Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallel algorithms”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,747 records · Page 97

Quantum Dynamics at Scale: Ultrafast Control of Emergent Functional Materials

Confluence of extreme-scale quantum dynamics simulations (i.e. quantum@scale) and cutting-edge x-ray free-electron laser experiments are revolutionizing materials science. An archetypal example is the exciting concept of using picosecond light pulses to control emergent material properties on demand in atomically-thin layered materials. This paper describes efforts to scale our quantum molecular dynamics engine toward the United States' first exaflop/s computer, under an Aurora Early Science Program project named "Metascalable layered material genome". Key algorithmic and computing techniques incorporated are: (1) globally-scalable and locally-fast solvers within a linear-scaling divide-conquer-recombine algorithmic framework; (2) algebraic `BLASification' of computational kernels; and (3) data alignment and loop restructuring, along with register and cache blocking, for enhanced vectorization and efficient memory access. The resulting weak-scaling parallel efficiency was 0.93 on 131,072 Intel Xeon Phi cores for a 56.6 million atom (or 169 million valence-electron) system, whereas the various code transformations achieved 5-fold speedup. The optimized simulation engine allowed us for the first time to establish a significant effect of substrate on the dynamics of layered material upon electronic excitation.

Algebraic BLASification↗

Scalable Cloning on Large-Scale GPU Platforms with Application to Time-Stepped Simulations on Grids

Cloning is a technique to efficiently simulate a tree of multiple what-if scenarios that are unraveled during the course of a base simulation. However, cloned execution is highly challenging to realize on large, distributed memory computing platforms, due to the dynamic nature of the computational load across clones, and due to the complex dependencies spanning the clone tree. In this paper, we present the conceptual simulation framework, algorithmic foundations, and runtime interface of CloneX, a new system we designed for scalable simulation cloning. It efficiently and dynamically creates whole logical copies of a dynamic tree of simulations across a large parallel system without full physical duplication of computation and memory. The performance of a prototype implementation executed on up to 1,024 graphical processing units of a supercomputing system has been evaluated with three benchmarks—heat diffusion, forest fire, and disease propagation models—delivering a speed up of over two orders of magnitude compared to replicated runs. Finally, the results demonstrate a significantly faster and scalable way to execute many what-if scenario ensembles of large simulations via cloning using the CloneX interface.

97 MATHEMATICS AND COMPUTING↗

A RADIATION TRANSFER SOLVER FOR ATHENA USING SHORT CHARACTERISTICS

We describe the implementation of a module for the Athena magnetohydrodynamics (MHD) code that solves the time-independent, multi-frequency radiative transfer (RT) equation on multidimensional Cartesian simulation domains, including scattering and non-local thermodynamic equilibrium (LTE) effects. The module is based on well known and well tested algorithms developed for modeling stellar atmospheres, including the method of short characteristics to solve the RT equation, accelerated Lambda iteration to handle scattering and non-LTE effects, and parallelization via domain decomposition. The module serves several purposes: it can be used to generate spectra and images, to compute a variable Eddington tensor (VET) for full radiation MHD simulations, and to calculate the heating and cooling source terms in the MHD equations in flows where radiation pressure is small compared with gas pressure. For the latter case, the module is combined with the standard MHD integrators using operator splitting: we describe this approach in detail, including a new constraint on the time step for stability due to radiation diffusion modes. Implementation of the VET method for radiation pressure dominated flows is described in a companion paper. We present results from a suite of test problems for both the RT solver itself and for dynamical problems that include radiative heating and cooling. These tests demonstrate that the radiative transfer solution is accurate and confirm that the operator split method is stable, convergent, and efficient for problems of interest. We demonstrate there is no need to adopt ad hoc assumptions of questionable accuracy to solve RT problems in concert with MHD: the computational cost for our general-purpose module for simple (e.g., LTE gray) problems can be comparable to or less than a single time step of Athena's MHD integrators, and only few times more expensive than that for more general (non-LTE) problems.

