Search NASA⌕ Search

SEARCH · Search NASA

Results for “Parallel optimization”

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

Wilson Dslash Kernel From Lattice QCD Optimization

Lattice Quantum Chromodynamics (LQCD) is a numerical technique used for calculations in Theoretical Nuclear and High Energy Physics. LQCD is traditionally one of the first applications ported to many new high performance computing architectures and indeed LQCD practitioners have been known to design and build custom LQCD computers. Lattice QCD kernels are frequently used as benchmarks (e.g. 168.wupwise in the SPEC suite) and are generally well understood, and as such are ideal to illustrate several optimization techniques. In this chapter we will detail our work in optimizing the Wilson-Dslash kernels for Intel Xeon Phi, however, as we will show the technique gives excellent performance on regular Xeon Architecture as well.

Lattice QCD, Wilson-Dslash, Parallel Computing, Op↗

Parallel IO Libraries for Managing HEP Experimental Data

The computing and storage requirements of the energy and intensity frontiers will grow significantly during the Run 4 & 5 and the HL-LHC era. Similarly, in the intensity frontier, with larger trig ger readouts during supernovae explosions, the Deep Underground Neutrino Experiment (DUNE) will have unique computing challenges that could be addressed by the use of parallel and accelerated dataprocessing capabilities. Most of the requirements of the energy and intensity frontier experiments rely on increasing the role of high performance computing (HPC) in the HEP community. In this presentation, we will describe our ongoing efforts that are focused on using HPC resources for the next generation HEP experiments. The HEPCCE (High Energy Physics-Center for Computational Excellence) IOS (Input/Output and Storage) group has been developing approaches to map HEP data to the HDF5 , an IO library optimized for the HPC platforms to store the intermediate HEP data. The complex HEP data products are serialized using ROOT to allow for experiment independent general mapping approaches of the HEP data to the HDF5 format. The mapping approaches can be optimized for high performance parallel IO. Similarly, simpler data can be directly mapped into the HDF5, which can also be suitable for offloading into the GPUs directly. We will present our works on both complex and simple data model models.

Bashyal, Amit↗

Parallel time integration software

This package implements an optimal-scaling multigrid solver for the (non) linear systems that arise from the discretization of problems with evolutionary behavior. Typically, solution algorithms for evolution equations are based on a time-marching approach, solving sequentially for one time step after the other. Parallelism in these traditional time-integrarion techniques is limited to spatial parallelism. However, current trends in computer architectures are leading twards system with more, but not faster. processors. Therefore, faster compute speeds must come from greater parallelism. One approach to achieve parallelism in time is with multigrid, but extending classical multigrid methods for elliptic poerators to this setting is a significant achievement. In this software, we implement a non-intrusive, optimal-scaling time-parallel method based on multigrid reduction techniques. The examples in the package demonstrate optimality of our multigrid-reduction-in-time algorithm (MGRIT) for solving a variety of parabolic equations in two and three sparial dimensions. These examples can also be used to show that MGRIT can achieve significant speedup in comparison to sequential time marching on modern architectures.

Source record↗

Parallel time integration software

This package implements an optimal-scaling multigrid solver for the (non) linear systems that arise from the discretization of problems with evolutionary behavior. Typically, solution algorithms for evolution equations are based on a time-marching approach, solving sequentially for one time step after the other. Parallelism in these traditional time-integrarion techniques is limited to spatial parallelism. However, current trends in computer architectures are leading twards system with more, but not faster. processors. Therefore, faster compute speeds must come from greater parallelism. One approach to achieve parallelism in time is with multigrid, but extending classical multigrid methods for elliptic poerators to this setting is a significant achievement. In this software, we implement a non-intrusive, optimal-scaling time-parallel method based on multigrid reduction techniques. The examples in the package demonstrate optimality of our multigrid-reduction-in-time algorithm (MGRIT) for solving a variety of parabolic equations in two and three sparial dimensions. These examples can also be used to show that MGRIT can achieve significant speedup in comparison to sequential time marching on modern architectures.

