental information on the effect formulation changes have upon neat resin and composite properties. The all-epoxy work involved formulations with various amounts and combinations of eight different epoxy resins, four different hardeners, fifteen different toughening agents, a filler, and a catalyst. The epoxy/bismaleimide effort improved formulations with various amounts and combinations of nine different resins, four different hardeners, eight different toughening agents, four different catalysts, and a filler. When a formulation appeared to offer the proper combination of properties required for a laminating resin Celion 3K-70P fabric was prepregged. Initial screening tests on composites primarily involved Gardner type impact and measurement of short beam shear strengths under dry and hot/wet conditions.

Biermann, T. F.↗

Database interfaces on NASA's heterogeneous distributed database system

The purpose of the ORACLE interface is to enable the DAVID program to submit queries and transactions to databases running under the ORACLE DBMS. The interface package is made up of several modules. The progress of these modules is described below. The two approaches used in implementing the interface are also discussed. Detailed discussion of the design of the templates is shown and concluding remarks are presented.

Huang, S. H. S.↗

Database interfaces on NASA's heterogeneous distributed database system

The purpose of Distributed Access View Integrated Database (DAVID) interface module (Module 9: Resident Primitive Processing Package) is to provide data transfer between local DAVID systems and resident Data Base Management Systems (DBMSs). The result of current research is summarized. A detailed description of the interface module is provided. Several Pascal templates were constructed. The Resident Processor program was also developed. Even though it is designed for the Pascal templates, it can be modified for templates in other languages, such as C, without much difficulty. The Resident Processor itself can be written in any programming language. Since Module 5 routines are not ready yet, there is no way to test the interface module. However, simulation shows that the data base access programs produced by the Resident Processor do work according to the specifications.

Huang, Shou-Hsuan Stephen↗

Database interfaces on NASA's heterogeneous distributed database system

The syntax and semantics of all commands used in the template are described. Template builders should consult this document for proper commands in the template. Previous documents (Semiannual reports) described other aspects of this project. Appendix 1 contains all substituting commands used in the system. Appendix 2 includes all repeating commands. Appendix 3 is a collection of DEFINE templates from eight different DBMS's.

Huang, Shou-Hsuan Stephen↗

Towards improved characterization of high-risk releases using heterogeneous indoor sensor systems

The sudden release of toxic contaminants that reach indoor spaces can be hazardous to building occupants. For an acutely toxic contaminant, the speed of the emergency response strongly influences the consequences to occupants. The design of a real time sensor system is made challenging both by the urgency and complex nature of the event, and by the imperfect sensors and models available to describe it. In this research, we use Bayesian modeling to combine information from multiple types of sensors to improve the characterization of a release. We discuss conceptual and algorithmic considerations for selecting and fusing information from disparate sensors. To explore system performance, we use both real tracer gas data from experiments in a three story building, along with synthetic data, including information from door position sensors. The added information from door position sensors is found to be useful for many scenarios, but not always. We discuss the physical conditions and design factors that affect these results, such as the influence of the door positions on contaminant transport. We highlight potential benefits of multisensor data fusion, challenges in realizing those benefits, and opportunities for further improvement.

54 ENVIRONMENTAL SCIENCES↗

Heterogeneous data-processing optimization with CLARA’s adaptive workflow orchestrator

The hardware landscape used in HEP and NP is changing from homogeneous multi-core systems towards heterogeneous systems with many different computing units, each with their own characteristics. To achieve maximum performance with data processing, the main challenge is to place the right computing on the right hardware. In this paper, we discuss CLAS12 charge particle tracking workflow orchestration that allows us to utilize both CPU and GPU to improve the performance. The tracking application algorithm was decomposed into micro-services that are deployed on CPU and GPU processing units, where the best features of both are intelligently combined to achieve maximum performance. In this heterogeneous environment, CLARA aims to match the requirements of each micro-service to the strength of a CPU or a GPU architecture. A predefined execution of a micro-service on a CPU or a GPU may not be the most optimal solution due to the streaming data-quantum size and the data-quantum transfer latency between CPU and GPU. So, the CLARA workflow orchestrator is designed to dynamically assign micro-service execution to a CPU or a GPU, based on the online benchmark results analyzed for a period of real-time data-processing.