79 ASTRONOMY AND ASTROPHYSICS↗

A Scalable Gaussian Process Analysis Algorithm for Biomass Monitoring

Biomass monitoring is vital for studying the carbon cycle of earth's ecosystem and has several significant implications, especially in the context of understanding climate change and its impacts. Recently, several change detection methods have been proposed to identify land cover changes in temporal profiles (time series) of vegetation collected using remote sensing instruments, but do not satisfy one or both of the two requirements of the biomass monitoring problem, i.e., {\em operating in online mode} and {\em handling periodic time series}. In this paper, we adapt Gaussian process regression to detect changes in such time series in an online fashion. While Gaussian process (GP) have been widely used as a kernel based learning method for regression and classification, their applicability to massive spatio-temporal data sets, such as remote sensing data, has been limited owing to the high computational costs involved. We focus on addressing the scalability issues associated with the proposed GP based change detection algorithm. This paper makes several significant contributions. First, we propose a GP based online time series change detection algorithm and demonstrate its effectiveness in detecting different types of changes in {\em Normalized Difference Vegetation Index} (NDVI) data obtained from a study area in Iowa, USA. Second, we propose an efficient Toeplitz matrix based solution which significantly improves the computational complexity and memory requirements of the proposed GP based method. Specifically, the proposed solution can analyze a time series of length $t$ in $O(t^2)$ time while maintaining a $O(t)$ memory footprint, compared to the $O(t^3)$ time and $O(t^2)$ memory requirement of standard matrix manipulation based methods. Third, we describe a parallel version of the proposed solution which can be used to simultaneously analyze a large number of time series. We study three different parallel implementations: using threads, MPI, and a hybrid implementation using threads and MPI. Experimental results show that the hybrid implementation scales better than the multi-threaded and MPI based implementations. The application of the proposed scalable algorithm is demonstrated in analyzing massive remote sensing observation data. The hybrid implementation, using 1536 computing cores, can analyze an NDVI data set for the Iowa study area in nearly 5 seconds, while a serial algorithm, using standard Cholesky decomposition based routines, takes several days to process the same data set.

09 BIOMASS FUELS↗

Efficient Synthesis of Graph Methods: a Dynamically Scheduled Architecture

RDF databases naturally map to a graph representation and employ languages, such as SPARQL, that implements queries as graph pattern matching routines. Graph methods exhibit an irregular behavior: they present unpredictable, fine-grained data accesses, and are synchronization inten- sive. Graph data structures expose large amounts of dy- namic parallelism, but are difficult to partition without gen- erating load unbalance. In this paper, we present a novel ar- chitecture to improve the synthesis of graph methods. Our design addresses the issues of these algorithms with two com- ponents: a Dynamic Task Scheduler (DTS), which reduces load unbalance and maximize resource utilization, and a Hi- erarchical Memory Interface controller (HMI), which pro- vides support for concurrent memory operations on multi- ported/multi-banked shared memories. We evaluate our ap- proach by generating the accelerators for a set of SPARQL queries from the Lehigh University Benchmark (LUBM). We first analyze the load unbalance of these queries, showing that execution time among tasks can differ even of order of magnitudes. We then synthesize the queries and com- pare the performance of the resulting accelerators against the current state of the art. Experimental results show that our solution provides a speedup over the serial implementa- tion close to the theoretical maximum and a speedup up to 3.45 over a baseline parallel implementation. We conclude our study by exploring the design space to achieve maximum memory channels utilization. The best design used at least three of the four memory channels for more than 90% of the execution time.

architecture↗

Verification of a Configuration Space Method for Evaluating the All-Orders Linear Kinetic Plasma Response to RF Power

Progress towards verification of a configuration space algorithm for evaluating the linear kinetic plasma current response to applied radio-frequency power is presented. This follows a previously published [1] verification of the parallel-only response, by verifying the perpendicular response required for much of the ion-and electron-cyclotron range of frequencies. In particular, we use the Colestock-Kashuba fast-wave to ion-Bernstein-wave benchmark [2] where we show that the Kinetic-J computational kernel is able to reproduce a known reference solution calculated via the AORSA full-wave code [3]. We also discuss the computational and scaling aspects of this novel approach, as well as the difficulties presented by the temporal scales in this multi-species problem.

