Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallelization”

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 289 records · Page 16

Parallelization and checkpointing of GPU applications through program transformation

GPUs have emerged as a powerful tool for accelerating general-purpose applications. The availability of programming languages that makes writing general-purpose applications for running on GPUs tractable have consolidated GPUs as an alternative for accelerating general purpose applications. Among the areas that have benefited from GPU acceleration are: signal and image processing, computational fluid dynamics, quantum chemistry, and, in general, the High Performance Computing (HPC) Industry. In order to continue to exploit higher levels of parallelism with GPUs, multi-GPU systems are gaining popularity. In this context, single-GPU applications are parallelized for running in multi-GPU systems. Furthermore, multi-GPU systems help to solve the GPU memory limitation for applications with large application memory footprint. Parallelizing single-GPU applications has been approached by libraries that distribute the workload at runtime, however, they impose execution overhead and are not portable. On the other hand, on traditional CPU systems, parallelization has been approached through application transformation at pre-compile time, which enhances the application to distribute the workload at application level and does not have the issues of library-based approaches. Hence, a parallelization scheme for GPU systems based on application transformation is needed. Like any computing engine of today, reliability is also a concern in GPUs. GPUs are vulnerable to transient and permanent failures. Current checkpoint/restart techniques are not suitable for systems with GPUs. Checkpointing for GPU systems present new and interesting challenges, primarily due to the natural differences imposed by the hardware design, the memory subsystem architecture, the massive number of threads, and the limited amount of synchronization among threads. Therefore, a checkpoint/restart technique suitable for GPU systems is needed. The goal of this work is to exploit higher levels of parallelism and to develop support for application-level fault tolerance in applications using multiple GPUs. Our techniques reduce the burden of enhancing single-GPU applications to support these features. To achieve our goal, this work designs and implements a framework for enhancing a single-GPU OpenCL application through application transformation.

97 MATHEMATICS AND COMPUTING↗

Series and parallel arc-fault circuit interrupter tests.

While the 2011 National Electrical Codeª (NEC) only requires series arc-fault protection, some arc-fault circuit interrupter (AFCI) manufacturers are designing products to detect and mitigate both series and parallel arc-faults. Sandia National Laboratories (SNL) has extensively investigated the electrical differences of series and parallel arc-faults and has offered possible classification and mitigation solutions. As part of this effort, Sandia National Laboratories has collaborated with MidNite Solar to create and test a 24-string combiner box with an AFCI which detects, differentiates, and de-energizes series and parallel arc-faults. In the case of the MidNite AFCI prototype, series arc-faults are mitigated by opening the PV strings, whereas parallel arc-faults are mitigated by shorting the array. A range of different experimental series and parallel arc-fault tests with the MidNite combiner box were performed at the Distributed Energy Technologies Laboratory (DETL) at SNL in Albuquerque, NM. In all the tests, the prototype de-energized the arc-faults in the time period required by the arc-fault circuit interrupt testing standard, UL 1699B. The experimental tests confirm series and parallel arc-faults can be successfully mitigated with a combiner box-integrated solution.

Johnson, Jay Dean↗

A Novel Algorithm for Solving the Multidimensional Neutron Transport Equation on Massively Parallel Architectures

We employ the Integral Transport Matrix Method (ITMM) as the kernel of new parallel solution methods for the discrete ordinates approximation of the within-group neutron transport equation. The ITMM abandons the repetitive mesh sweeps of the traditional source iterations (SI) scheme in favor of constructing stored operators that account for the direct coupling factors among all the cells' fluxes and between the cells' and boundary surfaces' fluxes. The main goals of this work are to develop the algorithms that construct these operators and employ them in the solution process, determine the most suitable way to parallelize the entire procedure, and evaluate the behavior and parallel performance of the developed methods with increasing number of processes, P. The fastest observed parallel solution method, Parallel Gauss-Seidel (PGS), was used in a weak scaling comparison with the PARTISN transport code, which uses the source iteration (SI) scheme parallelized with the Koch-baker-Alcouffe (KBA) method. Compared to the state-of-the-art SI-KBA with diffusion synthetic acceleration (DSA), this new method- even without acceleration/preconditioning-is completitive for optically thick problems as P is increased to the tens of thousands range. For the most optically thick cells tested, PGS reduced execution time by an approximate factor of three for problems with more than 130 million computational cells on P = 32,768. Moreover, the SI-DSA execution times's trend rises generally more steeply with increasing P than the PGS trend. Furthermore, the PGS method outperforms SI for the periodic heterogeneous layers (PHL) configuration problems. The PGS method outperforms SI and SI-DSA on as few as P = 16 for PHL problems and reduces execution time by a factor of ten or more for all problems considered with more than 2 million computational cells on P = 4.096.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

