Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “parallel, scaling”

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 1,369 records · Page 76

AutomaDeD: Automata-Based Debugging for Dissimilar Parallel Tasks

Today's largest systems have over 100,000 cores, with million-core systems expected over the next few years. This growing scale makes debugging the applications that run on them a daunting challenge. Few debugging tools perform well at this scale and most provide an overload of information about the entire job. Developers need tools that quickly direct them to the root cause of the problem. This paper presents AutomaDeD, a tool that identifies which tasks of a large-scale application first manifest a bug at a specific code region at a specific point during program execution. AutomaDeD creates a statistical model of the application's control-flow and timing behavior that organizes tasks into groups and identifies deviations from normal execution, thus significantly reducing debugging effort. In addition to a case study in which AutomaDeD locates a bug that occurred during development of MVAPICH, we evaluate AutomaDeD on a range of bugs injected into the NAS parallel benchmarks. Our results demonstrate that detects the time period when a bug first manifested itself with 90% accuracy for stalls and hangs and 70% accuracy for interference faults. It identifies the subset of processes first affected by the fault with 80% accuracy and 70% accuracy, respectively and the code region where where the fault first manifested with 90% and 50% accuracy, respectively.

97 MATHEMATICS AND COMPUTING↗

The GEWEX LandFlux project: Evaluation of model evaporation using tower-based and globally gridded forcing data

Determining the spatial distribution and temporal development of evaporation at regional and global scales is required to improve our understanding of the coupled water and energy cycles and to better monitor any changes in observed trends and variability of linked hydrological processes. With recent international efforts guiding the development of long-term and globally distributed flux estimates, continued product assessments are required to inform upon the selection of suitable model structures and also to establish the appropriateness of these multi-model simulations for global application. In support of the objectives of the Global Energy and Water Cycle Exchanges (GEWEX) LandFlux project, four commonly used evaporation models are evaluated against data from tower-based eddy-covariance observations, distributed across a range of biomes and climate zones. The selected schemes include the Surface Energy Balance System (SEBS) approach, the Priestley–Taylor Jet Propulsion Laboratory (PT-JPL) model, the Penman–Monteith-based Mu model (PM-Mu) and the Global Land Evaporation Amsterdam Model (GLEAM). Here we seek to examine the fidelity of global evaporation simulations by examining the multi-model response to varying sources of forcing data. To do this, we perform parallel and collocated model simulations using tower-based data together with a global-scale grid-based forcing product. Through quantifying the multi-model response to high-quality tower data, a better understanding of the subsequent model response to the coarse-scale globally gridded data that underlies the LandFlux product can be obtained, while also providing a relative evaluation and assessment of model performance. Using surface flux observations from 45 globally distributed eddy-covariance stations as independent metrics of performance, the tower-based analysis indicated that PT-JPL provided the highest overall statistical performance (0.72; 61 W m –2 ; 0.65), followed closely by GLEAM (0.68; 64 W m –2 ; 0.62), with values in parentheses representing the R 2 , RMSD and Nash–Sutcliffe efficiency (NSE), respectively. PM-Mu (0.51; 78 W m –2 ; 0.45) tended to underestimate fluxes, while SEBS (0.72; 101 W m –2 ; 0.24) overestimated values relative to observations. A focused analysis across specific biome types and climate zones showed considerable variability in the performance of all models, with no single model consistently able to outperform any other. Results also indicated that the global gridded data tended to reduce the performance for all of the studied models when compared to the tower data, likely a response to scale mismatch and issues related to forcing quality. Rather than relying on any single model simulation, the spatial and temporal variability at both the tower- and grid-scale highlighted the potential benefits of developing an ensemble or blended evaporation product for global-scale LandFlux applications. Hence, challenges related to the robust assessment of the LandFlux product are also discussed.