Source record↗

Randomized Sampling for Large Data Applications of SVM

A trend in machine learning is the application of existing algorithms to ever-larger datasets. Support Vector Machines (SVM) have been shown to be very effective, but have been difficult to scale to large-data problems. Some approaches have sought to scale SVM training by approximating and parallelizing the underlying quadratic optimization problem. This paper pursues a different approach. Our algorithm, which we call Sampled SVM, uses an existing SVM training algorithm to create a new SVM training algorithm. It uses randomized data sampling to better extend SVMs to large data applications. Experiments on several datasets show that our method is faster than and comparably accurate to both the original SVM algorithm it is based on and the Cascade SVM, the leading data organization approach for SVMs in the literature. Further, we show that our approach is more amenable to parallelization than Cascade SVM.

large data↗

Performance Portability Evaluation of Fluid-Structure Interaction Simulations on Heterogeneous Platforms

The rapid proliferation of heterogeneous programming languages and multi-vendor hardware has underscored the critical need to evaluate the performance portability of scientific applications. In this work, we present the systematic porting and optimization of a massively parallel fluid-structure interaction code across multiple heterogeneous programming frameworks for deployment on leadership-class supercomputers from major vendors. Our analysis focuses on at-scale performance for simulations involving hundreds of millions of deformable cells, executed on a combination of CPUs and GPUs spanning thousands of nodes on exascale machines. We benchmark the performance of each implementation, highlighting the trade-offs inherent in adopting diverse programming models. Key insights regarding the portability of CUDA on multi-vendor platforms, the superior multi-core CPU performance from SYCL, and architectural considerations on performance optimization are distilled from our experience, offering guidance to other users of high performance computing based on our findings.

Martin, Aristotle [Duke University]↗

A nonvoxel-based dose convolution/superposition algorithm optimized for scalable GPU architectures

Purpose: Real-time adaptive planning and treatment has been infeasible due in part to its high computational complexity. There have been many recent efforts to utilize graphics processing units (GPUs) to accelerate the computational performance and dose accuracy in radiation therapy. Data structure and memory access patterns are the key GPU factors that determine the computational performance and accuracy. In this paper, the authors present a nonvoxel-based (NVB) approach to maximize computational and memory access efficiency and throughput on the GPU. Methods: The proposed algorithm employs a ray-tracing mechanism to restructure the 3D data sets computed from the CT anatomy into a nonvoxel-based framework. In a process that takes only a few milliseconds of computing time, the algorithm restructured the data sets by ray-tracing through precalculated CT volumes to realign the coordinate system along the convolution direction, as defined by zenithal and azimuthal angles. During the ray-tracing step, the data were resampled according to radial sampling and parallel ray-spacing parameters making the algorithm independent of the original CT resolution. The nonvoxel-based algorithm presented in this paper also demonstrated a trade-off in computational performance and dose accuracy for different coordinate system configurations. In order to find the best balance between the computed speedup and the accuracy, the authors employed an exhaustive parameter search on all sampling parameters that defined the coordinate system configuration: zenithal, azimuthal, and radial sampling of the convolution algorithm, as well as the parallel ray spacing during ray tracing. The angular sampling parameters were varied between 4 and 48 discrete angles, while both radial sampling and parallel ray spacing were varied from 0.5 to 10 mm. The gamma distribution analysis method (γ) was used to compare the dose distributions using 2% and 2 mm dose difference and distance-to-agreement criteria, respectively. Accuracy was investigated using three distinct phantoms with varied geometries and heterogeneities and on a series of 14 segmented lung CT data sets. Performance gains were calculated using three 256 mm cube homogenous water phantoms, with isotropic voxel dimensions of 1, 2, and 4 mm. Results: The nonvoxel-based GPU algorithm was independent of the data size and provided significant computational gains over the CPU algorithm for large CT data sizes. The parameter search analysis also showed that the ray combination of 8 zenithal and 8 azimuthal angles along with 1 mm radial sampling and 2 mm parallel ray spacing maintained dose accuracy with greater than 99% of voxels passing the γ test. Combining the acceleration obtained from GPU parallelization with the sampling optimization, the authors achieved a total performance improvement factor of >175 000 when compared to our voxel-based ground truth CPU benchmark and a factor of 20 compared with a voxel-based GPU dose convolution method. Conclusions: The nonvoxel-based convolution method yielded substantial performance improvements over a generic GPU implementation, while maintaining accuracy as compared to a CPU computed ground truth dose distribution. Such an algorithm can be a key contribution toward developing tools for adaptive radiation therapy systems.

60 APPLIED LIFE SCIENCES↗

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

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

97 MATHEMATICS AND COMPUTING↗

Privateer

Privateer is a general-purpose data store that optimizes the tradeoff between storage space utilization and I/O performance. Privateer uses memory-mapped I/O with private mapping and an optimized writeback mechanism to maximize write parallelism and eliminate redundant writes; it also uses contentaddressable storage to optimize storage space via de-duplication.

