Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallelism”

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 937 records · Page 52

Maximal clique enumeration with data-parallel primitives

The enumeration of all maximal cliques in an undirected graph is a fundamental problem arising in several research areas. We consider maximal clique enumeration on shared-memory, multi-core architectures and introduce an approach consisting entirely of data-parallel operations, in an effort to achieve efficient and portable performance across different architectures. We study the performance of the algorithm via experiments varying over benchmark graphs and architectures. Overall, we observe that our algorithm achieves up to a 33-time speedup and 9-time speedup over state-of-the-art distributed and serial algorithms, respectively, for graphs with higher ratios of maximal cliques to total cliques. Further, we attain additional speedups on a GPU architecture, demonstrating the portable performance of our data-parallel design.

97 MATHEMATICS AND COMPUTING↗

PyOMP: Multithreaded Parallel Programming in Python

We know that Python is a widely used language in scientific computing. When the goal is high performance, however, Python lags far behind low-level languages such as C and Fortran. To support applications that stress performance, Python needs to access the full capabilities of modern CPUs. That means support for parallel multithreading. In this paper, we describe PyOMP, a system that enables OpenMP in Python. Programmers write code in Python with OpenMP, Numba generates code that compiles to LLVM, and the resulting programs run with performance that approaches that from code written with C and OpenMP. In this paper we provide an update on the PyOMP project and explain how to install it and use it to write parallel multithreaded code in Python.

97 MATHEMATICS AND COMPUTING↗

Parallel Evolutionary Optimization for Neuromorphic Network Training

One of the key impediments to the success of current neuromorphic computing architectures is the issue of how best to program them. Evolutionary optimization (EO) is one promising programming technique; in particular, its wide applicability makes it especially attractive for neuromorphic architectures, which can have many different characteristics. In this paper, we explore different facets of EO on a spiking neuromorphic computing model called DANNA. We focus on the performance of EO in the design of our DANNA simulator, and on how to structure EO on both multicore and massively parallel computing systems. We evaluate how our parallel methods impact the performance of EO on Titan, the U.S.'s largest open science supercomputer, and BOB, a Beowulf-style cluster of Raspberry Pi's. We also focus on how to improve the EO by evaluating commonality in higher performing neural networks, and present the result of a study that evaluates the EO performed by Titan.

97 MATHEMATICS AND COMPUTING↗

Parallel Implementation of Power System Dynamic Simulation

Dynamic simulation of power system transient stability is important for planning, monitoring, operation, and control of electrical power systems. However, modeling the system dynamics and network involves the computationally intensive time-domain solution of numerous differential and algebraic equations (DAE). This results in a transient stability implementation that may not maintain the real-time constraints of an online security assessment. This paper presents a parallel implementation of the dynamic simulation on a high-performance computing (HPC) platform using parallel simulation algorithms and computation architectures. It enables the simulation to run even faster than real time, enabling the “look-ahead” capability of upcoming stability problems in the power grid.

dynamic simulation↗

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↗

Image Gradient Decomposition for Parallel and Memory-Efficient Ptychographic Reconstruction

Ptychography is a popular microscopic imaging modality for many scientific discoveries and sets the record for highest image resolution. Unfortunately, the high image resolution for ptychographic reconstruction requires significant amount of memory and computations, forcing many applications to compromise their image resolution in exchange for a smaller memory footprint and a shorter reconstruction time. In this paper, we propose a novel image gradient decomposition method that significantly reduces the memory footprint for ptychographic reconstruction by tessellating image gradients and diffraction measurements into tiles. In addition, we propose a parallel image gradient decomposition method that enables asynchronous point-to-point communications and parallel pipelining with minimal overhead on a large number of GPUs. Our experiments on a Titanate material dataset (PbTiO3) with 16632 probe locations show that our Gradient Decomposition algorithm reduces memory footprint by 51 times. In addition, it achieves time-to-solution within 2.2 minutes by scaling to 4158 GPUs with a super-linear strong scaling efficiency at 364% compared to runtimes at 6 GPUs. This performance is 2.7 times more memory efficient, 9 times more scalable and 86 times faster than the state-of-the-art algorithm.

Wang, Xiao↗

Distributed-Memory Parallel Symmetric Nonnegative Matrix Factorization

We develop the first distributed-memory parallel implementation of Symmetric Nonnegative Matrix Factorization (SymNMF), a key data analytics kernel for clustering and dimensionality reduction. Our implementation includes two different algorithms for SymNMF, which give comparable results in terms of time and accuracy. The first algorithm is a parallelization of an existing sequential approach that uses solvers for non symmetric NMF. The second algorithm is a novel approach based on the Gauss-Newton method. It exploits second-order information without incurring large computational and memory costs. We evaluate the scalability of our algorithms on the Summit system at Oak Ridge National Laboratory, scaling up to 128 nodes (4,096 cores) with 70% efficiency. Additionally, we demonstrate our software on an image segmentation task.

