Search NASA⌕ Search

SEARCH · Search NASA

Results for “data parallel”

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 1,135 records · Page 63
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

Metagenome-assembled genomes from topsoils collected during NEON campaign in East River, CO (06/14/2018-06/28/2018)

The Watershed Function Science Focus Area (WF SFA) at Lawrence Berkeley National Lab is working to build a mechanistic understanding of the distribution and dynamics of biogeochemical processes in mountainous watersheds and their response to perturbation. In June 2018, the NEON (National Ecological Observatory Network) Airborne Observatory Platform (AOP) performed a taskable airborne imaging campaign to collect visible to shortwave infrared (VSWIR) imaging spectroscopy and LiDAR data across 330 km2 in the Upper East River at Crested Butte, CO. We conducted a parallel ground sampling campaign to sample vegetation traits, as well as soil physical, chemical, and microbiological characteristics. We collected these samples from 438 sites across 12 locations spanning much of the elevation, topographic, and geologic variability across the study area. A subset of 250 samples were used for soil metagenomics which is presented here. In addition, at each site, vegetation samples were collected to measure species-specific leaf water content and leaf mass area, foliar elemental composition and foliar CN stable isotope ratios. Soil samples were collected to measure soil physical properties which include bulk density and soil texture analysis. A suite of soil chemical properties was measured from the samples collected at each site, including pH, organic matter, concentrations exchangeable cations, total elemental composition, and the concentrations of extractable N pools (e.g. total free amino acids, ammonium, nitrate, dissolved organic N, and total dissolved N). Additionally, we have measured soil microbial biomass CN stoichiometry. Here, we present 1982 metagenome-assembled genomes (MAGs) for the bacterial and archaeal community from topsoil collected from during NEON 2018 campaign. All metagenomes were sequenced at JGI (Joint Genome Institute) (GOLD Study ID: Gs0149986). Metagenomes were assembled using JGI Metagenome Workflow (10.1128/mSystems.00804-20). The dataset includes (1) zip files for 1982 MAG fasta files (neon_genomes1-5.tar.gz, split into 5 tarballs to keep tarballs under 0.5 GB), (2) neon_Gs0149986_samples_soilproperties_metagenomes.csv: the sample information together with the accession numbers for the underlying metagenomes and the associated soil physical and chemical measurements in NMDC (National Microbiome Data Collaborative) compliant format, (3) neon_Gs0149986.kml: location bounding box file for the sampled locations, (4) samples.csv: sample metadata file used to register Internationall Generic Sample Numbers (IGSNs), (5) flmd.csv: file level metadata file, and (6) dd.csv: data dictionary file. This work was supported by the Watershed Function Science Focus Area at Lawrence Berkeley National Laboratory funded by the US Department of Energy, Office of Science, Biological and Environmental Research under Contract No. DE-AC02-05CH11231.

2018 NEON and 2025 CHESS Campaigns↗
DOE OSTI2015

Oxygen transport properties estimation by classical trajectory–direct simulation Monte Carlo

Coupling direct simulation Monte Carlo (DSMC) simulations with classical trajectory calculations is a powerful tool to improve predictive capabilities of computational dilute gas dynamics. The considerable increase in computational effort outlined in early applications of the method can be compensated by running simulations on massively parallel computers. In particular, Graphics Processing Unit acceleration has been found quite effective in reducing computing time of classical trajectory (CT)-DSMC simulations. The aim of the present work is to study dilute molecular oxygen flows by modeling binary collisions, in the rigid rotor approximation, through an accurate Potential Energy Surface (PES), obtained by molecular beams scattering. The PES accuracy is assessed by calculating molecular oxygen transport properties by different equilibrium and non-equilibrium CT-DSMC based simulations that provide close values of the transport properties. Comparisons with available experimental data are presented and discussed in the temperature range 300–900 K, where vibrational degrees of freedom are expected to play a limited (but not always negligible) role.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2017

A Computational Approach for Modeling Neutron Scattering Data from Lipid Bilayers

