Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel implementation”

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 541 records · Page 30
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

An in-situ conductometric apparatus for physicochemical characterization of solutions and in-line monitoring of separation processes at elevated temperatures and pressures

Specific conductance and frequency-dependent resistance (impedance) data are widely utilized for understanding the physicochemical characteristics of aqueous and non-aqueous fluids and for evaluating the performance of chemical processes. However, the implementation of such an in-situ probe in high-temperature and high-pressure environments is not trivial. This work provides a description of both the hardware and software associated with implementing a parallel-type in-situ electrochemical sensor. The sensor can be used for in-line monitoring of thermal desalination processes and for impedance measurements in fluids at high temperature and pressure. Further, a comparison between the experimental measurements on the specific conductance in aqueous sodium chloride solutions and the conductance model demonstrate that the methodology yields reasonable agreement with both the model and literature data. A combination of hardware components, a software-based correction for experimental artifacts, and computational fluid dynamics (CFD) calculations used in this work provide a sound basis for implementing such in-situ electrochemical sensors to measure frequency-dependent resistance spectra.

47 OTHER INSTRUMENTATION↗
DOE OSTI2023

Extending OpenSHMEM with Aggregation Support for Improved Message Rate Performance

OpenSHMEM is a highly efficient one-sided communication API that implements the PGAS parallel programming model, and is known for its low latency communication operations that can be mapped efficiently to RDMA capabilities of network interconnects. However, applications that use OpenSHMEM can be sensitive to point-to-point message rates, as many-to-many communication patterns can generate large amounts of small messages which tend to overwhelm network hardware that has predominantly been optimised for bandwidth over message rate. Additionally, many important emerging classes of problems such as data analytics are similarly troublesome for the irregular access patterns they employ. Message aggregation strategies have been proven to significantly enhance network performance, but their implementation often involves complex restructuring of user code, making them unwieldy. This paper shows how to combine the best qualities of message aggregation within the communication model of OpenSHMEM such that applications with small and irregular access patterns can improve network performance while maintaining their algorithmic simplicity. We do this by providing a path to a message aggregation framework called conveyors through a minimally intrusive OpenSHMEM extension introducing aggregation contexts that fit more naturally to the OpenSHMEM atomics, gets, and puts model. We test these extensions using four of the bale 3.0 applications which contain essential many-to-many access patterns to show how they can produce performance improvements of up to 65×.

Welch, Aaron↗
DOE OSTI2017

Concurrent conversation modeling and parallel simulation of the naming game in social networks

The Naming Game is an effective self-organization model to understand the emergence of linguistic consensus and to investigate the system dynamics in a variety of phenomena over social networks of autonomous agents. The Naming Game is an effective description for the evolution of consensus despite the absence of any central coordination or specialized initialization even in large-scale networks. While the classical game is effective in description, it was defined with inherently sequential evaluation semantics over the entire network. Here, we develop a new concurrent model as a relaxation of the classical formulation and express it in a discrete event style of evaluation. Further, with the uncovered concurrency that was absent in the classical algorithm, we map the concurrent model to parallel discrete event simulation. Using a prototype implementation, we present an initial parallel performance study on networks containing hundreds of thousands of individuals, with a decrease in simulation time in the best-observed case from 4800 seconds down to 1400 seconds.

Perumalla, Kalyan↗
NASA NTRS1992

A general purpose subroutine for fast fourier transform on a distributed memory parallel machine

One issue which is central in developing a general purpose Fast Fourier Transform (FFT) subroutine on a distributed memory parallel machine is the data distribution. It is possible that different users would like to use the FFT routine with different data distributions. Thus, there is a need to design FFT schemes on distributed memory parallel machines which can support a variety of data distributions. An FFT implementation on a distributed memory parallel machine which works for a number of data distributions commonly encountered in scientific applications is presented. The problem of rearranging the data after computing the FFT is also addressed. The performance of the implementation on a distributed memory parallel machine Intel iPSC/860 is evaluated.