Green, David↗

High speed two-dimensional imaging with an analog interface

A quantitative pulse count (event detection) algorithm with linearity to high count rates is accomplished by combining a high-speed, high frame rate camera with simple logic code run on a massively parallel processor such as a GPU. The parallel processor elements examine frames from the camera pixel by pixel to find and tag events or count pulses. The tagged events are combined to form a combined quantitative event image.

Waugh, Justin↗

Portability: A Necessary Approach for Future Scientific Software

Today's world of scientific software for High Energy Physics (HEP) is powered by x86 code, while the future will be much more reliant on accelerators like GPUs and FPGAs. The portable parallelization strategies (PPS) project of the High Energy Physics Center for Computational Excellence (HEP/CCE) is investigating solutions for portability techniques that will allow the coding of an algorithm once, and the ability to execute it on a variety of hardware products from many vendors, especially including accelerators. We think without these solutions, the scientific success of our experiments and endeavors is in danger, as software development could be expert driven and costly to be able to run on available hardware infrastructure. We think the best solution for the community would be an extension to the C++ standard with a very low entry bar for users, supporting all hardware forms and vendors. We are very far from that ideal though. We argue that in the future, as a community, we need to request and work on portability solutions and strive to reach this ideal.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Improving scalability of parallel CNN training by adaptively adjusting parameter update frequency

Synchronous SGD with data parallelism, the most popular parallelization strategy for CNN training, suffers from the expensive communication cost of averaging gradients among all workers. The iterative parameter updates of SGD cause frequent communications and it becomes the performance bottleneck. In this paper, we propose a lazy parameter update algorithm that adaptively adjusts the parameter update frequency to address the expensive communication cost issue. Our algorithm accumulates the gradients if the difference of the accumulated gradients and the latest gradients is sufficiently small. Here, the less frequent parameter updates reduce the per-iteration communication cost while maintaining the model accuracy. Our experimental results demonstrate that the lazy update method remarkably improves the scalability while maintaining the model accuracy. For ResNet50 training on ImageNet, the proposed algorithm achieves a significantly higher speedup (739.6 on 2048 Cori KNL nodes) as compared to the vanilla synchronous SGD (276.6) while the model accuracy is almost not affected (<0.2% difference).

97 MATHEMATICS AND COMPUTING↗

An MCNPX Monte Carlo model of a discrete spot scanning proton beam therapy nozzle

Purpose: The purposes of this study were to validate a discrete spot scanning proton beam nozzle using the Monte Carlo (MC) code MCNPX and use the MC validated model to investigate the effects of a low-dose envelope, which surrounds the beam's central axis, on measurements of integral depth dose (IDD) profiles. Methods: An accurate model of the discrete spot scanning beam nozzle from The University of Texas M. D. Anderson Cancer Center (Houston, Texas) was developed on the basis of blueprints provided by the manufacturer of the nozzle. The authors performed simulations of single proton pencil beams of various energies using the standard multiple Coulomb scattering (MCS) algorithm within the MCNPX source code and a new MCS algorithm, which was implemented in the MCNPX source code. The MC models were validated by comparing calculated in-air and in-water lateral profiles and percentage depth dose profiles for single pencil beams with their corresponding measured values. The models were then further tested by comparing the calculated and measured three-dimensional (3-D) dose distributions. Finally, an IDD profile was calculated with different scoring radii to determine the limitations on the use of commercially available plane-parallel ionization chambers to measure IDD. Results: The distance to agreement, defined as the distance between the nearest positions of two equivalent distributions with the same value of dose, between measured and simulated ranges was within 0.13 cm for both MCS algorithms. For low and intermediate pencil beam energies, the MC simulations using the standard MCS algorithm were in better agreement with measurements. Conversely, the new MCS algorithm produced better results for high-energy single pencil beams. The IDD profile calculated with cylindrical tallies with an area equivalent to the area of the largest commercially available ionization chamber showed up to 7.8% underestimation of the integral dose in certain depths of the IDD profile. Conclusions: The authors conclude that a combination of MCS algorithms is required to accurately reproduce experimental data of single pencil beams and 3-D dose distributions for the scanning beam nozzle. In addition, the MC simulations showed that because of the low-dose envelope, ionization chambers with radii as large as 4.08 cm are insufficient to accurately measure IDD profiles for a 221.8 MeV pencil beam in the scanning beam nozzle.