Biological cell membranes are responsible for a range of structural and dynamical phenomena crucial to a cell's well-being and its associated functions. Due to the complexity of cell membranes, lipid bilayer systems are often used as biomimetic models. These systems have led to signficant insights into vital membrane phenomena such as domain formation, passive permeation and protein insertion. Experimental observations of membrane structure and dynamics are, however, limited in resolution, both spatially and temporally. Importantly, computer simulations are starting to play a more prominent role in interpreting experimental results, enabling a molecular under- standing of lipid membranes. Particularly, the synergy between scattering experiments and simulations offers opportunities for new discoveries in membrane physics, as the length and time scales probed by molecular dynamics (MD) simulations parallel those of experiments. We also describe a coarse-grained MD simulation approach that mimics neutron scattering data from large unilamellar lipid vesicles over a range of bilayer rigidity. Specfically, we simulate vesicle form factors and membrane thickness fluctuations determined from small angle neutron scattering (SANS) and neutron spin echo (NSE) experiments, respectively. Our simulations accurately reproduce trends from experiments and lay the groundwork for investigations of more complex membrane systems.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Nonparallel transcriptional divergence during parallel adaptation

Abstract How underlying mechanisms bias evolution toward predictable outcomes remains an area of active debate. In this study, we leveraged phenotypic plasticity and parallel adaptation across independent lineages of Trinidadian guppies ( Poecilia reticulata ) to assess the predictability of gene expression evolution during parallel adaptation. Trinidadian guppies have repeatedly and independently adapted to high‐ and low‐predation environments in the wild. We combined this natural experiment with a laboratory breeding design to attribute transcriptional variation to the genetic influences of population of origin and developmental plasticity in response to rearing with or without predators. We observed substantial gene expression plasticity, as well as the evolution of expression plasticity itself, across populations. Genes exhibiting expression plasticity within populations were more likely to also differ in expression between populations, with the direction of population differences more likely to be opposite those of plasticity. While we found more overlap than expected by chance in genes differentially expressed between high‐ and low‐predation populations from distinct evolutionary lineages, the majority of differentially expressed genes were not shared between lineages. Our data suggest alternative transcriptional configurations associated with shared phenotypes, highlighting a role for transcriptional flexibility in the parallel phenotypic evolution of a species known for rapid adaptation.

Fischer, Eva K.↗
DOE OSTI2015

WE-EF-207-02: The Rotate-Plus-Shift C-Arm Trajectory: Theory and First Clinical Results

Purpose: The proposed method enables the acquisition of a complete dataset for 3D reconstruction of C-Arm data using less than 180° rotation. Methods: Typically a C–arm cone–beam CT scan is performed using a circle–like trajectory around a region of interest. Therefore an angular range of at least 180° plus fan–angle must be covered to ensure a completely sampled data set. This fact defines some constraints on the geometry and technical specifications of a C–arm system, for example a larger C radius or a smaller C opening respectively. This is even more important for mobile C-arm devices which are typically used in surgical applications.To overcome these limitations we propose a new trajectory which requires only 180° minusfan–angle of rotation for a complete data set. The trajectory consists of three parts: A rotation of the C around a defined iso–center and two translational movements parallel to the detector plane at the begin and at the end of the rotation (rotate plus shift trajectory). This enables the acquisition of a completely sampled dataset using only 180° minus fan–angle of rotation. Results: For the evaluation of the method we show simulated and measured data. The results show, that the rotate plus shift scan yields equivalent image quality compared to the short scan which is assumed to be the gold standard for C-arm CT today. Compared to the pure rotational scan over only 165°, the rotate plus shift scan shows strong improvements in image quality. Conclusion: The proposed method makes 3D imaging using C–arms with less than 180° rotation range possible. This enables integrating full 3D functionality into a C- arm device without any loss of handling and usability for 2D imaging.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2018

