Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallel in Time”

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 415 records · Page 23

Modeling Cooperative Threads to Project GPU Performance for Adaptive Parallelism

Most accelerators, such as graphics processing units (GPUs) and vector processors, are particularly suitable for accelerating massively parallel workloads. On the other hand, conventional workloads are developed for multi-core parallelism, which often scale to only a few dozen OpenMP threads. When hardware threads significantly outnumber the degree of parallelism in the outer loop, programmers are challenged with efficient hardware utilization. A common solution is to further exploit the parallelism hidden deep in the code structure. Such parallelism is less structured: parallel and sequential loops may be imperfectly nested within each other, neigh boring inner loops may exhibit different concurrency patterns (e.g. Reduction vs. Forall), yet have to be parallelized in the same parallel section. Many input-dependent transformations have to be explored. A programmer often employs a larger group of hardware threads to cooperatively walk through a smaller outer loop partition and adaptively exploit any encountered parallelism. This process is time-consuming and error-prone, yet the risk of gaining little or no performance remains high for such workloads. To reduce risk and guide implementation, we propose a technique to model workloads with limited parallelism that can automatically explore and evaluate transformations involving cooperative threads. Eventually, our framework projects the best achievable performance and the most promising transformations without implementing GPU code or using physical hardware. We envision our technique to be integrated into future compilers or optimization frameworks for autotuning.

Meng, Jiayuan↗

Computationally efficient implementation of combustion chemistry in parallel PDF calculations

In parallel calculations of combustion processes with realistic chemistry, the serial in situ adaptive tabulation (ISAT) algorithm [S.B. Pope, Computationally efficient implementation of combustion chemistry using in situ adaptive tabulation, Combustion Theory and Modelling, 1 (1997) 41-63; L. Lu, S.B. Pope, An improved algorithm for in situ adaptive tabulation, Journal of Computational Physics 228 (2009) 361-386] substantially speeds up the chemistry calculations on each processor. To improve the parallel efficiency of large ensembles of such calculations in parallel computations, in this work, the ISAT algorithm is extended to the multi-processor environment, with the aim of minimizing the wall clock time required for the whole ensemble. Parallel ISAT strategies are developed by combining the existing serial ISAT algorithm with different distribution strategies, namely purely local processing (PLP), uniformly random distribution (URAN), and preferential distribution (PREF). The distribution strategies enable the queued load redistribution of chemistry calculations among processors using message passing. They are implemented in the software x2f{sub m}pi, which is a Fortran 95 library for facilitating many parallel evaluations of a general vector function. The relative performance of the parallel ISAT strategies is investigated in different computational regimes via the PDF calculations of multiple partially stirred reactors burning methane/air mixtures. The results show that the performance of ISAT with a fixed distribution strategy strongly depends on certain computational regimes, based on how much memory is available and how much overlap exists between tabulated information on different processors. No one fixed strategy consistently achieves good performance in all the regimes. Therefore, an adaptive distribution strategy, which blends PLP, URAN and PREF, is devised and implemented. It yields consistently good performance in all regimes. In the adaptive parallel ISAT strategy, the type and extent of redistribution is determined 'on the fly' based on the prediction of future simulation time. Compared to the PLP/ISAT strategy where chemistry calculations are essentially serial, a speed-up factor of up to 30 is achieved. The study also demonstrates that the adaptive strategy has acceptable parallel scalability.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Parallel Nonnegative CP Decomposition of Dense Tensors

The CP tensor decomposition is a low-rank approximation of a tensor. We present a distributed-memory parallel algorithm and implementation of an alternating optimization method for computing a CP decomposition of dense tensors that can enforce nonnegativity of the computed low-rank factors. The principal task is to parallelize the Matricized-Tensor Times Khatri-Rao Product (MTTKRP) bottleneck subcomputation. The algorithm is computation efficient, using dimension trees to avoid redundant computation across MTTKRPs within the alternating method. Our approach is also communication efficient, using a data distribution and parallel algorithm across a multidimensional processor grid that can be tuned to minimize communication. We benchmark our software on synthetic as well as hyperspectral image and neuroscience dynamic functional connectivity data, demonstrating that our algorithm scales well to 100s of nodes (up to 4096 cores) and is faster and more general than the currently available parallel software.

