Results for “data structure” — Page 24

DOE OSTI2021

Understanding the Design Space of Sparse/Dense Multiphase Dataflows for Mapping Graph Neural Networks on Spatial Accelerators

Graph Neural Networks (GNNs) have garnered a lot of recent interest because of their success in learning representations from graph-structured data across several critical applications in cloud and HPC. Owing to their unique compute and memory characteristics that come from an interplay between dense and sparse phases of computations, the emergence of reconfigurable dataflow (aka spatial) accelerators offers promise for acceleration by mapping optimized dataflows (i.e., computation order and parallelism) for both phases. The goal of this work is to characterize and understand the design-space of dataflow choices for running GNNs on spatial accelerators in order for the compilers to optimize the dataflow based on the workload. Specifically, we propose a taxonomy to describe all possible choices for mapping the dense and sparse phases of GNNs spatially and temporally over a spatial accelerator, capturing both the intra-phase dataflow and the inter-phase (pipelined) dataflow. Using this taxonomy, we do deep-dives into the cost and benefits of several dataflows and perform case studies on implications of hardware parameters for dataflows and value of flexibility to support pipelined execution.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Understanding Lustre Internals. Second Edition

The Lustre file system has become a preferred storage resource for systems on the Top500 list, and it is often the file system of choice for small- to medium-sized HPC systems that require parallel shared access to data. Several resources exist to help users deploy and configure Lustre, but the same cannot be said for resources that explain the inner workings of the Lustre source code. A previous ORNL technical report entitled "Understanding Lustre Filesystem Internals" (ORNL/TM-2009/117) provided an excellent summary of Lustre subsystem operations. However, that report is over a decade old and is based on Lustre version 1.6. Since that report was published, Lustre has evolved significantly. Several subsystems underwent significant code changes and many new features have been added to the file system, bringing the current Lustre version up to 2.15.This report aims to document and explain the internal workings of the latest version of the Lustre file system. It will provide more complete and up-to-date information than the previous technical report and should serve as a foundational document for anyone interested in Lustre software development. Key data structures will be described along with the APIs used for interaction among the various Lustre subsystems. Although the Lustre software is constantly being developed, the details in this document should remain relevant for the forseeable future.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Cambium Documentation: Version 2021

Cambium is a tool that assembles structured data sets of simulated hourly emission, cost, and operational data for modeled futures of the U.S. electric sector with metrics designed to be useful for long-term decision-making. It was built to expand the metrics reported in the National Renewable Energy Laboratory's (NREL's) Standard Scenarios - an annually released set of projections of how the U.S. electric sector could evolve across a suite of different potential futures, looking forward through 2050 (Cole et al. 2021) Information about Cambium and related publications can be found at https://nrel.gov/analysis/cambium.html, and the Cambium data sets for the Standard Scenarios can be viewed and downloaded at https://cambium.nrel.gov/. In this documentation, we define the metrics reported in Cambium databases (Section 4) and document the Cambium-specific methods for calculating those metrics (Section 5). As this document is intended to cover multiple Cambium data releases, we do not document specific scenarios here - for individual data releases, readers should look for an accompanying report that describes the assumptions and data that underlie that specific release. For Cambium databases that are built on NREL's Standard Scenarios, for example, reports describing the scenarios can be found on the Standard Scenarios web site (https://www.nrel.gov/analysis/standard-scenarios.html).

24 POWER TRANSMISSION AND DISTRIBUTION↗

DOE OSTI2021

Developing Bloom Filters for Web Archives’ Holdings (Final Project Report)

The main goal of the project was to develop a framework for web archives to create Bloom filters based on their holdings of archived web resources. A Bloom filter (BF) is a data structure that, in our scenario, contains hash values of all (or a subset of) URLs, of which an archive has one or more archival copies. Two main use cases fall in scope for this project and are supported by a BF implementation: 1) Sharing of an archive’s holdings (URLs) and 2) Querying the holdings of one or more archives. Since URL strings are hashed before ingested into the BF, the index of an archive is not shared in plain text when the BF is shared with trusted parties. On the other hand, a BF implementation does allow queries for a URL to confirm if an archive indeed has one or more archival copies of that URL. This collaborative project between Los Alamos National Laboratory (LANL) and the Croatian Web Archive (HAW), from the National and University Library in Zagreb (NSK), developed by NSK and University of Zagreb University Computing Center SRCE), aimed at developing software to create BFs, evaluate the scalability of the approach, pilot a search service based on BFs, and design a framework for archives to share their filters with trusted parties. In the remainder of this document, we will report on the work completed with respect to the individual deliverables, outline aspects of future work, and conclude with our recommendations for the use of BFs for the web archiving community.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