Juggler: a dependence-aware task-based execution framework for GPUs

Scientific applications with single instruction, multiple data (SIMD) computations show considerable performance improvements when run on today's graphics processing units (GPUs). However, the existence of data dependences across thread blocks may significantly impact the speedup by requiring global synchronization across multiprocessors (SMs) inside the GPU. To efficiently run applications with interblock data dependences, we need fine-granular task-based execution models that will treat SMs inside a GPU as stand-alone parallel processing units. Such a scheme will enable faster execution by utilizing all internal computation elements inside the GPU and eliminating unnecessary waits during device-wide global barriers.In this paper, we propose Juggler, a task-based execution scheme for GPU workloads with data dependences. The Juggler framework takes applications embedding OpenMP 4.5 tasks as input and executes them on the GPU via an efficient in-device runtime, hence eliminating the need for kernel-wide global synchronization. Juggler requires no or little modification to the source code, and once launched, the runtime entirely runs on the GPU without relying on the host through the entire execution. We have evaluated Juggler on an NVIDIA Tesla P100 GPU and obtained up to 31% performance improvement against global barrier based implementation, with minimal runtime overhead.

Belviranli, Mehmet E.↗
DOE OSTI2017

GPU-accelerated Tersoff potentials for massively parallel Molecular Dynamics simulations

The Tersoff potential is one of the empirical many-body potentials that has been widely used in simulation studies at atomic scales. Unlike pair-wise potentials, the Tersoff potential involves three-body terms, which require much more arithmetic operations and data dependency. In this contribution, we have implemented the GPU-accelerated version of several variants of the Tersoff potential for LAMMPS, an open-source massively parallel Molecular Dynamics code. Compared to the existing MPI implementation in LAMMPS, the GPU implementation exhibits a better scalability and offers a speedup of 2.2X when run on 1000 compute nodes on the Titan supercomputer. On a single node, the speedup ranges from 2.0 to 8.0 times, depending on the number of atoms per GPU and hardware configurations. The most notable features of our GPU-accelerated version include its design for MPI/accelerator heterogeneous parallelism, its compatibility with other functionalities in LAMMPS, its ability to give deterministic results and to support both NVIDIA CUDA- and OpenCL-enabled accelerators. Our implementation is now part of the GPU package in LAMMPS and accessible for public use.

Computer Science↗
DOE OSTI2022

Distributed Real-time Plume Monitoring for Deep Sea Mineral Extraction​

In the emerging industry of deep-sea mining for minerals and deposits (e.g. polymetallic nodules for nickel, cobalt, copper, and manganese), more data is required to understand the effects of sediment plume generation and predict the distribution of disturbed sediment. There are two main sources of plume generation, the first being at the active mining site where the “collector” directly removes the top layer of the sea floor. The other is the “midwater plume” consisting of unwanted sediment that was collected during extraction that is pumped back into the aphotic zone. The vast majority of plume generation is caused by the collector, causing detrimental and long-lasting impacts on seafloor ecosystems due to the lack of wave activity or strong currents at the sea floor. Therefore, it is crucial to invest in the infrastructure to support the study and constant monitoring over a large area of the sea floor where plume generation is present. Due to the limited number of usable channels and power requirements, current subsea wireless communications technologies are not well suited to instrumenting the large areas of the sea floor needed to monitor plume migration. The scope of this effort is to transition experimental demonstrations of high-bandwidth, full-duplex scalable underwater laser communications to the seafloor in an open ocean environment. Specifically tackling challenges associated with the dynamic nature of the subsea world, including but not limited to, deployment logistics, sustainability, and range. The goal is to enable the internet of underwater things for deep sea industries by broadening the capabilities of subsea communications. By using high-precision laser transmitters, many of the challenges current subsea optical systems face can be circumvented, such as power consumption, interference, and bandwidth limitations. This approach lends itself to wireless interlinking multi-node networks, in series or parallel, facilitating the implementation of a wide array of sensor types. This interlinking allows all the data gathered from the network to be processed through a single hardline uplink to the surface, lowering the complexity required for near real-time data processing. Additionally, the laser control systems produce metadata that can be used to help characterize the water column between the nodes. Combining data from various sensors such as turbidity, temperature, current velocity with metadata such as beam attenuation and deflection can produce a high-resolution model of sea floor conditions around an active mining zone. The resulting near real-time model can be used to optimize location and flow rate of the mining operation to minimize and quantify the environmental impact.