Ballard, Grey↗

Parallelizing autotuning for HPC applications: Unveiling the potential of the speculation strategy in Bayesian optimization

In the exascale computing era, tuning High-Performance Computing (HPC) applications has become a significant computational challenge. Although Bayesian optimization (BO) has emerged as a promising tool for HPC performance tuning, the BO workflow is inherently sequential (i.e., one function evaluation at a time) and cannot leverage the huge amount of parallel resources present in modern supercomputers, resulting in a considerable underutilization of their computational capabilities. This paper explores the trade-off between search quality and parallelism in BO, investigating a diverse set of methods. Building upon both previous approaches from the literature and novel methodologies introduced in this work, our study provides a deep analysis to accelerate BO performance tuning. By examining a set of synthetic functions and practical HPC applications, our exploration analyzes the interaction among various BO methods for parallelization, the quantity of parallel resources, the runtime distribution of target HPC applications, and the costs associated with different search orchestration mechanisms that have been overlooked in previous studies. Compared to sequential BO, our novel methodology achieves comparable quality while demonstrating robust scalability in search time as the amount of parallel resources increases; it also outperforms a state-of-the-art tuner, which supports parallelization, achieving up to 3.67x faster search time. We provide high-value insights for practitioners seeking to leverage the power of parallel computing for efficient HPC application tuning. Additionally, to further assist researchers in accelerating the performance tuning of their HPC applications, we provide an extension of an existing open-source tuning framework that incorporates our methods.

Bayesian optimization↗

Real-time and post-hoc compression for data from Distributed Acoustic Sensing

Distributed Acoustic Sensing (DAS) is an emerging sensing technology that records the strain-rate along fiber optic cables at high spatial and temporal resolution. This technique is becoming a popular tool in seismology, hydrology, and other subsurface monitoring applications. However, due to the large coverage (10’s of km) and high density of measurements (1m spacing at 100’s of Hz), a DAS installation could produce terabytes of data records per day. Because many DAS instruments are deployed in remote locations, this large data size poses significant challenges to its transfer and storage. In this paper, we explore lossless compression methods to reduce the storage requirement in both real-time and post-hoc scenarios. Here we propose a two-stage compression method to improve the compression ratio and compression speed. This two-stage compression method could reduce the storage requirement by 40%, which is 20% more than other lossless methods, such as ZSTD. We demonstrate that the compression method could complete its operation well before the DAS instrument needs to output the next file, making it suitable for real-time DAS acquisition. We also implement a parallel compression method for a post-hoc scenario and demonstrate that our method could effectively utilize a parallel computer. With 256 CPU cores, our parallel compression method achieves the speed of 26GB/second.

58 GEOSCIENCES↗

Photon Detection System for DUNE Low-Energy Physics Study and the Demonstration of a Timing Resolution of a Few Nanoseconds Using ProtoDUNE-SP PDS

Photon detection systems (PDS) are an integral part of liquid-argon neutrino detectors. Besides providing the timing information for an event, which is necessary for reconstructing the drift coordinates of ionizing particle tracks, photon detectors can be effectively used for other purposes, including triggering events, background rejection, and calorimetric energy estimation. PDS in particular for the DUNE Far Detector Module 2 is designed to achieve a more extended optical coverage (→4 ) with new-generation large-size PD modules based on the ARAPUCA technology. This will provide enhanced opportunities for the study of low-energy neutrino physics using PDS. The ARAPUCA technology was extensively tested within the ProtoDUNE-SP detector operated at the CERN neutrino platform. Here, we present a study of the timing resolution of ARAPUCA detectors using light emitted from a sample of energetic cosmic ray muons traveling parallel to the PDS. An intrinsic timing resolution in the order of 3 ns is observed for the ARAPUCA detectors. The excellent timing resolution ability of PDS can be exploited for further enhancing physics studies using the DUNE far detectors.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