Iwabuchi, Keita↗

Rapid evaluation and quality control of next generation sequencing data with FaQCs

Background: Next generation sequencing (NGS) technologies that parallelize the sequencing process and produce thousands to millions, or even hundreds of millions of sequences in a single sequencing run, have revolutionized genomic and genetic research. Because of the vagaries of any platform's sequencing chemistry, the experimental processing, machine failure, and so on, the quality of sequencing reads is never perfect, and often declines as the read is extended. These errors invariably affect downstream analysis/application and should therefore be identified early on to mitigate any unforeseen effects. Results: Here we present a novel FastQ Quality Control Software (FaQCs) that can rapidly process large volumes of data, and which improves upon previous solutions to monitor the quality and remove poor quality data from sequencing runs. Both the speed of processing and the memory footprint of storing all required information have been optimized via algorithmic and parallel processing solutions. The trimmed output compared side-by-side with the original data is part of the automated PDF output. We show how this tool can help data analysis by providing a few examples, including an increased percentage of reads recruited to references, improved single nucleotide polymorphism identification as well as de novo sequence assembly metrics. Conclusion: FaQCs combines several features of currently available applications into a single, user-friendly process, and includes additional unique capabilities such as filtering the PhiX control sequences, conversion of FASTQ formats, and multi-threading. The original data and trimmed summaries are reported within a variety of graphics and reports, providing a simple way to do data quality control and assurance.

59 BASIC BIOLOGICAL SCIENCES↗

Layer-Parallel Training of Deep Residual Neural Networks

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

97 MATHEMATICS AND COMPUTING↗

EXTENDED SCALING LAWS IN NUMERICAL SIMULATIONS OF MAGNETOHYDRODYNAMIC TURBULENCE

Magnetized turbulence is ubiquitous in astrophysical systems, where it notoriously spans a broad range of spatial scales. Phenomenological theories of MHD turbulence describe the self-similar dynamics of turbulent fluctuations in the inertial range of scales. Numerical simulations serve to guide and test these theories. However, the computational power that is currently available restricts the simulations to Reynolds numbers that are significantly smaller than those in astrophysical settings. In order to increase computational efficiency and, therefore, probe a larger range of scales, one often takes into account the fundamental anisotropy of field-guided MHD turbulence, with gradients being much slower in the field-parallel direction. The simulations are then optimized by employing the reduced MHD equations and relaxing the field-parallel numerical resolution. In this work we explore a different possibility. We propose that there exist certain quantities that are remarkably stable with respect to the Reynolds number. As an illustration, we study the alignment angle between the magnetic and velocity fluctuations in MHD turbulence, measured as the ratio of two specially constructed structure functions. We find that the scaling of this ratio can be extended surprisingly well into the regime of relatively low Reynolds number. However, the extended scaling easily becomes spoiled when the dissipation range in the simulations is underresolved. Thus, taking the numerical optimization methods too far can lead to spurious numerical effects and erroneous representation of the physics of MHD turbulence, which in turn can affect our ability to identify correctly the physical mechanisms that are operating in astrophysical systems.

79 ASTRONOMY AND ASTROPHYSICS↗

Tuning HDF5 for Lustre File Systems

HDF5 is a cross-platform parallel I/O library that is used by a wide variety of HPC applications for the flexibility of its hierarchical object-database representation of scientific data. We describe our recent work to optimize the performance of the HDF5 and MPI-IO libraries for the Lustre parallel file system. We selected three different HPC applications to represent the diverse range of I/O requirements, and measured their performance on three different systems to demonstrate the robustness of our optimizations across different file system configurations and to validate our optimization strategy. We demonstrate that the combined optimizations improve HDF5 parallel I/O performance by up to 33 times in some cases running close to the achievable peak performance of the underlying file system and demonstrate scalable performance up to 40,960-way concurrency.

97 MATHEMATICS AND COMPUTING↗

Fast optimization and dose calculation in scanned ion beam therapy

Purpose: Particle therapy (PT) has advantages over photon irradiation on static tumors. An increased biological effectiveness and active target conformal dose shaping are strong arguments for PT. However, the sensitivity to changes of internal geometry complicates the use of PT for moving organs. In case of interfractionally moving objects adaptive radiotherapy (ART) concepts known from intensity modulated radiotherapy (IMRT) can be adopted for PT treatments. One ART strategy is to optimize a new treatment plan based on daily image data directly before a radiation fraction is delivered [treatment replanning (TRP)]. Optimizing treatment plans for PT using a scanned beam is a time consuming problem especially for particles other than protons where the biological effective dose has to be calculated. For the purpose of TRP, fast optimization and fast dose calculation have been implemented into the GSI in-house treatment planning system (TPS) TRiP98. Methods: This work reports about the outcome of a code analysis that resulted in optimization of the calculation processes as well as implementation of routines supporting parallel execution of the code. To benchmark the new features, the calculation time for therapy treatment planning has been studied. Results: Compared to the original version of the TPS, calculation times for treatment planning (optimization and dose calculation) have been improved by a factor of 10 with code optimization. The parallelization of the TPS resulted in a speedup factor of 12 and 5.5 for the original version and the code optimized version, respectively. Hence the total speedup of the new implementation of the authors' TPS yielded speedup factors up to 55. Conclusions: The improved TPS is capable of completing treatment planning for ion beam therapy of a prostate irradiation considering organs at risk in this has been overseen in the review process. Also see below 6 min.