58 GEOSCIENCES↗

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↗

The onset of ion heating during magnetic reconnection with a strong guide field

The onset of the acceleration of ions during magnetic reconnection is explored via particle-in-cell simulations in the limit of a strong ambient guide field that self-consistently and simultaneously follow the motions of protons and α particles. Heating parallel to the local magnetic field during reconnection with a guide field is strongly reduced compared with the reconnection of anti-parallel magnetic fields. The dominant heating of thermal ions during guide field reconnection results from pickup behavior of ions during their entry into reconnection exhausts and dominantly produces heating perpendicular rather than parallel to the local magnetic field. Pickup behavior requires that the ion transit time across the exhaust boundary (with a transverse scale of the order of the ion sound Larmor radius) be short compared with the ion cyclotron period. This translates into a threshold in the strength of reconnecting magnetic field that favors the heating of ions with high mass-to-charge. A simulation with a broad initial current layer produces a reconnecting system in which the amplitude of the reconnecting magnetic field just upstream of the dissipation region increases with time as reconnection proceeds. The sharp onset of perpendicular heating when the pickup threshold is crossed is documented. A comparison of the time variation of the parallel and perpendicular ion heating with that predicted based on the strength of the reconnecting field establishes the scaling of ion heating with ambient parameters both below and above the pickup threshold. The relevance to observations of ion heating in the solar corona is discussed.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Harnessing billions of tasks for a scalable portable hydrodynamic simulation of the merger of two stars

We propose a highly scalable demonstration of a portable asynchronous many-task programming model and runtime system applied to a grid-based adaptive mesh refinement hydrodynamic simulation of a double white dwarf merger with 14 levels of refinement that spans 17 orders of magnitude in astrophysical densities. The code uses the portable C++ parallel programming model that is embodied in the HPX library and being incorporated into the ISO C++ standard. The model reflects a significant shift from existing bulk synchronous parallel programming models under consideration for exascale systems. Through the use of the Futurization technique, seemingly sequential code is transformed into wait-free asynchronous tasks. We demonstrate the potential of our model by showing results from strong scaling runs on National Energy Research Scientific Computing Center’s Cori system (658,784 Intel Knight’s Landing cores) that achieve a parallel efficiency of 96.8% using billions of asynchronous tasks.

79 ASTRONOMY AND ASTROPHYSICS↗

Sage: parallel semi-asymmtric graph algorithms for NVRAMs

Non-volatile main memory (NVRAM) technologies provide an attractive set of features for large-scale graph analytics, including byte-addressability, low idle power, and improved memory-density. NVRAM systems today have an order of magnitude more NVRAM than traditional memory (DRAM). NVRAM systems could therefore potentially allow very large graph problems to be solved on a single machine, at a modest cost. However, a significant challenge in achieving high performance is in accounting for the fact that NVRAM writes can be much more expensive than NVRAM reads. In this paper, we propose an approach to parallel graph analytics using the Parallel Semi-Asymmetric Model (PSAM), in which the graph is stored as a read-only data structure (in NVRAM), and the amount of mutable memory is kept proportional to the number of vertices. Similar to the popular semi-external and semi-streaming models for graph analytics, the PSAM approach assumes that the vertices of the graph fit in a fast read-write memory (DRAM), but the edges do not. In NVRAM systems, our approach eliminates writes to the NVRAM, among other benefits. To experimentally study this new setting, we develop Sage, a parallel semi-asymmetric graph engine with which we implement provably-efficient (and often work-optimal) PSAM algorithms for over a dozen fundamental graph problems. We experimentally study Sage using a 48--core machine on the largest publicly-available real-world graph (the Hyperlink Web graph with over 3.5 billion vertices and 128 billion edges) equipped with Optane DC Persistent Memory, and show that Sage outperforms the fastest prior systems designed for NVRAM. Importantly, we also show that Sage nearly matches the fastest prior systems running solely in DRAM, by effectively hiding the costs of repeatedly accessing NVRAM versus DRAM.