HyMARC Seedling: Optimized Hydrogen Adsorbents via Machine Learning and Crystal Engineering

This final report is presented in two parts: In Part 1, the H 2 capacities of a diverse set of 918,734 metal-organic frameworks (MOFs) sourced from 19 databases is predicted via machine learning (ML). Using only 7 structural features as input, ML identifies 8,282 MOFs with the potential to exceed the capacities of state-of-the-art materials. The identified MOFs are predominantly hypothetical compounds having low densities (<0.31 g/cm3) in combination with high surface areas (>5,300 m2/g), void fractions ($0.90), and pore volumes (>3.3 cm3/g). The relative importance of the input features are characterized, and dependencies on the ML algorithm and training set size are quantified. The most important features for predicting H 2 uptake are pore volume (for gravimetric capacity) and void fraction (for volumetric capacity). The ML models are available on the web, allowing for rapid and accurate predictions of the hydrogen capacities of MOFs from limited structural data; the simplest models require only a single crystallographic feature. In part 2, ways to improve the poor powder packing density of MOFs is discussed. More specifically, a strategy that improves packing efficiency and volumetric hydrogen gas storage density dramatically through engineered morphologies and controlled-crystal size distributions is presented that holds promise for maximizing storage capacity for a given MOF. The packing density improvement, demonstrated for the benchmark sorbent MOF-5, leads to a significant enhancement of volumetric hydrogen storage performance relative to commercial MOF-5. System model projections demonstrate that engineering of crystal morphology/size or use of a bimodal distribution of cubic crystal sizes in tandem with system optimization can surpass the 25 g/L volumetric capacity of a typical 700 bar compressed storage system and exceed the DOE targets 2020 volumetric capacity (30 g/L). Finally, a critical link between improved powder packing density and reduced damage upon compaction is revealed leading to sorbents with both high surface area and high density.

08 HYDROGEN↗

DOE OSTI2020

Algorithmic and GPU enhancements for molecular dynamics in Cabana and LAMMPS

The focus of this milestone was to implement and test new algorithms and optimizations for improved GPU performance by molecular dynamics (MD) and related particle codes. The work was done in the CoPA Cabana library, the CabanaMD mini-app which uses Cabana to implement and benchmark prototypical MD models, and the LAMMPS MD code which is used by the ECP EXAALT project as well as for materials modeling generally. Efforts included (a) implementation of new or optimized algorithms for fundamental MD kernels like neighbor-list building and FFTs which are also widely used in other particle methods, (b) performance testing of data structure layouts for both simple pairwise and computationally intense neural network potentials, (c) optimizations aimed at improving GPU performance for models with small atom counts, and (d) optimizations specific to the SNAP interatomic potential due to its importance for the EXAALT project.

36 MATERIALS SCIENCE↗

DOE OSTI2022

PETSc/TAO Users Manual: Revision 3.18

This manual describes the use of the Portable, Extensible Toolkit for Scientific Computation (PETSc) and the Toolkit for Advanced Optimization (TAO) for the numerical solution of partial differential equations and related problems on high-performance computers. PETSc/TAO is a suite of data structures and routines that provide the building blocks for the implementation of large-scale application codes on parallel (and serial) computers. PETSc uses the MPI standard for all distributed memory communication.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

PETSc/TAO Users Manual (Rev. 3.19)

