Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel clustering”

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 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2014

Melting of “non-magic” argon clusters and extrapolation to the bulk limit

The melting of argon clusters Ar{sub N} is investigated by applying a parallel-tempering Monte Carlo algorithm for all cluster sizes in the range from 55 to 309 atoms. Extrapolation to the bulk gives a melting temperature of 85.9 K in good agreement with the previous value of 88.9 K using only Mackay icosahedral clusters for the extrapolation [E. Pahl, F. Calvo, L. Koči, and P. Schwerdtfeger, “Accurate melting temperatures for neon and argon from ab initio Monte Carlo simulations,” Angew. Chem., Int. Ed. 47, 8207 (2008)]. Our results for argon demonstrate that for the extrapolation to the bulk one does not have to restrict to magic number cluster sizes in order to obtain good estimates for the bulk melting temperature. However, the extrapolation to the bulk remains a problem, especially for the systematic selection of suitable cluster sizes.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗
DOE OSTI2016

Coupled cluster Green function: Model involving single and double excitations

In this paper, we report on the development of a parallel implementation of the coupled-cluster (CC) Green function formulation (GFCC) employing single and double excitations in the cluster operator (GFCCSD). A key aspect of this work is the determination of the frequency dependent self-energy, Σ(ω). The detailed description of the underlying algorithm is provided, including approximations used that preserve the pole structure of the full GFCCSD method, thereby reducing the computational costs while maintaining an accurate character of methodology. Furthermore, for systems with strong local correlation, our formulation reveals a diagonally dominate block structure where as the non-local correlation increases, the block size increases proportionally. Here, to demonstrate the accuracy of our approach, several examples including calculations of ionization potentials for benchmark systems are presented and compared against experiment.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2013

PARALLEL ELECTRIC FIELD SPECTRUM OF SOLAR WIND TURBULENCE

By searching through more than 10 satellite years of THEMIS and Cluster data, 3 reliable examples of parallel electric field turbulence in the undisturbed solar wind have been found. The perpendicular and parallel electric field spectra in these examples have similar shapes and amplitudes, even at large scales (frequencies below the ion gyroscale), where Alfvenic turbulence with no parallel electric field component is thought to dominate. The spectra of the parallel electric field fluctuations are power laws with exponents near -5/3 below the ion scales ({approx}0.1 Hz), and with a flattening of the spectrum in the vicinity of this frequency. At small scales (above a few Hz), the spectra are steeper than -5/3 with values in the range of -2.1 to -2.8. These steeper slopes are consistent with expectations for kinetic Alfven turbulence, although their amplitude relative to the perpendicular fluctuations is larger than expected.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2020

FPDeep: Scalable Acceleration of CNN Training on Deeply-Pipelined FPGA Clusters

In this paper, we propose a framework called FPDeep, which uses a hybrid of model and layer paral- lelism to configure distributed reconfigurable clusters to train DNNs. This approach has numerous benefits. First, the design does not suffer from batch size growth. Second, novel workload and weight partitioning leads to balanced loads of both among nodes. And third, the entire system is fine-grained pipeline. This leads to high parallelism and utilization and also minimizes the time features need to be cached while waiting for back-propagation.

Wang, Tianqi↗
DOE OSTI2022

aphBO-2GP-3B: a budgeted asynchronous parallel multi-acquisition functions for constrained Bayesian optimization on high-performing computing architecture

High-fidelity complex engineering simulations are often predictive, but also computationally expensive and often require substantial computational efforts. The mitigation of computational burden is usually enabled through parallelism in high-performance cluster (HPC) architecture. Optimization problems associated with these applications is a challenging problem due to the high computational cost of the high-fidelity simulations. In this paper, an asynchronous parallel constrained Bayesian optimization method is proposed to efficiently solve the computationally expensive simulation-based optimization problems on the HPC platform, with a budgeted computational resource, where the maximum number of simulations is a constant. The advantage of this method are three-fold. Firstly, the efficiency of the Bayesian optimization is improved, where multiple input locations are evaluated parallel in an asynchronous manner to accelerate the optimization convergence with respect to physical runtime. This efficiency feature is further improved so that when each of the inputs is finished, another input is queried without waiting for the whole batch to complete. Second, the proposed method can handle both known and unknown constraints. Third, the proposed method samples several acquisition functions based on their rewards using a modified GP-Hedge scheme. The proposed framework is termed aphBO-2GP-3B, which means asynchronous parallel hedge Bayesian optimization with two Gaussian processes and three batches. The numerical performance of the proposed framework aphBO-2GP-3B is comprehensively benchmarked using 16 numerical examples, compared against other 6 parallel Bayesian optimization variants and 1 parallel Monte Carlo as a baseline, and demonstrated using two real-world high-fidelity expensive industrial applications. The first engineering application is based on finite element analysis (FEA) and the second one is based on computational fluid dynamics (CFD) simulations.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2013