97 MATHEMATICS AND COMPUTING↗

Enabling machine learning-ready HPC ensembles with Merlin

With the growing complexity of computational and experimental facilities, many scientific researchers are turning to machine learning (ML) techniques to analyze large scale ensemble data. With complexities such as multi-component workflows, heterogeneous machine architectures, parallel file systems, and batch scheduling, care must be taken to facilitate this analysis in a high performance computing (HPC) environment. Here, we present Merlin, a workflow framework to enable large ML-friendly ensembles of scientific HPC simulations. By augmenting traditional HPC with distributed compute technologies, Merlin aims to lower the barrier for scientific subject matter experts to incorporate ML into their analysis. As a producer–consumer workflow model, Merlin enables multi-machine, cross-batch job, dynamically allocated yet persistent workflows capable of utilizing surge-compute resources. Key features of Merlin are a flexible HPC-centric interface, low per-task overhead, multi-tiered fault recovery, and a hierarchical sampling algorithm that allows for $\mathscr{O}$(N) task execution and $\mathscr{O}$(N ln N) task queuing to ensembles of millions of tasks. In addition to Merlin’s design, we test the algorithm’s performance in an HPC center and demonstrate the ability to enqueue 40 million simulations in 100 s, with a 30 millisecond per-task overhead that is independent of ensemble size. Finally, we describe some example applications that Merlin has enabled on leadership-class HPC resources, such as the ML-augmented optimization of nuclear fusion experiments and the calibration of infectious disease models to study the progression of and possible mitigation strategies for COVID-19.

97 MATHEMATICS AND COMPUTING↗

Ab Initio Reactive Computer Aided Molecular Design

Few would dispute that theoretical chemistry tools can now provide keen insights into chemical phenomena. Yet the holy grail of efficient and reliable prediction of complex reactivity has remained elusive. Fortunately, recent advances in electronic structure theory based on the concepts of both element- and rank-sparsity, coupled with the emergence of new highly parallel computer architectures, have led to a significant increase in the time and length scales which can be simulated using first principles molecular dynamics. This then opens the possibility of new discovery-based approaches to chemical reactivity, such as the recently proposed ab initio nanoreactor. Here, we argue that due to these and other recent advances, the holy grail of computational discovery for complex chemical reactivity is rapidly coming within our reach.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Parallelizable microfluidic dropmakers with multilayer geometry for the generation of double emulsions

Microfluidic devices enable the production of uniform double emulsions with control over droplet size and shell thickness. However, the limited production rate of microfluidic devices precludes the use of monodisperse double emulsions for industrial-scale applications, which require large quantities of droplets. To increase throughput, devices can be parallelized to contain many dropmakers operating simultaneously in one chip, but this is challenging to do for double emulsion dropmakers. Production of double emulsions requires dropmakers to have both hydrophobic and hydrophilic channels, requiring spatially precise patterning of channel surface wettability. Precise wettability patterning is difficult for devices containing multiple dropmakers, posing a significant challenge for parallelization. In this paper, we present a multilayer dropmaker geometry that greatly simplifies the process of producing microfluidic devices with excellent spatial control over channel wettability. Wettability patterning is achieved through the independent functionalization of channels in each layer prior to device assembly, rendering the dropmaker with a precise step between hydrophobic and hydrophilic channels. We report this device geometry enables uniform wettability patterning of parallelized dropmakers, providing a scalable approach for the production of double emulsions.

36 MATERIALS SCIENCE↗

Optically tunable mesoscale CdSe morphologies via inorganic phototropic growth

Inorganic phototropic growth using only spatially conformal illumination generated Se–Cd films that exhibited precise light-defined mesoscale morphologies including highly ordered, anisotropic, and periodic ridge and trench nanotextures over entire macroscopic substrates. Growth was accomplished via a light-induced electrochemical method using an optically and chemically isotropic solution, an unpatterned substrate, and unstructured, incoherent, low-intensity illumination in the absence of chemical directing agents or physical templates and masks. The morphologies were defined by the illumination inputs: the nanotexture long axes aligned parallel to the optical E-field vector, and the feature sizes and periods scaled with the wavelength. Optically based modeling of the growth closely reproduced the experimental results, confirming the film morphologies were fully determined by the light–matter interactions during growth. Solution processing of the Se–Cd films resulted in stoichiometric, crystalline CdSe films that also exhibited ordered nanotextures, demonstrating that inorganic phototropic growth can effect tunable, template-free generation of ordered CdSe nanostructures over macroscopic length scales.