Dubey, A.↗
NASA NTRS2000

Implementation of BT, SP, LU, and FT of NAS Parallel Benchmarks in Java

A number of Java features make it an attractive but a debatable choice for High Performance Computing. We have implemented benchmarks working on single structured grid BT,SP,LU and FT in Java. The performance and scalability of the Java code shows that a significant improvement in Java compiler technology and in Java thread implementation are necessary for Java to compete with Fortran in HPC applications.

Schultz, Matthew↗
NASA NTRS1994

An implementation of a tree code on a SIMD, parallel computer

We describe a fast tree algorithm for gravitational N-body simulation on SIMD parallel computers. The tree construction uses fast, parallel sorts. The sorted lists are recursively divided along their x, y and z coordinates. This data structure is a completely balanced tree (i.e., each particle is paired with exactly one other particle) and maintains good spatial locality. An implementation of this tree-building algorithm on a 16k processor Maspar MP-1 performs well and constitutes only a small fraction (approximately 15%) of the entire cycle of finding the accelerations. Each node in the tree is treated as a monopole. The tree search and the summation of accelerations also perform well. During the tree search, node data that is needed from another processor is simply fetched. Roughly 55% of the tree search time is spent in communications between processors. We apply the code to two problems of astrophysical interest. The first is a simulation of the close passage of two gravitationally, interacting, disk galaxies using 65,636 particles. We also simulate the formation of structure in an expanding, model universe using 1,048,576 particles. Our code attains speeds comparable to one head of a Cray Y-MP, so single instruction, multiple data (SIMD) type computers can be used for these simulations. The cost/performance ratio for SIMD machines like the Maspar MP-1 make them an extremely attractive alternative to either vector processors or large multiple instruction, multiple data (MIMD) type parallel computers. With further optimizations (e.g., more careful load balancing), speeds in excess of today's vector processing computers should be possible.

Olson, Kevin M.↗
NASA NTRS1988

Estimating water flow through a hillslope using the massively parallel processor

A new two-dimensional model of water flow in a hillslope has been implemented on the Massively Parallel Processor at the Goddard Space Flight Center. Flow in the soil both in the saturated and unsaturated zones, evaporation and overland flow are all modelled, and the rainfall rates are allowed to vary spatially. Previous models of this type had always been very limited computationally. This model takes less than a minute to model all the components of the hillslope water flow for a day. The model can now be used in sensitivity studies to specify which measurements should be taken and how accurate they should be to describe such flows for environmental studies.

Devaney, Judy E.↗
NASA NTRS1993

Generating local addresses and communication sets for data-parallel programs

Generating local addresses and communication sets is an important issue in distributed-memory implementations of data-parallel languages such as High Performance FORTRAN. We show that, for an array A affinely aligned to a template that is distributed across p processors with a cyclic(k) distribution and a computation involving the regular section A(l:h:s), the local memory access sequence for any processor is characterized by a finite state machine of at most k states. We present fast algorithms for computing the essential information about these state machines, and extend the framework to handle multidimensional arrays. We also show how to generate communication sets using the state machine approach. Performance results show that this solution requires very little run-time overhead and acceptable preprocessing time.

Chatterjee, Siddhartha↗
NASA NTRS1993

Generating local addresses and communication sets for data-parallel programs

Generating local addresses and communication sets is an important issue in distributed-memory implementations of data-parallel languages such as High Performance Fortran. We show that for an array A affinely aligned to a template that is distributed across p processors with a cyclic(k) distribution, and a computation involving the regular section A, the local memory access sequence for any processor is characterized by a finite state machine of at most k states. We present fast algorithms for computing the essential information about these state machines, and extend the framework to handle multidimensional arrays. We also show how to generate communication sets using the state machine approach. Performance results show that this solution requires very little runtime overhead and acceptable preprocessing time.