60 APPLIED LIFE SCIENCES↗

Exploiting Modern C++ for Portable Parallel Programming in Lattice QCD Applications

The evolution of ISO C++ standards increasingly serves the needs of scientific computing, offering potential benefits for developing portable applications. The recent revisions of C++ programming language, for instance, introduces a suite of algorithms capable of being executed on accelerators. Although this approach may not yield best performance, it can present a viable balance between code productivity and computational efficiency. In this report, we discuss the implementation of the HISQ operator utilizing a range of features from the C++17/20/23 standards and include an assessment of their performance.

Strelchenko, Alexei↗

Implementing Molecular Dynamics for Hybrid High Performance Computers - 1. Short Range Forces

The use of accelerators such as general-purpose graphics processing units (GPGPUs) have become popular in scientific computing applications due to their low cost, impressive floating-point capabilities, high memory bandwidth, and low electrical power requirements. Hybrid high performance computers, machines with more than one type of floating-point processor, are now becoming more prevalent due to these advantages. In this work, we discuss several important issues in porting a large molecular dynamics code for use on parallel hybrid machines - 1) choosing a hybrid parallel decomposition that works on central processing units (CPUs) with distributed memory and accelerator cores with shared memory, 2) minimizing the amount of code that must be ported for efficient acceleration, 3) utilizing the available processing power from both many-core CPUs and accelerators, and 4) choosing a programming model for acceleration. We present our solution to each of these issues for short-range force calculation in the molecular dynamics package LAMMPS. We describe algorithms for efficient short range force calculation on hybrid high performance machines. We describe a new approach for dynamic load balancing of work between CPU and accelerator cores. We describe the Geryon library that allows a single code to compile with both CUDA and OpenCL for use on a variety of accelerators. Finally, we present results on a parallel test cluster containing 32 Fermi GPGPUs and 180 CPU cores.

43 PARTICLE ACCELERATORS↗

Parallel transport sweeps on two-dimensional cartesian and hexagonal grids

This paper aims to provide a proof of concept for parallel transport sweeps on two-dimensional hexagonal grids for the discrete ordinates transport equation. While the method is an extension of the popular and well-established Koch-Baker-Alcoulffe (KBA) algorithm, there are significant differences between the cartesian and hexagonal grid and thereafter sweep. The most important is the three-way connectivity of hexagons within the grid which creates greater dependencies between the elements. The KBA method in structured orthogonal grids was first implemented in the DRAGON5 code and the method is first described here. The differences in implementation for the hexagonal grid are also described. Benchmark results are also presented, showing roughly 10 times speedup in computational times with roughly 100 processors, in both cases. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

TH-AB-207A-05: A Fully-Automated Pipeline for Generating CT Images Across a Range of Doses and Reconstruction Methods

Purpose: To evaluate the robustness of CAD or Quantitative Imaging methods, they should be tested on a variety of cases and under a variety of image acquisition and reconstruction conditions that represent the heterogeneity encountered in clinical practice. The purpose of this work was to develop a fully-automated pipeline for generating CT images that represent a wide range of dose and reconstruction conditions. Methods: The pipeline consists of three main modules: reduced-dose simulation, image reconstruction, and quantitative analysis. The first two modules of the pipeline can be operated in a completely automated fashion, using configuration files and running the modules in a batch queue. The input to the pipeline is raw projection CT data; this data is used to simulate different levels of dose reduction using a previously-published algorithm. Filtered-backprojection reconstructions are then performed using FreeCT-wFBP, a freely-available reconstruction software for helical CT. We also added support for an in-house, model-based iterative reconstruction algorithm using iterative coordinate-descent optimization, which may be run in tandem with the more conventional recon methods. The reduced-dose simulations and image reconstructions are controlled automatically by a single script, and they can be run in parallel on our research cluster. The pipeline was tested on phantom and lung screening datasets from a clinical scanner (Definition AS, Siemens Healthcare). Results: The images generated from our test datasets appeared to represent a realistic range of acquisition and reconstruction conditions that we would expect to find clinically. The time to generate images was approximately 30 minutes per dose/reconstruction combination on a hybrid CPU/GPU architecture. Conclusion: The automated research pipeline promises to be a useful tool for either training or evaluating performance of quantitative imaging software such as classifiers and CAD algorithms across the range of acquisition and reconstruction parameters present in the clinical environment. Funding support: NIH U01 CA181156; Disclosures (McNitt-Gray): Institutional research agreement, Siemens Healthcare; Past recipient, research grant support, Siemens Healthcare; Consultant, Toshiba America Medical Systems; Consultant, Samsung Electronics.