36 MATERIALS SCIENCE↗

Simulation of dilute polymeric fluids in a three-dimensional contraction using a multiscale FENE model

The multiscale FENE model is applied to a 3D square-square contraction flow problem. For this purpose, the stochastic Brownian configuration field method (BCF) has been coupled with our fully parallelized three-dimensional Navier-Stokes solver NaSt3DGPF. The robustness of the BCF method enables the numerical simulation of high Deborah number flows for which most macroscopic methods suffer from stability issues. The results of our simulations are compared with that of experimental measurements from literature and show a very good agreement. In particular, flow phenomena such as a strong vortex enhancement, streamline divergence and a flow inversion for highly elastic flows are reproduced. Due to their computational complexity, our simulations require massively parallel computations. Using a domain decomposition approach with MPI, the implementation achieves excellent scale-up results for up to 128 processors.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Mode-space energy distribution in instability-driven plasma turbulence

Energy transfer to damped modes in gyrokinetic ion temperature gradient driven turbulence is studied to understand the transfer dynamics and find scaling representations for the heavily populated mode space. Proper orthogonal and linear eigenmode decompositions are introduced and examined to assess whether modes are well-resolved and what scales they encompass. It is observed that damped modes across a range of inhomogeneous scales receive energy simultaneously and directly from the unstable mode, constituting a form of parallel transfer, distinct from the serial mode-to-mode transfer of the wavenumber cascade of hydrodynamic turbulence. Controlling for modes that are well resolved and labeling the modes of the linear decomposition in order of damping rate, energy transfer in the mode space satisfies an equipartition of the energy dissipation rate, leading to a simple rule for the distribution of energy in the space of damped modes. Energy dissipation rate equipartition is the form that the canonical nonlinear invariance of energy transfer assumes in a dissipation range with parallel rather than serial transfer.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Proposed linear energy transfer areal detector for protons using radiochromic film

Radiation therapy depends on predictably and reliably delivering dose to tumors and sparing normal tissues. Protons with kinetic energy of a few hundred MeV can selectively deposit dose to deep seated tumors without an exit dose, unlike x-rays. The better dose distribution is attributed to a phenomenon known as the Bragg peak. The Bragg peak is due to relatively high energy deposition within a given distance or high Linear Energy Transfer (LET). In addition, biological response to radiation depends on the dose, dose rate, and localized energy deposition patterns or LET. At present, the LET can only be measured at a given fixed point and the LET spatial distribution can only be inferred from calculations. The goal of this study is to develop and test a method to measure LET over extended areas. Traditionally, radiochromic films are used to measure dose distribution but not for LET distribution. We report the first use of these films for measuring the spatial distribution of the LET deposited by protons. The radiochromic film sensitivity diminishes for large LET. A mathematical model correlating the film sensitivity and LET is presented to justify relating LET and radiochromic film relative sensitivity. Protons were directed parallel to radiochromic film sandwiched between solid water slabs. This study proposes the scaled-normalized difference (SND) between the Treatment Planning system (TPS) and measured dose as the metric describing the LET. The SND is correlated with a Monte Carlo (MC) calculation of the LET spatial distribution for a large range of SNDs. A polynomial fit between the SND and MC LET is generated for protons having a single range of 20 cm with narrow Bragg peak. Coefficients from these fitted polynomial fits were applied to measured proton dose distributions with a variety of ranges. An identical procedure was applied to the protons deposited from Spread Out Bragg Peak and modulated by 5 cm. Gamma analysis is a method for comparing the calculated LET with the LET measured using radiochromic film at the pixel level over extended areas. Failure rates using gamma analysis are calculated for areas in the dose distribution using parameters of 25% of MC LET and 3 mm. The processed dose distributions find 5%–10% failure rates for the narrow 12.5 and 15 cm proton ranges and 10%–15% for proton ranges of 15, 17.5, and 20 cm and modulated by 5 cm. It is found through gamma analysis that the measured proton energy deposition in radiochromic film and TPS can be used to determine LET. This modified film dosimetry provides an experimental areal LET measurement that can verify MC calculations, support LET point measurements, possibly enhance biologically based proton treatment planning, and determine the polymerization process within the radiochromic film.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