Chatterjee, Siddhartha↗
NASA NTRS2014

Parallelization of Unsteady Adaptive Mesh Refinement for Unstructured Navier-Stokes Solvers

This paper explores the implementation of the MPI parallelization in a Navier-Stokes solver using adaptive mesh re nement. Viscous and inviscid test problems are considered for the purpose of benchmarking, as are implicit and explicit time advancement methods. The main test problem for comparison includes e ects from boundary layers and other viscous features and requires a large number of grid points for accurate computation. Ex- perimental validation against double cone experiments in hypersonic ow are shown. The adaptive mesh re nement shows promise for a staple test problem in the hypersonic com- munity. Extension to more advanced techniques for more complicated ows is described.

Schwing, Alan M.↗
NASA NTRS1987

A parallelized elliptic solver for reacting flows

A modified Newton algorithm for the solution of nonlinear elliptic boundary value problems via finite discretization methods is presented. A serial implementation of this algorithm which has recently been applied successfully to the computation of an axisymmetric over-ventilated subsonic laminar methane-air jet diffusion flame is described. Parallel implementation issues and a complexity theory are presented. Included as well are actual performance data for model systems obtained on the Intel Hypercube and a discussion of its implications for modeling realistic systems.

Keyes, David E.↗
NASA NTRS2003

Hierarchical Image Segmentation of Remotely Sensed Data using Massively Parallel GNU-LINUX Software

A hierarchical set of image segmentations is a set of several image segmentations of the same image at different levels of detail in which the segmentations at coarser levels of detail can be produced from simple merges of regions at finer levels of detail. In [1], Tilton, et a1 describes an approach for producing hierarchical segmentations (called HSEG) and gave a progress report on exploiting these hierarchical segmentations for image information mining. The HSEG algorithm is a hybrid of region growing and constrained spectral clustering that produces a hierarchical set of image segmentations based on detected convergence points. In the main, HSEG employs the hierarchical stepwise optimization (HSWO) approach to region growing, which was described as early as 1989 by Beaulieu and Goldberg. The HSWO approach seeks to produce segmentations that are more optimized than those produced by more classic approaches to region growing (e.g. Horowitz and T. Pavlidis, [3]). In addition, HSEG optionally interjects between HSWO region growing iterations, merges between spatially non-adjacent regions (i.e., spectrally based merging or clustering) constrained by a threshold derived from the previous HSWO region growing iteration. While the addition of constrained spectral clustering improves the utility of the segmentation results, especially for larger images, it also significantly increases HSEG s computational requirements. To counteract this, a computationally efficient recursive, divide-and-conquer, implementation of HSEG (RHSEG) was devised, which includes special code to avoid processing artifacts caused by RHSEG s recursive subdivision of the image data. The recursive nature of RHSEG makes for a straightforward parallel implementation. This paper describes the HSEG algorithm, its recursive formulation (referred to as RHSEG), and the implementation of RHSEG using massively parallel GNU-LINUX software. Results with Landsat TM data are included comparing RHSEG with classic region growing.

Tilton, James C.↗
NASA NTRS1992

Parallel solver for trajectory optimization search directions

A key algorithmic element of a real-time trajectory optimization hardware/software implementation is presented, the search step solver. This is one piece of an algorithm whose overall goal is to make nonlinear trajectory optimization fast enough to provide real-time commands during guidance of a vehicle such as an aeromaneuvering orbiter or the National Aerospace Plane. Many methods of nonlinear programming require the solution of a quadratic program (QP) at each iteration to determine the search step. In the trajectory optimization case, the QP has a special dynamic programming structure. The algorithm exploits this special structure with a divide- and conquer type of parallel implementation. The algorithm solves a (p.N)-stage problem on N processors in O(p + log2 N) operations. The algorithm yields a factor of 8 speed-up over the fastest known serial algorithm when solving a 1024-stage test problem on 32 processors.