Mons, Ishan↗
DOE OSTI2018

Structural disorder and elementary magnetic properties of triangular lattice ErMgGaO 4 single crystals

The single crystal growth, structure, and basic magnetic properties of ErMgGaO 4 are reported. The structure consists of triangular layers of magnetic ErO 6 octahedra separated by a double layer of randomly occupied non-magnetic (Ga,Mg)O 5 bipyramids. The Er atoms are positionally disordered. Magnetic measurements parallel and perpendicular to the c axis of a single crystal reveal dominantly antiferromagnetic interactions, with a small degree of magnetic anisotropy. A weighted average of the directional data suggests an antiferromagnetic Curie Weiss temperature of approximately -30 K. Below 10 K the temperature dependences of the inverse susceptibilities in the in-plane and perpendicular-to plane directions are parallel, indicative of an isotropic magnetic moment at low temperatures. In conclusion, no sign of magnetic ordering is observed above 1.8 K, suggesting that ErMgGaO 4 is a geometrically frustrated magnet.

36 MATERIALS SCIENCE↗
DOE OSTI2025

PPI DataHub Project Data Package: S. elongatus PCC 7942 Limited Proteolysis and Thermal Proteome Profiling Structural Proteomics (JM-PB-DP3)

The purpose of this experiment was to investigate structural alterations in proteins involved in central carbon metabolism and photosynthetic electron transfer pathways in Synechococcus elongatus PCC 7942. Sample data was obtained from S. elongatus cell lysates using three complementary mass spectrometry (MS) techniques using limited proteolysis (LiP-MS), thermal proteome profiling (TPP-MS), and redox enrichment (Redox-MS) in evaluating alterations solvent accessibility and structural stability caused by light perturbation at the molecular level. Experimentally processed sample data for LiP and TPP proteomic datasets were derived from the same cell culture stock, prepared simultaneously in parallel, and acquired by mass spectrometry. Processed datasets are openly accessible from the download button and contain secondary processed proteomic results files, computed outputs, and supporting metadata materials. Experimental samples processed for LiP-MS label-free quantification (LFQ) or TPP-MS tandem mass tag (TMT) 10-plex were acquired using a Q-Exactive HF-X mass spectrometer and processed/compiled using either MSGF+ (v2024.03.26) or ​​​​PlexedPiper for proteome evaluation. Additional software supporting downstream proteomic analysis include FragPipe (v.4.0), MSFragger (v.22.1), and an adapted Microbial Isolate LiP Analysis Workflow (located at Zenodo). Processed proteomic data downloads include a sample naming key, normalized quantification results files, and processed protein annotated abundance files.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Doppler Lidar Motion-Correction Wind Profiles (DLMCPROF-WIND) Value-Added Product Report

Wind speed and direction, together with pressure, temperature, and relative humidity, are the most fundamental atmospheric state parameters. Accurate measurement of these parameters is crucial for numerical weather prediction. Vertically resolved wind measurements in the atmospheric boundary layer are particularly important for modeling pollutant and aerosol transport. The U.S. Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) user facility currently operates several scanning coherent Doppler lidar (DL) systems that provide accurate height-resolved measurements of wind speed and direction. These instruments operate in the near infrared (IR;1.5 microns) and provide range-resolved measurements of radial velocity, attenuated aerosol backscatter, and signal-to-noise ratio (SNR). The systems are operated using a fixed scan schedule consisting of plan position indicator (PPI) or Doppler Beam-Swing (DBS) scans. PPI scans are performed by scanning the beam in azimuth while maintaining a fixed elevation angle, and DBS scans are similar but typically also include a vertical beam. Radial velocity data from these scans are processed to yield profiles of wind speed direction. For stationary ground-based deployments, the Doppler Lidar Horizontal Wind Profiles (DLPROF-WIND) Value-Added-Product (VAP) provides height- and time-resolved measurements of the winds (Newsom and Krishnamurthy 2022). For operation on a moving platform, modifications to the existing DLPROF-WIND VAP are required to compensate for the effects of the platform motion. This report describes a parallel VAP, DLMCPROF-WIND, that was developed for computing motion-compensated wind profiles from ARM Doppler lidar data acquired during the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC) campaign, when the lidar was deployed on the German icebreaker Polarstern.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2017

