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 127 records · Page 7

Self-consistent formation of parallel electric fields in the auroral zone

This paper presents results from a fully self-consistent kinetic particle simulation of the time-dependent formation of large scale parallel electric fields in the auroral zone. The results show that magnetic mirroring of the hot plasma that streams earthward from the magnetotail leads to a charge separation potential drop of many kilovolts, over an altitude range of a few thousand kilometers. Once the potential drop is formed, it remains relatively static and is maintained in time by the constant input of hot plasma from the tail; the parallel electric field accelerates ions away from Earth and ionospheric electrons towards the Earth. At altitudes above where the ions are mirror reflected and accelerated by the parallel electric field, low frequency waves are generated, possibly due to an ion/ion two-stream interaction.

Schriver, David↗

Domain decomposition in the GPU-accelerated Shift Monte Carlo code

The GPU solver within the Shift continuous-energy Monte Carlo neutron transport code has been extended to provide domain decomposition in addition to domain replication to enable the solution of problems with memory requirements exceeding the capacity of a single GPU. The strategy follows the Multiple Set, Overlapping Domain (MSOD) approach that is used in Shift’s CPU solver and integrates into the event-based algorithm used for Shift’s GPU solver. Furthermore, the ability to assign processors to spatial domains non-uniformly has been maintained. In this work, two different approaches for communicating particle data between domains are considered, and multiple criteria for load balancing problems have been investigated. Numerical results are presented for both fresh and depleted small modular nuclear reactor (SMR) cores. A parallel efficiency of approximately 80% was achieved with up to 16 spatial domains measured relative to full domain replication. A scaling study on the Summit supercomputer demonstrates a weak scaling parallel efficiency of over 90% on over 24000 GPUs.

97 MATHEMATICS AND COMPUTING↗

Program Correctness, Verification and Testing for Exascale (Corvette)

The goal of this project is to provide tools to assess the correctness of parallel programs written using hybrid parallelism. There is a dire lack of both theoretical and engineering know-how in the area of finding bugs in hybrid or large scale parallel programs, which our research aims to change. In the project we have demonstrated novel approaches in several areas: 1. Low overhead automated and precise detection of concurrency bugs at scale. 2. Using low overhead bug detection tools to guide speculative program transformations for performance. 3. Techniques to reduce the concurrency required to reproduce a bug using partial program restart/replay. 4. Techniques to provide reproducible execution of floating point programs. 5. Techniques for tuning the floating point precision used in codes.

97 MATHEMATICS AND COMPUTING↗

Dynamic Smagorinsky Modeled Large-Eddy Simulations of Turbulence Using Tetrahedral Meshes

Eddy-resolving numerical computations of turbulent flows are emerging as viable alternatives to Reynolds Averaged Navier-Stokes (RANS) calculations for flows with an intrinsically steady mean state due to the advances in large-scale parallel computing. In these computations, medium to large turbulent eddies are resolved by the numerics while the smaller or subgrid scales are either modeled or taken care of by the inherent numerical dissipation. To advance the state of the art of unstructured-mesh turbulence simulation capabilities, large eddy simulations (LES) using the dynamic Smagorinsky model (DSM) on tetrahedral meshes are carried out with the space-time conservation element, solution element (CESE) method. In contrast to what has been reported in the literature, the present implementation of dynamic models allows for active backscattering without any ad-hoc limiting of the eddy viscosity calculated from the subgrid-scale model. For the benchmark problems involving compressible isotropic turbulence decay as well as the shock/turbulent boundary layer interaction benchmark problems, no numerical instability associated with kinetic energy growth is observed and the volume percentage of the backscattering portion accounts for about 38-40% of the simulation domain. A slip-wall model in conjunction with the implemented DSM is used to simulate a relatively high Reynolds number Mach 2.85 turbulent boundary layer over a 30° ramp with several tetrahedral meshes and a wall-normal spacing of either Δγ& = 10 or Δγ& = 20. The computed mean wall pressure distribution, separation region size, mean velocity profiles, and Reynolds stress agree reasonably well with experimental data.