Eswar, Srinivas↗

Radiological Source Term Estimation and Isotopic Identification with Parallel Log Domain Particle Filters

This paper presents a parallel log-domain particle filtering algorithm combined with gamma spectrum unfolding to perform localization, identification, and evaluation of multiple point sources of various isotopes in an environment with attenuating obstacles. The method uses sets of precomputed attenuation kernels that map the attenuation characteristics of the environment. These kernels are specific to the energy level of a photopeak of interest. The spectral measurements are deconvolved into count measurements of each photopeak. These count measurements are fed into a set of parallel particle filters using attenuation kernels computed for that photopeak’s energy level. The individual regularized particle filters perform all likelihood calculations in the logarithmic domain to mitigate the effects of particle degeneracy. The output of each particle filter is combined to estimate which isotopes are present as well as their positions and strengths. The performance of the algorithm is characterized in a lab-scale environment using a mobile robot equipped with a gamma ray spectrometer in the presence of up to three different radioactive isotopes simultaneously. The sources were localized to within 10 cm, and their strengths were estimated within 10% of their true values. Furthermore, the isotopes were all correctly identified, and no spurious sources were reported.

42 ENGINEERING↗

Common-Mode Circulating Current Control of Paralleled Interleaved Three-Phase Two-Level Voltage-Source Converters With Discontinuous Space-Vector Modulation

This paper presents a control method to limit the common-mode (CM) circulating current between paralleled three-phase two-level voltage-source converters (VSCs) with discontinuous space-vector pulsewidth modulation (DPWM) and interleaved switching cycles. This CM circulating current can be separated into two separate components based on their frequency; the high-frequency component, close to the switching frequency, can be effectively limited by means of passive components; the low-frequency component, close to the fundamental frequency, embodies the jumping CM circulating current observed in parallel VSCs. This is the main reason why it is usually recommended not to implement discontinuous and interleaving PWM together. The origin of this low-frequency circulating current is analyzed in detail, and based on this, a method to eliminate its presence is proposed by impeding the simultaneous use of different zero vectors between the converters. This control method only requires six additional switching actions per line cycle, presenting a minimum impact on the converter thermal design. The analysis and the feasibility of the control method are verified by simulation and experimental results.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Identifying Redundant Flow Limits on Parallel Lines

Many power system optimization problems constrain line flows with limits specified in terms of the magnitudes of apparent power or current flows. The set of line-flow constraints may have redundancies (i.e., the feasible space may be unchanged upon removal of a subset of the line-flow constraints), which unnecessarily complicate optimization problems. This letter describes an algorithm for identifying redundant line-flow constraints corresponding to certain parallel lines. After formulating the constraints as ellipsoids in the voltage variables, redundancies are detected from the absence of intersections between pairs of ellipsoids corresponding to parallel lines. This algorithm is demonstrated using several large test cases.

Molzahn, Daniel K. (ORCID:0000000305835376)↗

Fast Parallel Stochastic Subspace Algorithms for Large-Scale Ambient Oscillation Monitoring

With the installation of synchrophasors widely across the power grid, measurement-based oscillation monitoring algorithms are becoming increasingly useful in identifying the real-time oscillatory modal properties in power systems. When the number of phasor measurement unit (PMU) channels grows, the computational time of many PMU data based algorithms is dominated by the computational burden in processing large-scale dense matrices. In order to overcome this limitation, this paper presents new formulations and computational strategies for speeding up an ambient oscillation monitoring algorithm, namely, stochastic subspace identification (SSI). Based on previous work, two fast singular value decomposition (SVD) approaches are first applied to the SVD evaluation within the SSI algorithm. Next, block structures are exploited so that the large-scale dense matrix computations can be processed in parallel. This helps in memory savings as well as in overall computational time. Experimental results from three sets of archived data of the western interconnection demonstrate that the new approaches can provide significant speedups while retaining modal estimation accuracy. With proposed fast parallel algorithms, the real-time oscillation monitoring of the large-scale system using hundreds of PMU measurements becomes feasible.

97 MATHEMATICS AND COMPUTING↗

Scalable parallel distance field construction for large-scale applications

Computing distance fields is fundamental to many scientific and engineering applications. Distance fields can be used to direct analysis and reduce data. In this paper, we present a highly scalable method for computing 3D distance fields on massively parallel distributed-memory machines. Anew distributed spatial data structure, named parallel distance tree, is introduced to manage the level sets of data and facilitate surface tracking overtime, resulting in significantly reduced computation and communication costs for calculating the distance to the surface of interest from any spatial locations. Our method supports several data types and distance metrics from real-world applications. We demonstrate its efficiency and scalability on state-of-the-art supercomputers using both large-scale volume datasets and surface models. We also demonstrate in-situ distance field computation on dynamic turbulent flame surfaces for a petascale combustion simulation. In conclusion, our work greatly extends the usability of distance fields for demanding applications.