A parallel workflow implementation for PEST version 13.6 in high-performance computing for WRF-Hydro version 5.0: a case study over the midwestern United States

The Weather Research and Forecasting Hydrological (WRF-Hydro) system is a state-of-the-art numerical model that models the entire hydrological cycle based on physical principles. As with other hydrological models, WRF-Hydro parameterizes many physical processes. Hence, WRF-Hydro needs to be calibrated to optimize its output with respect to observations for the application region. When applied to a relatively large domain, both WRF-Hydro simulations and calibrations require intensive computing resources and are best performed on multimode, multicore high-performance computing (HPC) systems. Typically, each physics-based model requires a calibration process that works specifically with that model and is not transferrable to a different process or model. The parameter estimation tool (PEST) is a flexible and generic calibration tool that can be used in principle to calibrate any of these models. In its existing configuration, however, PEST is not designed to work on the current generation of massively parallel HPC clusters. To address this issue, we ported the parallel PEST to HPCs and adapted it to work with WRF-Hydro. The porting involved writing scripts to modify the workflow for different workload managers and job schedulers, as well as to connect the parallel PEST to WRF-Hydro. To test the operational feasibility and the computational benefits of this first-of-its-kind HPC-enabled parallel PEST, we developed a case study using a flood in the midwestern United States in 2013. Results on a problem involving the calibration of 22 parameters show that on the same computing resources used for parallel WRF-Hydro, the HPC-enabled parallel PEST can speed up the calibration process by a factor of up to 15 compared with commonly used PEST in sequential mode. The speedup factor is expected to be greater with a larger calibration problem (e.g., more parameters to be calibrated or a larger size of study area).

54 ENVIRONMENTAL SCIENCES↗

OpenMP-style parallelism in data-centered multicore computing with R

R is a domain specific language widely used for data analysis by the statistics community as well as by researchers in finance, biology, social sciences, and many other disciplines. As R programs are linked to input data, the exponential growth of available data makes high-performance computing with R imperative. To ease the process of writing parallel programs in R, code transformation from a sequential program to a parallel version would bring much convenience to R users. In this paper, we present our work in semiautomatic parallelization of R codes with user-added OpenMPstyle pragmas. While such pragmas are used at the frontend, we take advantage of multiple parallel backends with different R packages. We provide flexibility for importing parallelism with plug-in components, impose built-in MapReduce for data processing, and also maintain code reusability. We illustrate the advantage of the on-the-fly mechanisms which can lead to significant applications in data-centered parallel computing.

Jiang, Lei↗

A model for optimizing file access patterns using spatio-temporal parallelism

For many years now, I/O read time has been recognized as the primary bottleneck for parallel visualization and analysis of large-scale data. In this paper, we introduce a model that can estimate the read time for a file stored in a parallel filesystem when given the file access pattern. Read times ultimately depend on how the file is stored and the access pattern used to read the file. The file access pattern will be dictated by the type of parallel decomposition used. We employ spatio-temporal parallelism, which combines both spatial and temporal parallelism, to provide greater flexibility to possible file access patterns. Using our model, we were able to configure the spatio-temporal parallelism to design optimized read access patterns that resulted in a speedup factor of approximately 400 over traditional file access patterns.

Boonthanome, Nouanesengsy↗

Self-balanced modulation and magnetic rebalancing method for parallel multilevel inverters

A self-balanced modulation method and a closed-loop magnetic flux rebalancing control method for parallel multilevel inverters. The combination of the two methods provides for balancing of the magnetic flux of the inter-cell transformers (ICTs) of the parallel multilevel inverters without deteriorating the quality of the output voltage. In various embodiments a parallel multi-level inverter modulator is provide including a multi-channel comparator to generate a multiplexed digitized ideal waveform for a parallel multi-level inverter and a finite state machine (FSM) module coupled to the parallel multi-channel comparator, the FSM module to receive the multiplexed digitized ideal waveform and to generate a pulse width modulated gate-drive signal for each switching device of the parallel multi-level inverter. The system and method provides for optimization of the output voltage spectrum without influence the magnetic balancing.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Towards Generic Parallel Programming in Computer Science Education with Kokkos