60 APPLIED LIFE SCIENCES↗

Locally Resolved Electron Emission Area and Unified View of Field Emission from Ultrananocrystalline Diamond Films

One of the common problems in case of field emission from polycrystalline diamond films, which typically have uniform surface morphology, is uncertainty in determining exact location of electron emission sites across the surface. Although several studies have suggested that grain boundaries are the main electron emission source, it is not particularly clear what makes some sites emit more than the others. It is also practically unclear how one could quantify the actual electron emission area and therefore field emission current per unit area. In this paper we study the effect of actual, locally resolved, field emission (FE) area on electron emission characteristics of uniform planar highly conductive nitrogen-incorporated ultrananocrystalline diamond ((N)UNCD) field emitters. It was routinely found that field emission from as-grown planar (N)UNCD films is always confined to a counted number of discrete emitting centers across the surface which varied in size and electron emissivity. It was established that the actual FE area critically depends on the applied electric field, as well as that the actual FE area and the overall electron emissivity improve with sp2 fraction present in the film irrespectively of the original substrate roughness and morphology. To quantify the actual FE area and its dependence on the applied electric field, imaging experiments were carried out in a vacuum system in a parallel-plate configuration with a specialty anode phosphor screen. Electron emission micrographs were taken concurrently with I-V characteristics measurements. In addition, a novel automated image processing algorithm was developed to process extensive imaging datasets and calculate emission area per image. By doing so, it was determined that the emitting area was always significantly smaller than the FE cathode surface area. Namely, the actual FE area would change from 5×10-3 % to 1.5 % of the total cathode area with the applied electric field increased. Finally and most importantly, it was shown that when I-E curves as measured in the experiment were normalized by the field-dependent emission area, the resulting j-E curves demonstrated a strong kink and significant deviation from Fowler-Nordheim (FN) law, and eventually saturated at a current density of ~100 mA/cm2 . This value was nearly identical for all (N)UNCD films measured in this study, regardless of the substrate.

36 MATERIALS SCIENCE↗

Maximum entropy models of ecosystem functioning

Using organism-level traits to deduce community-level relationships is a fundamental problem in theoretical ecology. This problem parallels the physical one of using particle properties to deduce macroscopic thermodynamic laws, which was successfully achieved with the development of statistical physics. Drawing on this parallel, theoretical ecologists from Lotka onwards have attempted to construct statistical mechanistic theories of ecosystem functioning. Jaynes’ broader interpretation of statistical mechanics, which hinges on the entropy maximisation algorithm (MaxEnt), is of central importance here because the classical foundations of statistical physics do not have clear ecological analogues (e.g. phase space, dynamical invariants). However, models based on the information theoretic interpretation of MaxEnt are difficult to interpret ecologically. Here I give a broad discussion of statistical mechanical models of ecosystem functioning and the application of MaxEnt in these models. Emphasising the sample frequency interpretation of MaxEnt, I show that MaxEnt can be used to construct models of ecosystem functioning which are statistical mechanical in the traditional sense using a savanna plant ecology model as an example.

54 ENVIRONMENTAL SCIENCES↗

Optimal control of a repowered vehicle: Plug-in fuel cell against plug-in hybrid electric powertrain

This paper describes two different powertrain configurations for the repowering of a conventional vehicle, equipped with an internal combustion engine (ICE). A model of a mid-sized ICE-vehicle is realized and then modified to model both a parallel plug-in hybrid electric powertrain and a proton electrolyte membrane (PEM) fuel cell (FC) hybrid powertrain. The vehicle behavior under the application of an optimal control algorithm for the energy management is analyzed for the different scenarios and results are compared.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