On the inward drift of runaway electrons during the plateau phase of runaway current

The well observed inward drift of current carrying runaway electrons during runaway plateau phase after disruption is studied by considering the phase space dynamic of runaways in a large aspect ratio toroidal system. We consider the case where the toroidal field is unperturbed and the toroidal symmetry of the system is preserved. The balance between the change in canonical angular momentum and the input of mechanical angular momentum in such a system requires runaways to drift horizontally in configuration space for any given change in momentum space. The dynamic of this drift can be obtained by integrating the modified Euler-Lagrange equation over one bounce time. It is then found that runaway electrons will always drift inward as long as they are decelerating. This drift motion is essentially non-linear, since the current is carried by runaways themselves, and any runaway drift relative to the magnetic axis will cause further displacement of the axis itself. A simplified analytical model is constructed to describe such inward drift both in the ideal wall case and no wall case, and the runaway current center displacement as a function of parallel momentum variation is obtained. The time scale of such displacement is estimated by considering effective radiation drag, which shows reasonable agreement with the observed displacement time scale. Furthermore, this indicates that the phase space dynamic studied here plays a major role in the horizontal displacement of runaway electrons during plateau phase. (C) 2016 AIP Publishing LLC.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

IMS Rapid Response 2024 Summary Report: A Machine Learning Potential for the Periodic Table

Stockpile stewardship and nuclear waste remediation are inherently chemically complex, involving practically the full diversity of the periodic table, but existing methods are too expensive or not functional for a large diversity of atom types. Overall, the field of machine learning interatomic potentials (MLIPs) has advanced dramatically in 2024 with large high-accuracy datasets existing for bulk, surface, and organic chemical systems and new online leaderboards for diverse chemistry. To participate in, and bring LANL interests into this ecosystem, here, we have built upon existing technologies created by LANL to create a framework capable of creating machine learning interatomic potentials (MLIPs) for over 90 atom types. Our results have created a massively diverse coordination complex training dataset more than 3 times the size of existing datasets, parallelized MLIP training over multiple GPUs, enabling the training of an MLIP spanning the periodic table at 20 times the speed of prior training on 32 GPUs. These advances are substantial towards creation on foundational MLIPs for LANL-specific application areas.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Application of the Differential Evolution Method to Solving Inverse Transport Problems

The differential evolution method, a powerful stochastic optimization algorithm that mimics the process of evolution in nature, is applied to inverse transport problems with several unknown parameters of mixed types, including interface location identification, source composition identification, and material mass density identification, in spherical and cylindrical radioactive source/shield systems. In spherical systems, measurements of leakages of discrete gamma-ray lines are assumed, while in cylindrical systems, measurements of scalar fluxes of discrete lines at points outside the system are assumed. The performance of the differential evolution algorithm is compared to the Levenberg-Marquardt method, a standard gradient-based technique, and the covariance matrix adaptation evolution strategy, another stochastic technique, on a variety of numerical test problems with several (i.e., three or more) unknown parameters. Numerical results indicate that differential evolution is the most adept method for finding the global optimum for these problems. In spherical geometry, differential evolution implemented serially is run-time competitive with gradient-based methods, while a parallel version of differential evolution would be run-time competitive with gradient-based techniques in cylindrical geometry. A hybrid differential evolution/Levenberg-Marquardt method is also introduced, and numerical results indicate that it can be a fast and robust optimizer for inverse transport problems.