Extreme-Scale Algorithms & Software Resilience (EASIR) Architecture-Aware Algorithms for Scalable Performance and Resilience on Heterogeneous Architectures

This project addresses both communication-avoiding algorithms, and reproducible floating-point computation. Communication, i.e. moving data, either between levels of memory or processors over a network, is much more expensive per operation than arithmetic (measured in time or energy), so we seek algorithms that greatly reduce communication. We developed many new algorithms for both dense and sparse, and both direct and iterative linear algebra, attaining new communication lower bounds, and getting large speedups in many cases. We also extended this work in several ways: (1) We minimize writes separately from reads, since writes may be much more expensive than reads on emerging memory technologies, like Flash, sometimes doing asymptotically fewer writes than reads. (2) We extend the lower bounds and optimal algorithms to arbitrary algorithms that may be expressed as perfectly nested loops accessing arrays, where the array subscripts may be arbitrary affine functions of the loop indices (eg A(i), B(i,j+k, k+3*m-7, …) etc.). (3) We extend our communication-avoiding approach to some machine learning algorithms, such as support vector machines. This work has won a number of awards. We also address reproducible floating-point computation. We define reproducibility to mean getting bitwise identical results from multiple runs of the same program, perhaps with different hardware resources or other changes that should ideally not change the answer. Many users depend on reproducibility for debugging or correctness. However, dynamic scheduling of parallel computing resources, combined with nonassociativity of floating point addition, makes attaining reproducibility a challenge even for simple operations like summing a vector of numbers, or more complicated operations like the Basic Linear Algebra Subprograms (BLAS). We describe an algorithm that computes a reproducible sum of floating point numbers, independent of the order of summation. The algorithm depends only on a subset of the IEEE Floating Point Standard 754-2008, uses just 6 words to represent a “reproducible accumulator,” and requires just one read-only pass over the data, or one reduction in parallel. New instructions based on this work are being considered for inclusion in the future IEEE 754-2018 floating-point standard, and new reproducible BLAS are being considered for the next version of the BLAS standard.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2016