Hierarchical Petascale Simulation Framework For Stress Corrosion Cracking

A number of major accomplishments resulted from the project. These include: • Data Structures, Algorithms, and Numerical Methods for Reactive Molecular Dynamics. We have developed a range of novel data structures, algorithms, and solvers (amortized ILU, Spike) for use with ReaxFF and charge equilibration. • Parallel Formulations of ReactiveMD (Purdue ReactiveMolecular Dynamics Package, PuReMD, PuReMD-GPU, and PG-PuReMD) for Messaging, GPU, and GPU Cluster Platforms. We have developed efficient serial, parallel (MPI), GPU (Cuda), and GPU Cluster (MPI/Cuda) implementations. Our implementations have been demonstrated to be significantly better than the state of the art, both in terms of performance and scalability. • Comprehensive Validation in the Context of Diverse Applications. We have demonstrated the use of our software in diverse systems, including silica-water, silicon-germanium nanorods, and as part of other projects, extended it to applications ranging from explosives (RDX) to lipid bilayers (biomembranes under oxidative stress). • Open Source Software Packages for Reactive Molecular Dynamics. All versions of our soft- ware have been released over the public domain. There are over 100 major research groups worldwide using our software. • Implementation into the Department of Energy LAMMPS Software Package. We have also integrated our software into the Department of Energy LAMMPS software package.

36 MATERIALS SCIENCE↗
DOE OSTI2016

IMPACC: A Tightly Integrated MPI+OpenACC Framework Exploiting Shared Memory Parallelism

We propose IMPACC, an MPI+OpenACC framework for heterogeneous accelerator clusters. IMPACC tightly integrates MPI and OpenACC, while exploiting the shared memory parallelism in the target system. IMPACC dynamically adapts the input MPI+OpenACC applications on the target heterogeneous accelerator clusters to fully exploit target system-specific features. IMPACC provides the programmers with the unified virtual address space, automatic NUMA-friendly task-device mapping, efficient integrated communication routines, seamless streamlining of asynchronous executions, and transparent memory sharing. We have implemented IMPACC and evaluated its performance using three heterogeneous accelerator systems, including Titan supercomputer. Results show that IMPACC can achieve easier programming, higher performance, and better scalability than the current MPI+OpenACC model.

Kim, Jungwon↗
DOE OSTI2017

Leószilárdite, the first Na,Mg-containing uranyl carbonate from the Markey Mine, San Juan County, Utah, USA

Leószilárdite (IMA2015-128), Na 6 Mg(UO 2 ) 2 (CO 3 ) 6 ·6H 2 O, was found in the Markey Mine, Red Canyon, White Canyon District, San Juan County, Utah, USA, in areas with abundant andersonite, natrozippeite, gypsum, anhydrite, and probable hydromagnesite along with other secondary uranium minerals bayleyite, čejkaite and johannite. The new mineral occurs as aggregates of pale yellow bladed crystals flattened on {001} and elongated along [010], individually reaching up to 0.2 mm long. More commonly it occurs as pale yellow pearlescent masses to 2 mm consisting of very small plates. Leószilárdite fluoresces green under both longwave and shortwave ultraviolet light, and is translucent with a white streak, hardness of 2 (Mohs), and brittle tenacity with uneven fracture. The new mineral is readily soluble in room temperature H 2 O. Crystals have perfect cleavage along {001}, and exhibit the forms {110},{001},{100},{101} and {101}. Optically, leószilárdite is biaxial (-), α= 1.504(1), β= 1.597(1), γ = 1.628(1) (white light); 2V (meas.) = 57(1)°, 2V (calc.) = 57.1°; dispersion r > v , slight. Pleochroism: X = colourless, Y and Z = light yellow; X < Y ≈ Z . The average of six wavelength dispersive spectroscopic analyses provided Na 2 O 14.54, MgO 3.05, UO 3 47.95, CO 2 22.13, H 2 O 9.51, total 97.18 wt.%. The empirical formula is Na 5.60 Mg 0.90 U 2 O 28 C 6 H 12.60 , based on 28 O apfu. Leószilárdite is monoclinic, C2/m, a = 11.6093(21), b = 6.7843(13), c = 15.1058(28) Å, β = 91.378(3)°, V = 1189.4(4) Å 3 and Z = 2. The crystal structure ( R 1 = 0.0387 for 1394 reflections with I obs > 4σ I ), consists of uranyl tricarbonate anion clusters [(UO 2 )(CO 3 ) 3 ] 4- held together in part by irregular chains of NaO 5 (H 2 O) polyhedra sub parallel to [010]. Individual uranyl tricarbonate clusters are also linked together by three-octahedron units consisting of two Na-centred octahedra that share the opposite faces of a Mg-centred octahedron at the centre (Na–Mg–Na), and have the composition Na 2 MgO 12 (H 2 O) 4 . The name of the new mineral honours the Hungarian-American physicist, inventor and biologist Dr. Leó Szilárd (1898–1964).