99 GENERAL AND MISCELLANEOUS↗

Developing an ELM Ecosystem Dynamics Model on GPU with OpenACC

Porting a complex scientific code, such as the E3SM land model (ELM), onto a new computing architecture is challenging. The paper presents design strategies and technical approaches to develop an ELM ecosystem dynamics model with compiler directives (OpenACC) on NVIDIA GPUs. The code has been refactored with advanced OpenACC features (such as deepcopy and routine directives) to reduce memory consumption and to increase the levels of parallelism through parallel loop reconstruction and new data structures. As a result, the optimized parallel implementation achieved more than a 140-time speedup (50 ms vs 7600 ms), compared to a naive implementation that uses OpenACC routine directive and parallelizes the code across existing loops on a single NVIDIA V100. On a fully loaded computing node with 44 CPUs and 6 GPUs, the code achieved over a 3.0-times speedup, compared to the original code on the CPU. Furthermore, the memory footprint of the optimized parallel implementation is 300 MB, which is around 15% of the 2.15 GB of memory consumed by a naive implementation. This study is the first effort to develop the ELM component on GPUs efficiently to support ultra-high-resolution land simulations at continental scales.

Schwartz, Peter↗

Multiscale Mathematics for Biomass Conversion to Renewable Hydrogen

Our two key accomplishments in the first three years were towards the development of, (1) a mathematically rigorous and at the same time computationally flexible framework for parallelization of Kinetic Monte Carlo methods, and its implementation on GPUs, and (2) spatial multilevel coarse-graining methods for Monte Carlo sampling and molecular simulation. A common underlying theme in both these lines of our work is the development of numerical methods which are at the same time both computationally efficient and reliable, the latter in the sense that they provide controlled-error approximations for coarse observables of the simulated molecular systems. Finally, our key accomplishment in the last year of the grant is that we started developing (3) pathwise information theory-based and goal-oriented sensitivity analysis and parameter identification methods for complex high-dimensional dynamics and in particular of nonequilibrium extended (high-dimensional) systems. We discuss these three research directions in some detail below, along with the related publications.

97 MATHEMATICS AND COMPUTING↗

Do optimization methods in deep learning applications matter?

With advances in deep learning, exponential data growth and increasing model complexity, developing efficient optimization methods are attracting much research attention. Several implementations favor the use of Conjugate Gradient (CG) and Stochastic Gradient Descent (SGD) as being practical and elegant solutions to achieve quick convergence, however, these optimization processes also present many limitations in learning across deep learning applications. Recent research is exploring higher-order optimization functions as better approaches, but these present very complex computational challenges for practical use. Comparing first and higher-order optimization functions, in this paper, our experiments reveal that Levemberg-Marquardt (LM) significantly supersedes optimal convergence but suffers from very large processing time increasing the training complexity of both, classification and reinforcement learning problems. Our experiments compare off-the-shelf optimization functions(CG, SGD, LM and L-BFGS) in standard CIFAR, MNIST, CartPole and FlappyBird experiments.The paper presents arguments on which optimization functions to use and further, which functions would benefit from parallelization efforts to improve pretraining time and learning rate convergence.

Ozyildirim, Buse Melis↗

Ignition Delay Times and CO Composition during Ethanol Combustion in Miniature Shock Tube

Ignition delay times and composition of CO were measured simultaneously in a miniature high repetition rate shock tube (HRRST) using chemiluminescence and tunable diode laser absorption spectroscopy. Mixtures of 1% ethanol with oxygen and balance argon were tested with equivalence ratios 0.5, 1, and 1.2, over pressure range of 1.6 to 1.8 bar, and temperature range of 1180 to 1400K. Ignition delay times in this range were compared with those in larger shock tubes using several different techniques. One method included using the Livengood-Wu integral (and improved estimates from measured pressure trace). Another method involved direct modeling of conditions in the miniature shock tube using the measured pressure trace. CO composition was measured in parallel with the ignition delay time experiments. Comparison was made with mechanisms in the literature. Experiments were also performed doping CO into the Ar bath gas and conducting two-line CO measurements using TDLAS before ignition. With proper calibration, this measurement is sensitive to temperature.