This manual describes the use of the Portable, Extensible Toolkit for Scientific Computation (PETSc) and the Toolkit for Advanced Optimization (TAO) for the numerical solution of partial differential equations and related problems on high-performance computers. PETSc/TAO is a suite of data structures and routines that provide the building blocks for the implementation of large-scale application codes on parallel (and serial) computers. PETSc uses the MPI standard for all distributed memory communication. PETSc/TAO includes a large suite of parallel linear solvers, nonlinear solvers, time integrators, and opti mization that may be used in application codes written in Fortran, C, C++, and Python (via petsc4py; see Getting Started). PETSc provides many of the mechanisms needed within parallel application codes, such as parallel matrix and vector assembly routines. The library is organized hierarchically, enabling users to employ the level of abstraction that is most appropriate for a particular problem. By using techniques of object-oriented programming, PETSc provides enormous flexibility for users. PETSc is a sophisticated set of software tools; as such, for some users it initially has a much steeper learning curve than packages such as MATLAB or a simple subroutine library. In particular, for individuals without some computer science background, experience programming in C, C++, python, or Fortran and experience using a debugger such as gdb or lldb, it may require a significant amount of time to take full advantage of the features that enable efficient software use. However, the power of the PETSc design and the algorithms it incorporates may make the efficient implementation of many application codes simpler than “rolling them” yourself. For many tasks a package such as MATLAB is often the best tool; PETSc is not intended for the classes of problems for which effective MATLAB code can be written. There are several packages, built on PETSc, that may satisfy your needs without requiring directly using PETSc. We recommend reviewing these packages functionality before starting to code directly with PETSc. PETSc can be used to provide a “MPI parallel linear solver” in an otherwise sequential, or OpenMP parallel code. This approach cannot provide extremely large improvements in the application time by utilizing large numbers of MPI processes but can still improve the performance. Certainly all parts of a previously sequential code need not be parallelized but the matrix generation portion must be parallelized to expect true scalability to large numbers of MPI processes. See PCMPI for details on how to utilize the PETSc MPI linear solver server. Since PETSc is under continued development, small changes in usage and calling sequences of routines will occur. PETSc has been supported for twenty-five years; see mailing list information on our website for information on contacting support.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

PETSc/TAO Users Manual (Rev. 3.20)

This manual describes the use of the Portable, Extensible Toolkit for Scientific Computation (PETSc) and the Toolkit for Advanced Optimization (TAO) for the numerical solution of partial differential equations and related problems on high-performance computers. PETSc/TAO is a suite of data structures and routines that provide the building blocks for the implementation of large-scale application codes on parallel (and serial) computers. PETSc uses the MPI standard for all distributed memory communication.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

Guidelines for Nuclear Structure Evaluators

This document sets forth the policies and procedures for preparing mass chain evaluations for the Evaluated Nuclear Structure Data File (ENSDF). Also included are some appendices which are referred to or augment material in the guidelines. See section on Appendices for details and references to additional material.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2024

PETSc/TAO Users Manual V.3.21

This manual describes the use of the Portable, Extensible Toolkit for Scientific Computation (PETSc) and the Toolkit for Advanced Optimization (TAO) for the numerical solution of partial differential equations (PDEs) and related problems on high-performance computers. PETSc/TAO is a suite of data structures and routines that provide the building blocks for implementing large-scale application codes on parallel (and serial) computers. PETSc uses the MPI standard for all distributed memory communication. PETSc/TAO includes a large suite of parallel linear solvers, nonlinear solvers, time integrators, and optimizers that may be used in application codes written in Fortran, C, C++, and Python (via petsc4py; see Getting Started ). The library is organized hierarchically, enabling users to employ the abstraction level most appropriate for a particular problem. By using techniques of object-oriented programming, PETSc provides enormous flexibility for users.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Sandia Toolkit Manual (V.5.21.1)

This report provides documentation for the Sandia Toolkit (STK) modules. STK modules are intended to provide infrastructure that assists the development of computational engineering software such as finite-element analysis applications. STK includes modules for unstructured-mesh data structures, reading/writing mesh files, geometric proximity search, transfers, MPMD coupling support, and various other utilities. This document contains a chapter for each module, and each chapter contains overview descriptions and usage examples. Usage examples are primarily code listings which are generated from working test programs that are included in the STK code-base. A goal of this approach is to ensure that the usage examples will not fall out of date.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Sandia Toolkit Manual Version 5.15.6

This report provides documentation for the Sandia Toolkit (STK) modules. STK modules are intended to provide infrastructure that assists the development of computational engineering software such as finite-element analysis applications. STK includes modules for unstructured-mesh data structures, reading/writing mesh files, geometric proximity search, and various utilities. This document contains a chapter for each module, and each chapter contains overview descriptions and usage examples. Usage examples are primarily code listings which are generated from working test programs that are included in the STK code-base. A goal of this approach is to ensure that the usage examples will not fall out of date.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Comprehensive Review of 2$β$ Decay Half-Lives