97 MATHEMATICS AND COMPUTING↗

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↗

Experimental and Numerical Analysis of Additively Manufactured Coupons With Parallel Channels and Inline Wall Jets

Present paper developed a series of additively manufactured parallel cooling channels with streamwise wall jets inside, which could be suitable for double-wall and near-wall cooling configurations in gas turbine hot section components. The tested coupons consisted of parallel channels, each channel further divided into small chambers using several spanwise separation walls. Height of these walls was kept less than channel height, thus forming a slot with one of the end walls. Coolant entered from one side of channel and formed streamwise wall jet while crossing through the slot over to the downstream chamber. The test coupons were additively manufactured by selective laser sintering (SLS) technique using Inconel 718 alloy. Steady-state heat transfer experiments with constant wall temperature boundary condition were performed to analyze effect of pitch between subsequent slots and blockage ratio (ratio of separation wall height to channel height) on heat transfer. The channel Reynolds number ranged from 1800 to 5000. Numerical simulations were performed using ANSYS cfx solver with SST k–ω turbulence model to obtain detailed understanding of existing flow field. Experimental results showed heat transfer enhancement of up to 6.5 times that of a smooth channel for the highest blockage ratio of 0.75. Numerical results revealed complex flow field which consisted of wall jets along with impingement, separation, and recirculation zones in each chamber. For all configurations, gain in heat transfer was accompanied with high pressure drops. However, coupled with the high heat transfer, this design could lead to potential coolant savings.

Engineering↗

Hardware-in-the-Loop Investigation of Emissions Challenges in Hybrid Medium- and Heavy-Duty Powertrains Using a Pre-Production Diesel-Electric Parallel Hybrid System With and Without Stop-Start Operation

Hybrid electric powertrains are a growing market in medium- and heavy-duty applications. There is a lack of available information to understand the challenges in the integration of engine platforms into electrified powertrains, such as cold-start, restart, and load-reduction effects on emissions and emission control devices. Results from the Heavy Heavy-Duty Diesel Truck (HHDDT) cycle using a conventional medium-duty diesel engine were compared with those of a parallel hybrid architecture. Oak Ridge National Laboratory in collaboration with the US Department of Energy and Odyne Systems, LLC developed a powertrain in a hardware-in-the-loop environment, integrating the Odyne Systems, LLC medium-duty parallel hybrid system, which was used for the hybrid portion of this study. Experiments under the HHDDT cycle showed increasing improvements in fuel consumption and engine-out emissions with the integration of stop/start, hybrid, and hybrid with stop/start. However, the effects of load reduction and exhaust temperature on the thermal management strategy have shown an increase in fueling in the second part of the HHDDT cycle. Four configurations of medium-duty electrification were studied and contributed to building a unique data set containing combustion, emissions, and system integration data. Each electrification level was compared with the conventional baseline. The calibration of the conventional engine was not altered for this study. Opportunities to tailor the combustion process were identified with the stop/start strategy.

Lerin, Chloe↗

Extraction of depth-dependent perturbation factors for parallel-plate chambers in electron beams using a plastic scintillation detector

Purpose: This work presents the experimental extraction of the overall perturbation factor P{sub Q} in megavoltage electron beams for NACP-02 and Roos parallel-plate ionization chambers using a plastic scintillation detector (PSD). Methods: The authors used a single scanning PSD mounted on a high-precision scanning tank to measure depth-dose curves in 6, 12, and 18 MeV clinical electron beams. The authors also measured depth-dose curves using the NACP-02 and PTW Roos chambers. Results: The authors found that the perturbation factors for the NACP-02 and Roos chambers increased substantially with depth, especially for low-energy electron beams. The experimental results were in good agreement with the results of Monte Carlo simulations reported by other investigators. The authors also found that using an effective point of measurement (EPOM) placed inside the air cavity reduced the variation of perturbation factors with depth and that the optimal EPOM appears to be energy dependent. Conclusions: A PSD can be used to experimentally extract perturbation factors for ionization chambers. The dosimetry protocol recommendations indicating that the point of measurement be placed on the inside face of the front window appear to be incorrect for parallel-plate chambers and result in errors in the R{sub 50} of approximately 0.4 mm at 6 MeV, 1.0 mm at 12 MeV, and 1.2 mm at 18 MeV.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

On the wall perturbation correction for a parallel-plate NACP-02 chamber in clinical electron beams