Gyurjyan, Vardan↗

Resolving and Measuring Diffusion in Complex Interfaces: Exploring New Capabilities

This exploratory LDRD targeted the use of a new high resolution spectroscopic diffusion capabilities developed at Sandia to resolve transport processes at interfaces in heterogeneous polymer materials. In particular, the combination of high resolution magic angle spinning (HRMAS) nuclear magnetic resonance (NMR) spectroscopy with pulsed field gradient (PFG) diffusion experiments were used to directly explore interface diffusion within heterogeneous polymer composites, including measuring diffusion for individual chemical species in multi-component mixtures. Several different types of heterogeneous polymer systems were studied using these HRMAS NMR diffusion capabilities to probe the resolution limitations, determine the spatial length scales involved, and explore the general applicability to specific heterogeneous systems. The investigations pursued included a) the direct measurement of the diffusion for poly(dimethyl siloxane) polymer (PDMS) on nano-porous materials, b) measurement of penetrant diffusion in additive manufactures (3D printed) processed PDMS composites, and c) the measurement of diffusion in swollen polymers/penetrant mixtures within nano-confined aluminum oxide membranes. The NMR diffusion results obtained were encouraging and allowed for an improved understanding of diffusion and transport processes at the molecular level, while at the same time demonstrating that the spatial heterogeneity that can be resolved using HRMAS NMR PFG diffusion experiment must be larger than ~μm length scales, expect for polymer transport within nanoporous carbons where additional chemical resolution improves the resolvable heterogeneous length scale to hundreds of nm.

36 MATERIALS SCIENCE↗

Electronic–Vibrational Coupling and Electron Transfer

This Feature Article describes our recent work on implementing pump-degenerate four-wave mixing for studying electronic vibrational coupling in solid-state systems and in particular in interfacial electron transfer systems. Interfacial electron transfer (ET) reactions constitute key physical phenomena central to a variety of light-induced energy transport and conversion processes such as vision, photocatalysis, photovoltaics, energy storage, molecular electronics, etc. Heterogeneous material systems like organic/inorganic interfaces are of particular interest due to the variety of possible modifications to their properties. Although ET dynamics has been studied intensively in heterogeneous systems, many questions remain unanswered. The significance of electronic–vibrational coupling and coherence is one aspect that has yet to be fully explored. It is well-known that ET between a molecule and a semiconductor can be accompanied by vibrational excitation due to Franck–Condon overlap of the excited state and cation. The energy necessary for exciting these high-energy vibrational modes is taken from the excess energy of the transferred electron and leads to an ET spectrum. This spectrum carries information that is associated with the vibrational wavepacket that is involved in the ET process, including the coherent beating that results from periodic attempts to overcome the transition state. The dynamics of the wavepacket, and in particular its coherence time, provides insight into the strength of the coupling between electron donor and acceptor states and the adiabaticity of the process. This information is important but not readily available for interfacial ET systems comprising a large number of possible electron acceptor states because neither the rate of the reaction nor the electronic spectrum allows unambiguous assignment. These recent studies give access to vibrational wave packet dynamics triggered by ET reaction and have the potential to allow direct measurement of Huang–Rys parameters and study fundamental properties of interfacial ET like the adiabaticity and the amount of coherence—electronic and vibrational—in the system.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Electronic and Vibrational Coherence in Heterogeneous Electron Transfer (Final Report)

Interfacial electron transfer reactions constitute key physical phenomena central to a variety of energy related transport and conversion processes such as catalysis, photocatalysis, photovoltaics, energy storage, molecular electronics, etc. Heterogeneous material systems like organic/inorganic interfaces are of particular interest because they provide great potential to tailor properties according to application specific requirements. While electron transfer dynamics has been studied intensively in heterogeneous systems many questions remain unanswered. The significance of electronic-vibrational coupling and coherence is one aspect that has yet to be fully explored. The objective of this project is to develop and implement new ultrafast spectroscopic techniques based on pump four-wave mixing spectroscopy (pump-FWM) for studying the dynamics of vibrational coupling and coherence in heterogeneous electron transfer (HET) at molecule/semiconductor interfaces. The implementation of the spectroscopic setup allows to explore different four-wave mixing techniques and thus to identify a suitable approach to study the significance of vibrational as well as electronic coherence in HET.

14 SOLAR ENERGY↗

IRIS Reimagined: Advancements in Intelligent Runtime System for Task-Based Programming

Task-based programming models are gaining traction in scientific computing. IRIS is a portable runtime system that exploits multiple heterogeneous programming systems and can discover available resources and manage multiple diverse programming systems (e.g., CUDA, Hexagon, HIP, Level Zero, OpenCL, and OpenMP) simultaneously. It accounts for the constraints of task dependencies and provides customizable scheduling policies to map those tasks to heterogeneous devices. In this paper, we present new capabilities added to IRIS to improve its portability for heterogeneous programming, build-friendliness, and performance efficiency. The new additions include vendor-specific kernel support, a runtime system with a foreign function interface to eliminate writing wrapper or boilerplate code for heterogeneous kernels, an easy-to-use and configurable CMake-based build environment, automatic and efficient data transfers and orchestration, and the Hunter and DAGGER toolchains to evaluate IRIS’s task scheduling algorithms.

Miniskar, Narasinga Rao↗