Parallel patterns, views, and spaces are promising abstractions to capture the programmer's intent as well as the contextual information that can be used by an underlying runtime to efficiently map software to parallel hardware. These abstractions can be valuable in cases where an algorithm must accommodate requirements of code and performance portability across hardware architectures and vendor programming models. Kokkos is a parallel programming model for host- and accelerator architectures that relies on these abstractions and targets these requirements. It consists of a pure C++ interface, a specification, and a programming library. The programming library exposes patterns and types and maps them to an underlying abstract machine model. The abstract machine model offers a generic view of parallel hardware. While Kokkos is gaining popularity in large-scale HPC applications at some DOE laboratories, we believe that the implemented concepts are of interest to a broader audience including academia as they may contribute to a generic, vendor, and architecture-independent education of parallel programming. In this work, we give an insight into the design considerations of this programming model and list important abstractions. Further, we document best practices obtained from giving virtual classes on Kokkos and give pointers to resources that the reader may consider valuable for a lecture on generic parallel programming for students with preexisting knowledge on this matter.

Ciesko, Jan↗

Simulation and Optimizing Groundwater Investigations, Conceptual Design Analysis, Model Calibration and Remedial Design Using Parallel Numerical Methods: Theory and Application - 16578

Optimization of environmental restoration efforts at complex groundwater sites involves the development of a conceptual and often numerical flow and transport models. Typically, a limited number of simulations are done to assess the impact of parameter uncertainty, to evaluate remedial alternatives and develop a remedy, or to modify an existing remedy. A limited number of simulations are performed because each simulation can take upwards of several hours per simulation and until now the ability to conduct the 100's to 1000's of simulations needed to derive an optimal remedy was beyond the reach of the everyday practitioner. HGL has developed and implemented new parallel simulation and optimization methods in their software and has successfully deployed applications on high performance Cloud-based computing platforms. This advancement removes the impediment of burdensome and unpredictable runtime requirements these types of projects have typically exhibited using heuristic trial-and-error or even automated serial computation approaches. A parallel numerical solver has been added to the subsurface flow and transport simulator MODFLOW-SURFACT{sup TM} and has demonstrated reductions in solution time by as much as a factor of 9.8. This is accomplished by distributing the numerical computations of every time step over multiple processors. The integrated parallel simulation and optimization method uses the breadth and depth search approach; iteratively producing sets of tens or 100's of candidate solutions that are evaluated concurrently instead sequentially (i.e. one model evaluation at a time). Solutions within 80-100% of the globally optimal solution are regularly found expending only 5-10% of the usual computational effort using this new parallel optimization method (reducing computation effort by 10 to 20 times). Distributing optimal search runs across multiple processors additionally reduces calendar time to solutions. Practitioners evaluating multiple conceptual site models, model calibration by auto-calibration methods and remedial design by heuristics can enjoy significant reductions in model solution time by using the parallel numerical matrix solver in the physics-based simulator. Combining investigation and design analysis with breadth and depth parallel optimization methods, projects at the fringe of practical optimization, along with projects that were beyond the reach of formal optimization methods are now wholly feasible and tractable optimization problems. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Techniques for recovering from errors when executing software applications on parallel processors

In various embodiments, a software program uses hardware features of a parallel processor to checkpoint a context associated with an execution of a software application on the parallel processor. The software program uses a preemption feature of the parallel processor to cause the parallel processor to stop executing instructions in accordance with the context. The software program then causes the parallel processor to collect state data associated with the context. After generating a checkpoint based on the state data, the software program causes the parallel processor to resume executing instructions in accordance with the context.

Hukerikar, Saurabh↗

GSplit: Scaling Graph Neural Network Training on Large Graphs via Split-Parallelism

Graph neural networks (GNNs), an emerging class of machine learning models for graphs, have gained popularity for their superior performance in various graph analytical tasks. Mini-batch training is commonly used to train GNNs on large graphs, and data parallelism is the standard approach to scale mini-batch training across multiple GPUs. Data parallel approaches contain redundant work as subgraphs sampled by different GPUs contain significant overlap. To address this issue, we introduce a hybrid parallel mini-batch training paradigm called Split parallelism. Split parallelism avoids redundant work by splitting the sampling, loading, and training of each mini-batch across multiple GPUs. Split parallelism, however, introduces communication overheads that can be more than the savings from removing redundant work. We further present a lightweight partitioning algorithm that probabilistically minimizes these overheads. We implement spllit parllelism in GSplit and show that it outperforms state-of-the-art mini-batch training systems like DGL, Quiver, and P3.