Sutar, Ashish↗

Accelerated molecular dynamics simulations of dislocation climb in nickel

We report the mechanical behavior of materials operating under high temperatures is strongly influenced by creep mechanisms such as dislocation climb, which is controlled by the diffusion of vacancies. However, atomistic simulations of these mechanisms have traditionally been impractical due to the long time scales required. To overcome these time scale challenges, we use Parallel Trajectory Splicing (ParSplice), an accelerated molecular dynamics method, to simulate dislocation climb in nickel. We focus on modeling the activity of a vacancy near a jog on an edge dislocation in order to observe vacancy pipe diffusion and vacancy absorption at the jog. From rigorously constructed trajectories encompassing more than 2000 vacancy absorption events over a simulation time of more than 4μs at 900 K, a comprehensive sampling of available atomistic mechanisms is collated and analyzed further with molecular statics calculations. We estimate average rates for pipe diffusion and vacancy absorption into the jog using data from the dynamic and static calculations, finding very good agreement. Our results strongly suggest that the dominant mechanism for vacancy absorption by jogs is via biased diffusion to the dislocation core followed by fast pipe diffusion to the jog.

36 MATERIALS SCIENCE↗

Organizing the Extremely Large LSST Database forReal-Time Astronomical Processing

The Large Synoptic Survey Telescope (LSST) will catalog billions of astronomical objects and trillions of sources, all of which will be stored and managed by a database management system. One of the main challenges is real-time alert generation. To generate alerts, up to 100K new difference detections have to be cross-correlated with the huge historical catalogs, and then further processed to prune false alerts. This paper explains the challenges, the implementation of the LSST Association Pipeline and the database organization strategies we are planning to use to meet the real-time requirements, including data partitioning, parallelization, and pre-loading.

79 ASTRONOMY AND ASTROPHYSICS↗

Virtual machine-based simulation platform for mobile ad-hoc network-based cyber infrastructure

In modeling and simulating complex systems such as mobile ad-hoc networks (MANETs) in de-fense communications, it is a major challenge to reconcile multiple important considerations: the rapidity of unavoidable changes to the software (network layers and applications), the difficulty of modeling the critical, implementation-dependent behavioral effects, the need to sustain larger scale scenarios, and the desire for faster simulations. Here we present our approach in success-fully reconciling them using a virtual time-synchronized virtual machine(VM)-based parallel ex-ecution framework that accurately lifts both the devices as well as the network communications to a virtual time plane while retaining full fidelity. At the core of our framework is a scheduling engine that operates at the level of a hypervisor scheduler, offering a unique ability to execute multi-core guest nodes over multi-core host nodes in an accurate, virtual time-synchronized manner. In contrast to other related approaches that suffer from either speed or accuracy issues, our framework provides MANET node-wise scalability, high fidelity of software behaviors, and time-ordering accuracy. The design and development of this framework is presented, and an ac-tual implementation based on the widely used Xen hypervisor system is described. Benchmarks with synthetic and actual applications are used to identify the benefits of our approach. The time inaccuracy of traditional emulation methods is demonstrated, in comparison with the accurate execution of our framework verified by theoretically correct results expected from analytical models of the same scenarios. In the largest high fidelity tests, we are able to perform virtual time-synchronized simulation of 64-node VM-based full-stack, actual software behaviors of MANETs containing a mix of static and mobile (unmanned airborne vehicle) nodes, hosted on a 32-core host, with full fidelity of unmodified ad-hoc routing protocols, unmodified application executables, and user-controllable physical layer effects including inter-device wireless signal strength, reachability, and connectivity.

97 MATHEMATICS AND COMPUTING↗

An S N Algorithm for Modern Architectures

LANL discrete ordinates transport packages are required to perform large, computationally intensive time-dependent calculations on massively parallel architectures, where even a single such calculation may need many months to complete. While KBA methods scale out well to very large numbers of compute nodes, we are limited by practical constraints on the number of such nodes we can actually apply to any given calculation. Instead, we describe a modified KBA algorithm that allows realization of the reductions in solution time offered by both the current, and future, architectural changes within a compute node.

97 MATHEMATICS AND COMPUTING↗