CESE method↗

Programming Probabilistic Structural Analysis for Parallel Processing Computer

The ultimate goal of this research program is to make Probabilistic Structural Analysis (PSA) computationally efficient and hence practical for the design environment by achieving large scale parallelism. The paper identifies the multiple levels of parallelism in PSA, identifies methodologies for exploiting this parallelism, describes the development of a parallel stochastic finite element code, and presents results of two example applications. It is demonstrated that speeds within five percent of those theoretically possible can be achieved. A special-purpose numerical technique, the stochastic preconditioned conjugate gradient method, is also presented and demonstrated to be extremely efficient for certain classes of PSA problems.

Sues, Robert H.↗

Parallel Computing for Probabilistic Response Analysis of High Temperature Composites

The objective of this Phase I research was to establish the required software and hardware strategies to achieve large scale parallelism in solving PCM problems. To meet this objective, several investigations were conducted. First, we identified the multiple levels of parallelism in PCM and the computational strategies to exploit these parallelisms. Next, several software and hardware efficiency investigations were conducted. These involved the use of three different parallel programming paradigms and solution of two example problems on both a shared-memory multiprocessor and a distributed-memory network of workstations.

Sues, R. H.↗

Discrete Event Modeling and Massively Parallel Execution of Epidemic Outbreak Phenomena

In complex phenomena such as epidemiological outbreaks, the intensity of inherent feedback effects and the significant role of transients in the dynamics make simulation the only effective method for proactive, reactive or post-facto analysis. The spatial scale, runtime speed, and behavioral detail needed in detailed simulations of epidemic outbreaks make it necessary to use large-scale parallel processing. Here, an optimistic parallel execution of a new discrete event formulation of a reaction-diffusion simulation model of epidemic propagation is presented to facilitate in dramatically increasing the fidelity and speed by which epidemiological simulations can be performed. Rollback support needed during optimistic parallel execution is achieved by combining reverse computation with a small amount of incremental state saving. Parallel speedup of over 5,500 and other runtime performance metrics of the system are observed with weak-scaling execution on a small (8,192-core) Blue Gene / P system, while scalability with a weak-scaling speedup of over 10,000 is demonstrated on 65,536 cores of a large Cray XT5 system. Scenarios representing large population sizes exceeding several hundreds of millions of individuals in the largest cases are successfully exercised to verify model scalability.

99 GENERAL AND MISCELLANEOUS↗

Extended temperature-accelerated dynamics: Enabling long-time full-scale modeling of large rare-event systems

A new method, the Extended Temperature-Accelerated Dynamics (XTAD), is introduced for modeling long-timescale evolution of large rare-event systems. The method is based on the Temperature-Accelerated Dynamics approach [M. Sørensen and A. Voter, J. Chem. Phys. 112, 9599 (2000)], but uses full-scale parallel molecular dynamics simulations to probe a potential energy surface of an entire system, combined with the adaptive on-the-fly system decomposition for analyzing the energetics of rare events. The method removes limitations on a feasible system size and enables to handle simultaneous diffusion events, including both large-scale concerted and local transitions. Due to the intrinsically parallel algorithm, XTAD not only allows studies of various diffusion mechanisms in solid state physics, but also opens the avenue for atomistic simulations of a range of technologically relevant processes in material science, such as thin film growth on nano- and microstructured surfaces.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Best Practices and Lessons Learned from Deploying and Operating Large-Scale Data-Centric Parallel File Systems, In: SC '14: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis

The Oak Ridge Leadership Computing Facility (OLCF) has deployed multiple large-scale parallel file systems (PFS) to support its operations. During this process, OLCF acquired significant expertise in large-scale storage system design, file system software development, technology evaluation, benchmarking, procurement, deployment, and operational practices. Based on the lessons learned from each new PFS deployment, OLCF improved its operating procedures, and strategies. This paper provides an account of our experience and lessons learned in acquiring, deploying, and operating large-scale parallel file systems. We believe that these lessons will be useful to the wider HPC community.