The double-beta (2β)-decay is the rarest nuclear physics process, and its experimental half-lives (T 1/2 ) exceed the age of the Universe from nine to fourteen orders of magnitude. Double-beta decay was observed, and its half-life was measured in 14 parent nuclei using direct, radiochemical, and geochemical methods. The decay observables are analyzed using the Evaluated Nuclear Structure Data File (ENSDF) procedures, and the recommended T 1/2 were deduced. Using the calculated values of phase factors, the effective nuclear matrix elements were extracted and compared with available data. Thousands of theoretical and experimental works have been dedicated to these topics in the last 85 years, and we present two data sets of recommended values to encapsulate the results.

2β-decay↗

DOE OSTI2024

Diffraction Measurement of Reaction Products in Shock Compressed TATB on the NIF

The shock-induced reaction of high explosives to gaseous and solid reaction products is a rapid, complex process. Understanding solid reaction product structure and formation kinetics is essential in determining the high-pressure equation of state of these multicomponent systems. We use the National Ignition Facility (NIF) to shock compress ~500-µm thick pressed powder high-explosive TATB samples to 70-135 GPa and collect in situ structural data on detonation byproducts using the TARDIS X-ray diffraction diagnostic. Velocimetry is used to record the transmitted compression wave profile, which is well described by Cheetah hydrocode simulations coupled with a reactive flow model, providing strong evidence of reaction in the TATB sample. While an unambiguous determination of the product phases was not possible owing to the low signal-to-noise quality of the diffraction signal, X-ray diffraction data of the product phases formed within the first 50 ns of the reaction is most consistent with a mixture of amorphous products and crystalline hexagonal diamond.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DOE OSTI2022

Merger of the Dissolved Gas Method into RELAP5-3D

The Dissolved Gas Model (DGM) was developed for Bettis Atomic Power Laboratory in their own special version of RELAP5-3D known as NUPAC by a subcontractor to. Incorporation of this important code capability in RELAP5-3D was necessary for collaboration between INL and Bettis on future code development efforts. The changes necessary to implement DGM are fundamental, changing the very solver technology at the core of RELAP5-3D, and pervasive, requiring the addition of new subprograms and data, alteration of existing data structures, and modification of more than one hundred subprograms. Moreover, NUPAC and RELAP5-3D had evolved separately and in different directions for a decade and were significantly dissimilar, so the effort was both complex and time-consuming. To ensure its correctness, a large test set comprised thousands of input cases had been constructed and had to run correctly to prove DGM was implemented correctly. A novel method for transporting and adjusting the updates for DGM into RELAP5-3D was undertaken. Tens of thousands of lines of code were either changed or added resulting in a 100% successful implementation of DGM in RELAP5-3D.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

DOE OSTI2024

PETSc/TAO Users Manual Revision 3.22

This manual describes the use of the Portable, Extensible Toolkit for Scientific Computation (PETSc) and the Toolkit for Advanced Optimization (TAO) for the numerical solution of partial differential equations (PDEs) and related problems on high-performance computers. PETSc/TAO is a suite of data structures and routines that provide the building blocks for implementing large-scale application codes on parallel (and serial) computers. PETSc uses the MPI standard for all distributed memory communication.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Discovering Dark Matter Clumps and Primordial Particles with Galaxies

Cosmological observations and galaxy dynamics seem to imply that five out of six parts in mass of all matter in the universe is composed of dark matter, which is not accounted for by the Standard Model of particles. The cold dark matter (CDM) paradigm has been extremely successful at describing observations on large-cosmological scale. However, many different dark matter candidates have the same observable effects as CDM on large-length scales. One possible avenue to distinguish between these different models is to look on much smaller-length scales, where usually dark matter models become distinguishable. This research developed the theoretical framework and statistical tools needed to map the detailed distribution of dark matter on subgalactic scales using strong gravitational lensing. A second goal of this research was to build new statistical techniques that efficiently exploit the full power of cosmic-structure data from next-generation surveys. Galaxy clustering on large scales provides significant cosmological information through the power spectrum. Additional information can be gained from higher-order statistics. This research provided new ways to understand the initial conditions of the universe from current and upcoming cosmological surveys.

79 ASTRONOMY AND ASTROPHYSICS↗