Expanding on Successful Concepts, Models, and Organization

In her letter to the editor1 regarding our recent Feature Article “Completing the Link between Exposure Science and Toxicology for Improved Environmental Health Decision Making: The Aggregate Exposure Pathway Framework” 2, Dr. von Göetz expressed several concerns about terminology, and the perception that we propose the replacement of successful approaches and models for exposure assessment with a concept. We are glad to have the opportunity to address these issues here. If the goal of the AEP framework was to replace existing exposure models or databases for organizing exposure data with a concept, we would share Dr. von Göetz concerns. Instead, the outcome we promote is broader use of an organizational framework for exposure science. The framework would support improved generation, organization, and interpretation of data as well as modeling and prediction, not replacement of models. The field of toxicology has seen the benefits of wide use of one or more organizational frameworks (e.g., mode and mechanism of action, adverse outcome pathway). These frameworks influence how experiments are designed, data are collected, curated, stored and interpreted and ultimately how data are used in risk assessment. Exposure science is poised to similarly benefit from broader use of a parallel organizational framework, which Dr. von Göetz correctly points out, is currently used in the exposure modeling community. In our view, the concepts used so effectively in the exposure modeling community, expanded upon in the AEP framework, could see wider adoption by the field as a whole. The value of such a framework was recognized by the National Academy of Sciences.3 Replacement of models, databases, or any application with the AEP framework was not proposed in our article. The positive role broader more consistent use of such a framework might have in enabling and advancing “general activities such as data acquisition, organization…,” and exposure modeling was discussed in some detail. Like Dr. von Göetz, we recognized the challenges associated with acceptance of the terminology, definitions, and structure proposed in the paper. To address these challenges, an expert workshop was held in May, 2016 to consider and revise the “basic elements” outlined in the paper. The attendees produced revisions to the terminology (e.g., key events) that align with terminology currently in use in the field. We were also careful in our paper to acknowledge a point raised by Dr. von Göetz, that the term AEP implies aggregation, providing these clarifications: “The simplest form of an AEP represents a single source and a single pathway and may more commonly be referred to as an exposure pathway,”; and “An aggregate exposure pathway may represent multiple sources and transfer through single pathways to the TSE, single sources and transfer through multiple pathways to the target site exposure (TSE), or any combination of these.” These clarifications address the concern that the AEP term is not accurate or logical, and further expands upon the word “aggregate” in a broader context. Our use of AEP is consistent with the definition for “aggregate exposure”, which refers to the combined exposures to a single chemical across multiple routes and pathways.3 The AEP framework embraces existing methods for collection, prediction, organization, and interpretation of human and ecological exposure data cited by Dr. von Göetz. We remain hopeful that wider recognition and use of an organizing concept for exposure information across the exposure science, toxicology and epidemiology communities advances the development of the kind of infrastructure and models Dr. von Göetz discusses. This outcome would be a step forward, rather than a step backward.

exposure science↗
DOE OSTI2018

Locality-aware and load-balanced static task scheduling for MapReduce

Task scheduling for MapReduce jobs has been an active area of research with the objective of decreasing the amount of data transferred during the shuffle phase via exploiting data locality. In the literature, generally only the scheduling of reduce tasks is considered with the assumption that scheduling of map tasks is already determined by the input data placement. However, in cloud or HPC deployments of MapReduce, the input data is located in a remote storage and scheduling map tasks gains importance. Here, we propose models for simultaneous scheduling of map and reduce tasks in order to improve data locality and balance the processors’ loads in both map and reduce phases. Our approach is based on graph and hypergraph models which correctly encode the interactions between map and reduce tasks. Partitions produced by these models are decoded to schedule map and reduce tasks. A two-constraint formulation utilized in these models enables balancing processors’ loads in both map and reduce phases. The partitioning objective in the hypergraph models correctly encapsulates the minimization of data transfer when a local combine step is performed prior to shuffle, whereas the partitioning objective in the graph models achieve the same feat when a local combine is not performed. Here, we show the validity of our scheduling on the MapReduce parallelizations of two important kernel operations – sparse matrix–vector multiplication (SpMV) and generalized sparse matrix–matrix multiplication (SpGEMM) – that are widely encountered in big data analytics and scientific computations. Compared to random scheduling, our models lead to tremendous savings in data transfer by reducing data traffic from several hundreds of megabytes to just a few megabytes in the shuffle phase and consequently leading up to 2.6x and 4.2x speedup for SpMV and SpGEMM, respectively.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Distributed Data-Driven Optimization for Voltage Regulation in Distribution Systems

Here, this paper proposes a distributed data-driven optimization framework for voltage regulation in distribution systems. The recursive kernel regression and alternating direction method of multipliers (ADMM) are selected to cover the system learning and distributed optimization tasks. The proposed distributed data-driven framework is capable of having a rapid response to system or load changes while considering the operation optimality. Besides, the distributed algorithm parallels the computation tasks and reduces the computational expense of a single agent. To validate the performance of the proposed method, a hypothetical 7-Bus system and the IEEE 123-Bus system are selected to show the effectiveness of the proposed data-driven framework. According to the numerical study results, the proposed method offers great flexibility for selecting customized kernel models for different regions and can effectively improve the system voltage profile in a distributed manner.