Computer Science↗

Exploiting variability for energy optimization of parallel programs

Here in this paper we present optimizations that use DVFS mechanisms to reduce the total energy usage in scientific applications. Our main insight is that noise is intrinsic to large scale parallel executions and it appears whenever shared resources are contended. The presence of noise allows us to identify and manipulate any program regions amenable to DVFS. When compared to previous energy optimizations that make per core decisions using predictions of the running time, our scheme uses a qualitative approach to recognize the signature of executions amenable to DVFS. By recognizing the "shape of variability" we can optimize codes with highly dynamic behavior, which pose challenges to all existing DVFS techniques. We validate our approach using offline and online analyses for one-sided and two-sided communication paradigms. We have applied our methods to NWChem, and we show best case improvements in energy use of 12% at no loss in performance when using online optimizations running on 720 Haswell cores with one-sided communication. With NWChem on MPI two-sided and offline analysis, capturing the initialization, we find energy savings of up to 20%, with less than 1% performance cost.

97 MATHEMATICS AND COMPUTING↗

pcircle - A Suite of Scalable Parallel File System Tools

Most of the software related to file system are written for conventional local file system, they are serialized and can't take advantage of the benefit of a large scale parallel file system. "pcircle" software builds on top of ubiquitous MPI in cluster computing environment and "work-stealing" pattern to provide a scalable, high-performance suite of file system tools. In particular - it implemented parallel data copy and parallel data checksumming, with advanced features such as async progress report, checkpoint and restart, as well as integrity checking.

WANG, FEIYI↗

pcircle - A Suite of Scalable Parallel File System Tools

Most of the software related to file system are written for conventional local file system, they are serialized and can't take advantage of the benefit of a large scale parallel file system. "pcircle" software builds on top of ubiquitous MPI in cluster computing environment and "work-stealing" pattern to provide a scalable, high-performance suite of file system tools. In particular - it implemented parallel data copy and parallel data checksumming, with advanced features such as async progress report, checkpoint and restart, as well as integrity checking.

WANG, FEIYI↗

Generation of Reversible C++ Code for Optimistic Parallel Discrete Event Simulation

The reversible execution of C/C++ code has been a target of research and engineering for more than a decade as reversible computation has become a central notion in large-scale parallel discrete event simulation (PDES). The simulation models that are implemented for PDES are of increasing complexity and size and require various language features to support abstraction, encapsulation, and composition when building a simulation model. In this paper, we focus on parallel simulation models that are written with user-defined C++ abstractions and abstractions of the C++ Standard Library. We present an approach based on incremental state saving for establishing reversibility of C++ and an evaluation for a kinetic Monte-Carlo simulation implemented in C++. Although a significant runtime overhead is introduced with our technique, it is an enormous win that it allows using the entire C++ language, and has that code automatically transformed into reversible code to enable parallel execution with the Rensselaer’s optimistic simulation system (ROSS).

97 MATHEMATICS AND COMPUTING↗

Time-parallel solutions to differential equations via functional optimization

We describe an approach to solving a generic time-dependent differential equation (DE) that recasts the problem as a functional optimization one. The techniques employed to solve for the functional minimum, which we relate to the Sobolev Gradient method, allow for large-scale parallelization in time, and therefore potential faster “wall-clock” time computing on machines with significant parallel computing capacity. We are able to come up with numerous different discretizations and approximations for our optimization-derived equations, each of which either puts an existing approach, the Parareal method, in an optimization context, or provides a new time-parallel (TP) scheme with potentially faster convergence to the DE solution. We describe how the approach is particularly effective for solving multiscale DEs and present TP schemes that incorporate two different solution scales. Sample results are provided for three differential equations, solved with TP schemes, and we discuss how the choice of TP scheme can have an orders of magnitude effect on the accuracy or convergence rate.

97 MATHEMATICS AND COMPUTING↗

A Scalable Graph Analytics Framework for Programming with Big Data in R (pbdR)

Many disciplines such as biology, economics, engineering, physics, and the social sciences represent their data as graphs to capture patterns, trends, and associations. There are are many commercially available graph libraries in different programming languages to analyze these complex graphs. But there is no distributed graph library package in R - the popular statistical programming language to analyze graphs that bigger than a single machine's memory. Many domain experts prefer R over the numerous other alternatives. Towards this, we present a distributed graph analytics framework for R called programming with big graph using R (pBGR.) Our proposed framework leverages the Programming with Big Data in R (pbdR) ecosystem that provides scalable R packages for distributed computing in data science. We present an early prototype implementation of this framework using the distributed-memory parallel graph library CombBLAS and evaluate the framework's performance on leadership class computing platforms. Our experimental results demonstrate that the proposed framework is capable of performing large-scale parallel graph mining through the easyto-use R language. This enhanced graph processing capability coupled with other statistical tools already available in R, should be valuable to many domain experts.