The reduction of magnetic reconnection outflow jets to sub-Alfvénic speeds

The outflow velocity of jets produced by collisionless magnetic reconnection is shown to be reduced by the ion exhaust temperature in fully kinetic particle in cell simulations and in situ satellite observations. We derive a scaling relationship for the outflow velocity based on the upstream Alfvén speed and the parallel ion exhaust temperature, which is verified in kinetic simulations and observations. The outflow speed reduction is shown to be due to the firehose instability criterion, and so, for large enough guide fields, this effect is suppressed and the outflow speed reaches the upstream Alfvén speed based on the reconnecting component of the magnetic field.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

PARTICLE ACCELERATION AND WAVE EXCITATION IN QUASI-PARALLEL HIGH-MACH-NUMBER COLLISIONLESS SHOCKS: PARTICLE-IN-CELL SIMULATION

We herein investigate shock formation and particle acceleration processes for both protons and electrons in a quasi-parallel high-Mach-number collisionless shock through a long-term, large-scale, particle-in-cell simulation. We show that both protons and electrons are accelerated in the shock and that these accelerated particles generate large-amplitude Alfvénic waves in the upstream region of the shock. After the upstream waves have grown sufficiently, the local structure of the collisionless shock becomes substantially similar to that of a quasi-perpendicular shock due to the large transverse magnetic field of the waves. A fraction of protons are accelerated in the shock with a power-law-like energy distribution. The rate of proton injection to the acceleration process is approximately constant, and in the injection process, the phase-trapping mechanism for the protons by the upstream waves can play an important role. The dominant acceleration process is a Fermi-like process through repeated shock crossings of the protons. This process is a “fast” process in the sense that the time required for most of the accelerated protons to complete one cycle of the acceleration process is much shorter than the diffusion time. A fraction of the electrons are also accelerated by the same mechanism, and have a power-law-like energy distribution. However, the injection does not enter a steady state during the simulation, which may be related to the intermittent activity of the upstream waves. Upstream of the shock, a fraction of the electrons are pre-accelerated before reaching the shock, which may contribute to steady electron injection at a later time.

79 ASTRONOMY AND ASTROPHYSICS↗

Controlled interactions of femtosecond light filaments in air

We experimentally demonstrate coherent control of two light filaments in air generated by ultrafast laser pulses, as proposed by Xi et al. [Phys. Rev. Lett. 96, 025003 (2006)]. We show that depending on the relative phase and the incidence angle, the filaments can experience fusion, repulsion, energy redistribution, and spiral motion for propagation in air. By translating the initial beams on subwavelength scales, we achieve {approx}1-mm transverse deflection of approximately 0.5-mJ energy for parallel propagation and 7 deg. angular deflection for spiral motion in 3-m propagation. Our approach using the relative beam phases to reliably control propagation of femtosecond light filaments is potentially applicable to remote sensing and lightning guiding.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Island model for parallel evolutionary optimization of spiking neuromorphic computing

Parallel genetic algorithms (PGAs) can be used to accelerate optimization by exploiting large-scale computational resources. In this work, we describe a PGA framework for evolving spiking neural networks (SNNs) for neuromorphic hardware implementation. The PGA framework is based on an islands model with migration. We show that using this framework, better SNNs for neuromorphic systems can be evolved faster.

Schuman, Catherine↗

Sirocco Storage Server v. pre-alpha 0.1

Sirocco is a parallel storage system under development, designed for write-intensive workloads on large-scale HPC platforms. It implements a keyvalue object store on top of a set of loosely federated storage servers that cooperate to ensure data integrity and performance. It includes support for a range of different types of storage transactions. This software release constitutes a conformant storage server, along with the client-side libraries to access the storage over a network.

Curry, MatthewL.↗