Psiaki, M. L.↗
DOE OSTI2025

TorchBraid: High-Performance Layer-Parallel Training of Deep Neural Networks with MPI and GPU Acceleration

TorchBraid is a high-performance implementation of layer-parallel training for deep neural networks (DNNs) supporting MPI-based parallelism and GPU acceleration. Layer-parallel training has been developed to overcome the serialization inherent in forward and backward propagation of DNNs that limits utilization of computational resources in the strong scaling limit. To achieve this, TorchBraid integrates the PyTorch neural network framework with the state-of-the-art XBraid time-parallel library. Furthermore, this article presents the use and performance of TorchBraid, in addition to solutions for overcoming the algorithmic challenges inherent in combining automatic differentiation with layer-parallel. Results are presented with and without GPU acceleration for the Tiny ImageNet and MNIST image classification data sets, as well as recurrent neural networks. Overall, TorchBraid enables fast training of DNNs, both in a strong and weak scaling context. In addition to the TorchBraid software, several new advances in applying layer-parallel algorithms are detailed. Integration of layer-parallel with data-parallel algorithms is presented for the first time, showing the computational advantages of the combination. Standard deep learning techniques, like batch-normalization, are developed for layer-parallel training. Finally, a new approach combining layer-parallel with spatial coarsening in order to accelerate training for 3D image classification shows roughly a 10× speedup over serial execution.

Layer-parallel↗
DOE OSTI2022

ppdx : Automated modeling of protein–protein interaction descriptors for use with machine learning

This paper describes ppdx, a python workflow tool that combines protein sequence alignment, homology modeling, and structural refinement, to compute a broad array of descriptors for characterizing protein–protein interactions. The descriptors can be used to predict various properties of interest, such as protein–protein binding affinities, or inhibitory concentrations (IC 50 ), using approaches that range from simple regression to more complex machine learning models. The software is highly modular. It supports different protocols for generating structures, and 95 descriptors can be currently computed. More protocols and descriptors can be easily added. The implementation is highly parallel and can fully exploit the available cores in a single workstation, or multiple nodes on a supercomputer, allowing many systems to be analyzed simultaneously. As an illustrative application, ppdx is used to parametrize a model that predicts the IC 50 of a set of antigens and a class of antibodies directed to the influenza hemagglutinin stalk.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Computing the QRPA level density with the finite amplitude method

Here, we describe a new algorithm to calculate the vibrational nuclear level density of an atomic nucleus. Fictitious perturbation operators that probe the response of the system are generated by drawing their matrix elements from some probability distribution function. We use the Finite Amplitude Method to explicitly compute the response for each such sample. With the help of the Kernel Polynomial Method, we build an estimator of the vibrational level density and provide the upper bound of the relative error in the limit of infinitely many random samples. The new algorithm can give accurate estimates of the vibrational level density. Since it is based on drawing multiple samples of perturbation operators, its computational implementation is naturally parallel and scales like the number of available processing units.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2019

Analytic band-to-trap tunneling model including band offset for heterojunction devices

In this paper, we present an analytic band-to-trap tunneling model based on the open boundary scattering approach. The new model has three major advantages: (i) It includes not only the well-known electric field effect, but more importantly, the effect of heterojunction band offset. This feature allows us to simulate both electric field and band offset enhanced carrier recombination near a heterojunction in heterostructures. (ii) Its analytic form enables straightforward implementation into a parallel Technology Computer Aided Design device and circuit simulators. (iii) The developed method can be used for any potentials which can be approximated to a good degree such that the Schrödinger equation with open boundary conditions results in piecewise analytic wave functions. Simulation results of an InGaP/GaAs heterojunction bipolar transistor (HBT) reveal that the proposed model predicts significantly increased base currents, because the tunneling of holes in the base to traps in the emitter is greatly enhanced by the emitter-base band offset. Lastly, this finding, which is not captured by existing band-to-trap tunneling models, is consistent with the experimental observation for an InGaP/GaAs HBT after neutron irradiation.

36 MATERIALS SCIENCE↗