Hasan, S M Shamimul↗

Xyce parallel electronic simulator : reference guide.

This document is a reference guide to the Xyce Parallel Electronic Simulator, and is a companion document to the Xyce Users Guide. The focus of this document is (to the extent possible) exhaustively list device parameters, solver options, parser options, and other usage details of Xyce. This document is not intended to be a tutorial. Users who are new to circuit simulation are better served by the Xyce Users Guide. The Xyce Parallel Electronic Simulator has been written to support, in a rigorous manner, the simulation needs of the Sandia National Laboratories electrical designers. It is targeted specifically to run on large-scale parallel computing platforms but also runs well on a variety of architectures including single processor workstations. It also aims to support a variety of devices and models specific to Sandia needs. This document is intended to complement the Xyce Users Guide. It contains comprehensive, detailed information about a number of topics pertinent to the usage of Xyce. Included in this document is a netlist reference for the input-file commands and elements supported within Xyce; a command line reference, which describes the available command line arguments for Xyce; and quick-references for users of other circuit codes, such as Orcad's PSpice and Sandia's ChileSPICE.

99 GENERAL AND MISCELLANEOUS↗

An Asynchronous Many-Task Implementation of In-Situ Statistical Analysis using Legion.

In this report, we propose a framework for the design and implementation of in-situ analy- ses using an asynchronous many-task (AMT) model, using the Legion programming model together with the MiniAero mini-application as a surrogate for full-scale parallel scientific computing applications. The bulk of this work consists of converting the Learn/Derive/Assess model which we had initially developed for parallel statistical analysis using MPI [PTBM11], from a SPMD to an AMT model. In this goal, we propose an original use of the concept of Legion logical regions as a replacement for the parallel communication schemes used for the only operation of the statistics engines that require explicit communication. We then evaluate this proposed scheme in a shared memory environment, using the Legion port of MiniAero as a proxy for a full-scale scientific application, as a means to provide input data sets of variable size for the in-situ statistical analyses in an AMT context. We demonstrate in particular that the approach has merit, and warrants further investigation, in collaboration with ongoing efforts to improve the overall parallel performance of the Legion system.

Pebay, Philippe Pierre↗

Magnetic field line random walk in isotropic turbulence with varying mean field

In astrophysical plasmas, the magnetic field line random walk (FLRW) plays an important role in guiding particle transport. The FLRW behavior is scaled by the Kubo number R=(b/B{sub 0})(ℓ{sub ∥}/ℓ{sub ⊥}) for rms magnetic fluctuation b, large-scale mean field B{sub 0}, and coherence scales parallel (ℓ{sub ∥}) and perpendicular (ℓ{sub ⊥}) to B{sub 0}. Here we use a nonperturbative analytic framework based on Corrsin’s hypothesis, together with direct computer simulations, to examine the R-scaling of the FLRW for varying B {sub 0} with finite b and isotropic fluctuations with ℓ{sub ∥}/ℓ{sub ⊥}=1, instead of the well-studied route of varying ℓ{sub ∥}/ℓ{sub ⊥} for b ≪B{sub 0}. The FLRW for isotropic magnetic fluctuations is also of astrophysical interest regarding transport processes in the interstellar medium. With a mean field, fluctuations may have variance anisotropy, so we consider limiting cases of isotropic variance and transverse variance (with b {sub z} = 0). We obtain analytic theories, and closed-form solutions for extreme cases. Padé approximants are provided to interpolate all versions of theory and simulations to any B {sub 0}. We demonstrate that, for isotropic turbulence, Corrsin-based theories generally work well, and with increasing R there is a transition from quasilinear to Bohm diffusion. This holds even with b {sub z} = 0, when different routes to R→∞ are mathematically equivalent; in contrast with previous studies, we find that a Corrsin-based theory with random ballistic decorrelation works well even up to R = 400, where the effects of trapping are barely perceptible in simulation results.

79 ASTRONOMY AND ASTROPHYSICS↗