Mineralogy↗
DOE OSTI2019

Managing genomic variant calling workflows with Swift/T

Bioinformatics research is frequently performed using complex workflows with multiple steps, fans, merges, and conditionals. This complexity makes management of the workflow difficult on a computer cluster, especially when running in parallel on large batches of data: hundreds or thousands of samples at a time. Scientific workflow management systems could help with that. Many are now being proposed, but is there yet the “best” workflow management system for bioinformatics? Such a system would need to satisfy numerous, sometimes conflicting requirements: from ease of use, to seamless deployment at peta- and exa-scale, and portability to the cloud. We evaluated Swift/T as a candidate for such role by implementing a primary genomic variant calling workflow in the Swift/T language, focusing on workflow management, performance and scalability issues that arise from production-grade big data genomic analyses. In the process we introduced novel features into the language, which are now part of its open repository. Additionally, we formalized a set of design criteria for quality, robust, maintainable workflows that must function at-scale in a production setting, such as a large genomic sequencing facility or a major hospital system. The use of Swift/T conveys two key advantages. (1) It operates transparently in multiple cluster scheduling environments (PBS Torque, SLURM, Cray aprun environment, etc.), thus a single workflow is trivially portable across numerous clusters. (2) The leaf functions of Swift/T permit developers to easily swap executables in and out of the workflow, which makes it easy to maintain and to request resources optimal for each stage of the pipeline. While Swift/T’s data-level parallelism eliminates the need to code parallel analysis of multiple samples, it does make debugging more difficult, as is common for implicitly parallel code. Nonetheless, the language gives users a powerful and portable way to scale up analyses in many computing architectures.

Science & Technology - Other Topics↗
DOE OSTI2017

The Frontier Fields: Survey Design and Initial Results

What are the faintest distant galaxies we can see with the Hubble Space Telescope ( HST ) now, before the launch of the James Webb Space Telescope ? This is the challenge taken up by the Frontier Fields, a Director’s discretionary time campaign with HST and the Spitzer Space Telescope to see deeper into the universe than ever before. The Frontier Fields combines the power of HST and Spitzer with the natural gravitational telescopes of massive high-magnification clusters of galaxies to produce the deepest observations of clusters and their lensed galaxies ever obtained. Six clusters—Abell 2744, MACSJ0416.1-2403, MACSJ0717.5+3745, MACSJ1149.5+2223, Abell S1063, and Abell 370—have been targeted by the HST ACS/WFC and WFC3/IR cameras with coordinated parallel fields for over 840 HST orbits. The parallel fields are the second-deepest observations thus far by HST with 5 σ point-source depths of ∼29th ABmag. Galaxies behind the clusters experience typical magnification factors of a few, with small regions magnified by factors of 10–100. Therefore, the Frontier Field cluster HST images achieve intrinsic depths of ∼30–33 mag over very small volumes. Spitzer has obtained over 1000 hr of Director’s discretionary imaging of the Frontier Field cluster and parallels in IRAC 3.6 and 4.5 μ m bands to 5 σ point-source depths of ∼26.5, 26.0 ABmag. We demonstrate the exceptional sensitivity of the HST Frontier Field images to faint high-redshift galaxies, and review the initial results related to the primary science goals.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2023

Porting Fragmentation Methods to Graphical Processing Units Using an OpenMP Application Programming Interface: Offloading the Fock Build for Low Angular Momentum Functions

Here, a framework to offload four-index two-electron repulsion integrals to graphical processing units (GPUs) using OpenMP is discussed. The method has been applied to the Fock build for low angular momentum s and p functions in both the restricted Hartree–Fock (RHF) and in the effective fragment molecular orbital (EFMO) framework. Benchmark calculations for the GPU code for the pure RHF method show an increasing speedup relative to the existing OpenMP CPU code in GAMESS from 1.04 to 52× for clusters of 70–569 water molecules. The parallel efficiency on 24 NVIDIA V100 GPU boards also increases when increasing the system size: from 75 to 94% for water clusters that contain 303–1120 molecules. In the EFMO framework, the GPU Fock build shows a high linear scalability up to 4608 V100s with a parallel efficiency of 96% for calculations on a solvated mesoporous silica nanoparticle system with ~67,000 basis functions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2015

Implementation of Parallel Dynamic Simulation on Shared-Memory vs. Distributed-Memory Environments