Lim, Seung-Hwan [ORNL] (ORCID:0000000194616866)↗

The Case for Strong Scaling in Deep Learning: Training Large 3D CNNs with Hybrid Parallelism

Here, we present scalable hybrid-parallel algorithms for training large-scale 3D convolutional neural networks. Deep learning-based emerging scientific workflows often require model training with large, high-dimensional samples, which can make training much more costly and even infeasible due to excessive memory usage. We solve these challenges by extensively applying hybrid parallelism throughout the end-to-end training pipeline, including both computations and I/O. Our hybrid-parallel algorithm extends the standard data parallelism with spatial parallelism, which partitions a single sample in the spatial domain, realizing strong scaling beyond the mini-batch dimension with a larger aggregated memory capacity. We evaluate our proposed training algorithms with two challenging 3D CNNs, CosmoFlow and 3D U-Net. Our comprehensive performance studies show that good weak and strong scaling can be achieved for both networks using up to 2K GPUs. More importantly, we enable training of CosmoFlow with much larger samples than previously possible, realizing an order-of-magnitude improvement in prediction accuracy.

97 MATHEMATICS AND COMPUTING↗

Instability Issue of Paralleled Dies in an SiC Power Module in Solid-State Circuit Breaker Applications

Paralleled dies in a power module could have instability issues during high current switching transients. Here, the instability is caused by the differential-mode oscillation among paralleled MOSFETs. Conventional analyses of paralleled MOSFETs’ stability are normally limited to a single operating point, which ignores the influences of the switching trajectory and nonlinear device parameters on stability. This article reveals that the switching trajectory can significantly influence parallel stability. The analysis is improved by solving eigenvalues of state-space modeling system matrices of all operating points that the switching trajectory goes through considering nonlinear device parameters. Higher voltage and current stresses result in greater real parts of complex eigenvalues, which explains why the paralleled MOSFETs are more unstable with higher voltage and current stresses. To improve stability in solid-state circuit breaker applications, we propose a method to manipulate the switching trajectory to avoid the unstable region where the conventional hard switching trajectory normally goes through. Experimental results show that the turn- off current capability can be increased from ~five times of rated current with the gate oscillation using the conventional turn- off trajectory to ~ten times of rated current without the gate oscillation using the optimal turn- off trajectory.

42 ENGINEERING↗

Layer-Parallel Training of Deep Residual Neural Networks

Residual neural networks (ResNets) are a promising class of deep neural networks that have shown excellent performance for a number of learning tasks, e.g., image classification and recognition. Mathematically, ResNet architectures can be interpreted as forward Euler discretizations of a nonlinear initial value problem whose time-dependent control variables represent the weights of the neural network. Hence, training a ResNet can be cast as an optimal control problem of the associated dynamical system. For similar time-dependent optimal control problems arising in engineering applications, parallel-in-time methods have shown notable improvements in scalability. This paper demonstrates the use of those techniques for efficient and effective training of ResNets. The proposed algorithms replace the classical (sequential) forward and backward propagation through the network layers with a parallel nonlinear multigrid iteration applied to the layer domain. This adds a new dimension of parallelism across layers that is attractive when training very deep networks. From this basic idea, we derive multiple layer-parallel methods. The most efficient version employs a simultaneous optimization approach where updates to the network parameters are based on inexact gradient information in order to speed up the training process. Finally, using numerical examples from supervised classification, we demonstrate that the new approach achieves a training performance similar to that of traditional methods, but enables layer-parallelism and thus provides speedup over layer-serial methods through greater concurrency.

97 MATHEMATICS AND COMPUTING↗

A parallel p ‐adaptive discontinuous Galerkin method for the Euler equations with dynamic load‐balancing on tetrahedral grids

Abstract A novel p ‐adaptive discontinuous Galerkin (DG) method has been developed to solve the Euler equations on three‐dimensional tetrahedral grids. Hierarchical orthogonal basis functions are adopted for the DG spatial discretization while a third order TVD Runge‐Kutta method is used for the time integration. A vertex‐based limiter is applied to the numerical solution in order to eliminate oscillations in the high order method. An error indicator constructed from the solution of order and is used to adapt degrees of freedom in each computational element, which remarkably reduces the computational cost while still maintaining an accurate solution. The developed method is implemented with under the Charm++ parallel computing framework. Charm++ is a parallel computing framework that includes various load‐balancing strategies. Implementing the numerical solver under Charm++ system provides us with access to a suite of dynamic load balancing strategies. This can be efficiently used to alleviate the load imbalances created by p ‐adaptation. A number of numerical experiments are performed to demonstrate both the numerical accuracy and parallel performance of the developed p ‐adaptive DG method. It is observed that the unbalanced load distribution caused by the parallel p ‐adaptive DG method can be alleviated by the dynamic load balancing from Charm++ system. Due to this, high performance gain can be achieved. For the testcases studied in the current work, the parallel performance gain ranged from 1.5× to 3.7×. Therefore, the developed p ‐adaptive DG method can significantly reduce the total simulation time in comparison to the standard DG method without p ‐adaptation.