24 POWER TRANSMISSION AND DISTRIBUTION↗
DOE OSTI2022

A New Integrated Analysis Suite for Fast-Ion Study in KSTAR

Here, an integrated workflow for fast-ion analysis was developed by adapting the One Modeling Framework for Integrated Task (OMFIT) workflow manager to support a standard and unified analysis platform for KSTAR users. The newly established analysis suite offers a graphical user interface–based workflow to enable users to readily access and handle experimental data archived in various data formats and servers. Further, users can analyze the data by importing modules designed for conducting certain tasks, such as profile fitting, equilibrium reconstruction, and postprocessing of tokamak data. The procedures for preparing the inputs for fast-ion simulations are streamlined by a common workflow manager, which enables the parallel processing of various tasks to efficiently analyze large fast-ion datasets. The OMFIT platform comprises a flexible Python-based application that enables users to freely manipulate the Python scripts for applications that are unavailable in the standard workflow. The framework also offers mapping tools to translate the output data into the Integrated Modeling and Analysis Suite format to maintain application compatibility for future ITER burning plasma experiments.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2024

A fast and accurate domain decomposition nonlinear manifold reduced order model

Here, this paper integrates nonlinear-manifold reduced order models (NM-ROMs) with domain decomposition (DD). NM ROMs approximate the full order model (FOM) state in a nonlinear-manifold by training a shallow, sparse autoencoder using FOM snapshot data. These NM-ROMs can be advantageous over linear-subspace ROMs (LS-ROMs) for problems with slowly decaying Kolmogorov n-width. However, the number of NM-ROM parameters that need to be trained scales with the size of the FOM. Moreover, for “extreme-scale” problems, the storage of high-dimensional FOM snapshots alone can make ROM training expensive. To alleviate the training cost, this paper applies DD to the FOM, computes NM-ROMs on each subdomain, and couples them to obtain a global NM-ROM. This approach has several advantages: Subdomain NM-ROMs can be trained in parallel, involve fewer parameters to be trained than global NM-ROMs, require smaller subdomain FOM dimensional training data, and can be tailored to subdomain specific features of the FOM. The shallow, sparse architecture of the autoencoder used in each subdomain NM-ROM allows application of hyper-reduction (HR), reducing the complexity caused by nonlinearity and yielding computational speedup of the NM-ROM. This paper provides the first application of NM-ROM (with HR) to a DD problem. In particular, this paper details an algebraic DD reformulation of the FOM, training a NM-ROM with HR for each sub domain, and a sequential quadratic programming (SQP) solver to evaluate the coupled global NM-ROM. Theoretical convergence results for the SQP method and a priori and a posteriori error estimates for the DD NM-ROM with HR are provided. The proposed DD NM-ROM with HR approach is numerically compared to a DD LS-ROM with HR on the 2D steady-state Burgers’ equation, showing an order of magnitude improvement in accuracy of the proposed DD NM-ROM over the DD LS-ROM.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2018

Roofline scaling trajectories: A method for parallel application and architectural performance analysis

The end of Dennard scaling signaled a shift in HPC supercomputer architectures from systems built from single-core processor architectures to systems built from multicore and eventually manycore architectures. This transition substantially complicated performance optimization and analysis as new programming models were created, new scaling methodologies deployed, and on-chip contention became a bottleneck to performance. Existing distributed memory performance models like logP and logGP were unable to capture this contention. The Roofline model was created to address this contention and its interplay with locality. However, to date, the Roofline model has focused on full-node concurrency. In this paper, we extend the Roofline model to capture the effects of concurrency on data locality and on-chip contention. We demonstrate the value of this new technique by evaluating the NAS parallel benchmarks on both multicore and manycore architectures under both strong-And weak-scaling regimes. In order to quantify the interplay between programming model and locality, we evaluate scaling under both the OpenMP and flat MPI programming models.

Ibrahim, K↗