Purpose: In recent years, several Monte Carlo studies have been published concerning the perturbation corrections of a parallel-plate chamber in clinical electron beams. In these studies, a strong depth dependence of the relevant correction factors (p{sub wall} and p{sub cav}) for depth beyond the reference depth is recognized and it has been shown that the variation with depth is sensitive to the choice of the chamber's effective point of measurement. Recommendations concerning the positioning of parallel-plate ionization chambers in clinical electron beams are not the same for all current dosimetry protocols. The IAEA TRS-398 as well as the IPEM protocol and the German protocol DIN 6800-2 interpret the depth of measurement within the phantom as the water equivalent depth, i.e., the nonwater equivalence of the entrance window has to be accounted for by shifting the chamber by an amount {Delta}z. This positioning should ensure that the primary electrons traveling from the surface of the water phantom through the entrance window to the chamber's reference point sustain the same energy loss as the primary electrons in the undisturbed phantom. The objective of the present study is the determination of the shift {Delta}z for a NACP-02 chamber and the calculation of the resulting wall perturbation correction as a function of depth. Moreover, the contributions of the different chamber walls to the wall perturbation correction are identified. Methods: The dose and fluence within the NACP-02 chamber and a wall-less air cavity is calculated using the Monte Carlo code EGSnrc in a water phantom at different depths for different clinical electron beams. In order to determine the necessary shift to account for the nonwater equivalence of the entrance window, the chamber is shifted in steps {Delta}z around the depth of measurement. The optimal shift {Delta}z is determined from a comparison of the spectral fluence within the chamber and the bare cavity. The wall perturbation correction is calculated as the ratio between doses for the complete chamber and a wall-less air cavity. Results: The high energy part of the fluence spectra within the chamber strongly varies even with small chamber shifts, allowing the determination of {Delta}z within micrometers. For the NACP-02 chamber a shift {Delta}z=-0.058 cm results. This value is independent of the energy of the primary electrons as well as of the depth within the phantom and it is in good agreement with the value recommended in the German dosimetry protocol. Applying this shift, the calculated wall perturbation correction as a function of depth is varying less than 1% from zero up to the half value depth R{sub 50} for electron energies in the range of 6-21 MeV. The remaining depth dependence can mainly be attributed to the scatter properties of the entrance window. When neglecting the nonwater equivalence of the entrance window, the variation of p{sub wall} with depth is up to 10% and more, especially for low electron energies. Conclusions: The variation of the wall perturbation correction for the NACP-02 chamber in clinical electron beams strongly depends on the positioning of the chamber. Applying a shift {Delta}z=-0.058 cm toward the focus ensures that the primary electron spectrum within the chamber bears the largest resemblance to the fluence of a wall-less cavity. Hence, the influence of the chamber walls on the perturbation correction can be separated out and the residual variation of p{sub wall} with depth is minimized.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

SU-E-T-448: On the Perturbation Factor P-cav of the Markus Parallel Plate Ion Chambers in Clinical Electron Beams, Monte Carlo Based Reintegration of An Historical Experiment

Purpose: All present dosimetry protocols recommend well-guarded parallel-plate ion chambers for electron dosimetry. For the guard-less Markus chamber an energy dependent fluence perturbation correction pcav is given. This perturbation correction was experimentally determined by van der Plaetsen by comparison of the read-out of a Markus and a NACP chamber, which was assumed to be “perturbation-free”. Aim of the present study is a Monte Carlo based reiteration of this experiment. Methods: Detailed models of four parallel-plate chambers (Roos, Markus, NACP and Advanced Markus) were designed using the Monte Carlo code EGSnrc and placed in a water phantom. For all chambers the dose to the active volume filled with low density water was calculated for 13 clinical electron spectra (E{sub 0}=6-21 MeV) at the depth of maximum and at the reference depth under reference conditions. In all cases the chamber's reference point was positioned at the depth of measurement. Moreover, the dose to water DW was calculated in a small water voxel positioned at the same depth. Results: The calculated dose ratio D{sub NACP}/D{sub Markus}, which according to van der Plaetsen reflects the fluence perturbation correction of the Markus chamber, deviates less from unity than the values given by van der Plaetsen's but exhibits a similar energy dependence. The same holds for the dose ratios of the other well guarded chambers. But, in comparison to water, the Markus chamber reveals the smallest overall perturbation correction which is nearly energy independent at both investigated depths. Conclusion: The simulations principally confirm the energy dependence of the dose ratio D{sub NACP}/D{sub Markus} as published by van der Plaetsen. But, as shown by our simulations of the ratio D{sub W}/D{sub Markus}, the conclusion drawn in all dosimetry protocols is questionable: in contrast to all well-guarded chambers the guard-less Markus chamber reveals the smallest overall perturbation correction and also the smallest energy dependence.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