97 MATHEMATICS AND COMPUTING↗

NWChem: scalable parallel computational chemistry

NWChem is a general purpose computational chemistry code specifically designed to run on distributed memory parallel computers. The core functionality of the code focuses on molecular dynamics, Hartree-Fock and density functional theory methods for both plane-wave basis sets as well as Gaussian basis sets, tensor contraction engine based coupled cluster capabilities and combined quantum mechanics/molecular mechanics descriptions. It was realized from the beginning that scalable implementations of these methods required a programming paradigm inherently different from what message passing approaches could offer. In response a global address space library, the Global Array Toolkit, was developed. The programming model it offers is based on using predominantly one-sided communication. This model underpins most of the functionality in NWChem and the power of it is exemplified by the fact that the code scales to tens of thousands of processors. In this paper the core capabilities of NWChem are described as well as their implementation to achieve an efficient computational chemistry code with high parallel scalability. NWChem is a modern, open source, computational chemistry code1 specifically designed for large scale parallel applications2. To meet the challenges of developing efficient, scalable and portable programs of this nature a particular code design was adopted. This code design involved two main features. First of all, the code is build up in a modular fashion so that a large variety of functionality can be integrated easily. Secondly, to facilitate writing complex parallel algorithms the Global Array toolkit was developed. This toolkit allows one to write parallel applications in a shared memory like approach, but offers additional mechanisms to exploit data locality to lower communication overheads. This framework has proven to be very successful in computational chemistry but is applicable to any engineering domain. Within the context created by the features above NWChem has grown into a general purpose computational chemistry code that supports a wide variety of energy expressions and capabilities to calculate properties based there upon. The main energy expressions are classical mechanics force fields, Hartree-Fock and DFT both for finite systems and condensed phase systems, coupled cluster, as well as QM/MM. For most energy expressions single point calculations, geometry optimizations, excited states, and other properties are available. Below we briefly discuss each of the main energy expressions and the critical points involved in scalable implementations thereof.

99 GENERAL AND MISCELLANEOUS↗

A non-intrusive parallel-in-time adjoint solver with the XBraid library

In this paper, an adjoint solver for the multigrid-in-time software library XBraid is presented. XBraid provides a non-intrusive approach for simulating unsteady dynamics on multiple processors while parallelizing not only in space but also in the time domain (XBraid: Parallel multigrid in time, http://llnl.gov/casc/xbraid). It applies an iterative multigrid reduction in time algorithm to existing spatially parallel classical time propagators and computes the unsteady solution parallel in time. Techniques from Automatic Differentiation are used to develop a consistent discrete adjoint solver which provides sensitivity information of output quantities with respect to design parameter changes. The adjoint code runs backwards through the primal XBraid actions and accumulates gradient information parallel in time. It is highly non-intrusive as existing adjoint time propagators can easily be integrated through the adjoint interface. The adjoint code is validated on advection-dominated flow with periodic upstream boundary condition. We report it provides similar strong scaling results as the primal XBraid solver and offers great potential for speeding up the overall computational costs for sensitivity analysis using multiple processors.

97 MATHEMATICS AND COMPUTING↗

ExaHDF5: Delivering Efficient Parallel I/O on Exascale Computing Systems

Scientific applications at exascale generate and analyze massive amounts of data. A critical requirement of these applications is the capability to access and manage this data efficiently on exascale systems. As such, parallel I/O, the key technology enables moving data between compute nodes and storage, faces monumental challenges from new applications, memory, and storage architectures considered in the designs of exascale systems. As the storage hierarchy is expanding to include node-local persistent memory, burst buffers, etc., as well as disk-based storage, data movement among these layers must be efficient. Parallel I/O libraries of the future should be capable of handling file sizes of many terabytes and beyond. In this paper, we describe new capabilities we have developed in Hierarchical Data Format version 5 (HDF5), the most popular parallel I/O library for scientific applications. HDF5 is one of the most used libraries at the leadership computing facilities for performing parallel I/O on existing HPC systems. The state-of-the-art features we describe include: Virtual Object Layer (VOL), Data Elevator, asynchronous I/O, full-featured single-writer and multiple-reader (Full SWMR), and parallel querying. In this paper, we introduce these features, their implementations, and the performance and feature benefits to applications and other libraries.

97 MATHEMATICS AND COMPUTING↗