Power system dynamic simulation computes the system response to a sequence of large disturbance, such as sudden changes in generation or load, or a network short circuit followed by protective branch switching operation. It consists of a large set of differential and algebraic equations, which is computational intensive and challenging to solve using single-processor based dynamic simulation solution. High-performance computing (HPC) based parallel computing is a very promising technology to speed up the computation and facilitate the simulation process. This paper presents two different parallel implementations of power grid dynamic simulation using Open Multi-processing (OpenMP) on shared-memory platform, and Message Passing Interface (MPI) on distributed-memory clusters, respectively. The difference of the parallel simulation algorithms and architectures of the two HPC technologies are illustrated, and their performances for running parallel dynamic simulation are compared and demonstrated.

dynamic simulation↗
DOE OSTI2018

Tomographic analysis of tangential viewing cameras (invited)

Many tokamaks now use visible light cameras to observe plasma-wall interactions and integrated line emission. The DIII-D coherence imaging spectroscopy diagnostic cameras image interferograms that encode line integrated velocity. By modeling the 2D camera image pixels as line of sight integrals through an axisymmetric discrete grid, it is possible to do tomographic analysis to determine the local plasma line emissivity and parallel velocity. Approaches to solve the inverse problem posed by these tangential viewing cameras are presented. The inversion begins with calculation of the sparse response matrix that encompasses all the geometry and diagnostic information and reduces the process of image formation to a sparse matrix-vector multiply. Our report includes techniques for determining the detailed geometry of the camera views and methods for handling physical quantities that vary spatially. Additionally, the size of the response matrix has driven the development of capability to distribute the coarse parallel calculation across a heterogeneous cluster of computers on the Energy Sciences Network. Iterative techniques are then used to solve the sparse matrix-vector linear system

47 OTHER INSTRUMENTATION↗
DOE OSTI2021

Explicit structure-preserving geometric particle-in-cell algorithm in curvilinear orthogonal coordinate systems and its applications to whole-device 6D kinetic simulations of tokamak physics

Explicit structure-preserving geometric particle-in-cell (PIC) algorithm in curvilinear orthogonal coordinate systems is developed. The work reported represents a further development of the structure-preserving geometric PIC algorithm achieving the goal of practical applications in magnetic fusion research. The algorithm is constructed by discretizing the field theory for the system of charged particles and electromagnetic field using Whitney forms, discrete exterior calculus, and explicit non-canonical symplectic integration. In addition to the truncated infinitely dimensional symplectic structure, the algorithm preserves exactly many important physical symmetries and conservation laws, such as local energy conservation, gauge symmetry and the corresponding local charge conservation. As a result, the algorithm possesses the long-term accuracy and fidelity required for first-principles-based simulations of the multiscale tokamak physics. The algorithm has been implemented in the SymPIC code, which is designed for high-efficiency massively-parallel PIC simulations in modern clusters. The code has been applied to carry out whole-device 6D kinetic simulation studies of tokamak physics. A self-consistent kinetic steady state for fusion plasma in the tokamak geometry is numerically found with a predominately diagonal and anisotropic pressure tensor. The state also admits a steady-state sub-sonic ion flow in the range of 10 km s -1 , agreeing with experimental observations and analytical calculations Kinetic ballooning instability in the self-consistent kinetic steady state is simulated. It is shown that high-n ballooning modes have larger growth rates than low-n global modes, and in the nonlinear phase the modes saturate approximately in 5 ion transit times at the 2% level by the E × B flow generated by the instability. These results are consistent with early and recent electromagnetic gyrokinetic simulations.

43 PARTICLE ACCELERATORS↗
DOE OSTI2025

RWRtoolkit: multi-omic network analysis using random walks on multiplex networks in any species

Abstract We introduce RWRtoolkit, a multiplex generation, exploration, and statistical package built for R and command-line users. RWRtoolkit enables the efficient exploration of large and highly complex biological networks generated from custom experimental data and/or from publicly available datasets, and is species agnostic. A range of functions can be used to find topological distances between biological entities, determine relationships within sets of interest, search for topological context around sets of interest, and statistically evaluate the strength of relationships within and between sets. The command-line interface is designed for parallelization on high-performance cluster systems, which enables high-throughput analysis such as permutation testing. Several tools in the package have also been made available for use in reproducible workflows via the KBase web application.

Kainer, David (ORCID:0000000172714676)↗
DOE OSTI2024

PPO And Friends

PPO and Friends (PPOAF) is a pytorch implementation of proximal policy optimization for single- and multi-agent reinforcement learning (the PPO), along with several optimizations and add-ons (the Friends) to enable efficient MPI-parallelized model training on HPC clusters.

Maguire, AlisterO↗