60 APPLIED LIFE SCIENCES↗

A Massively Parallel Implementation of the CCSD(T) Method Using the Resolution-of-the-Identity Approximation and a Hybrid Distributed/Shared Memory Parallelization Model

In this work, a parallel algorithm is described for the coupled-cluster singles and doubles method augmented with a perturbative correction for triple excitations [CCSD(T)] using the resolution-of-the-identity (RI) approximation for two-electron repulsion integrals (ERIs). The algorithm bypasses the storage of four-center ERIs by adopting an integral-direct strategy. The CCSD amplitude equations are given in a compact quasi-linear form by factorizing them in terms of amplitude-dressed three-center intermediates. A hybrid MPI/OpenMP parallelization scheme is employed, which uses the OpenMP-based shared memory model for intranode parallelization and the MPI-based distributed memory model for internode parallelization. Parallel efficiency has been optimized for all terms in the CCSD amplitude equations. Two different algorithms have been implemented for the rate-limiting terms in the CCSD amplitude equations that entail and -scaling computational costs, where N O and N V denote the number of correlated occupied and virtual orbitals, respectively. One of the algorithms assembles the four-center ERIs requiring N V 4 and N O 2 N V 2 -scaling memory costs in a distributed manner on a number of MPI ranks, while the other algorithm completely bypasses the assembling of quartic memory-scaling ERIs and thus largely reduces the memory demand. It is demonstrated that the former memory-expensive algorithm is faster on a few hundred cores, while the latter memory-economic algorithm shows a better strong scaling in the limit of a few thousand cores. The program is shown to exhibit a near-linear scaling, in particular for the compute-intensive triples correction step, on up to 8000 cores. The performance of the program is demonstrated via calculations involving molecules with 24–51 atoms and up to 1624 atomic basis functions. As the first application, the complete basis set (CBS) limit for the interaction energy of the π-stacked uracil dimer from the S66 data set has been investigated. This work reports the first calculation of the interaction energy at the CCSD(T)/aug-cc-pVQZ level without local orbital approximation. The CBS limit for the CCSD correlation contribution to the interaction energy was found to be -8.01 kcal/mol, which agrees very well with the value -7.99 kcal/mol reported by Schmitz, Hättig, and Tew [ Phys. Chem. Chem. Phys. 2014 , 16 , 22167-22178]. The CBS limit for the total interaction energy was estimated to be -9.64 kcal/mol.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Accurate modeling of switched reluctance machine based on hybrid trained WNN

According to the strong nonlinear electromagnetic characteristics of switched reluctance machine (SRM), a novel accurate modeling method is proposed based on hybrid trained wavelet neural network (WNN) which combines improved genetic algorithm (GA) with gradient descent (GD) method to train the network. In the novel method, WNN is trained by GD method based on the initial weights obtained per improved GA optimization, and the global parallel searching capability of stochastic algorithm and local convergence speed of deterministic algorithm are combined to enhance the training accuracy, stability and speed. Based on the measured electromagnetic characteristics of a 3-phase 12/8-pole SRM, the nonlinear simulation model is built by hybrid trained WNN in Matlab. The phase current and mechanical characteristics from simulation under different working conditions meet well with those from experiments, which indicates the accuracy of the model for dynamic and static performance evaluation of SRM and verifies the effectiveness of the proposed modeling method.

97 MATHEMATICS AND COMPUTING↗

TeraChem: Accelerating electronic structure and ab initio molecular dynamics with graphical processing units

Developed over the past decade, TeraChem is an electronic structure and ab initio molecular dynamics software package designed from the ground up to leverage graphics processing units (GPUs) to perform large-scale ground and excited state quantum chemistry calculations in the gas and the condensed phase. TeraChem’s speed stems from the reformulation of conventional electronic structure theories in terms of a set of individually optimized high-performance electronic structure operations (e.g., Coulomb and exchange matrix builds, one- and two-particle density matrix builds) and rank-reduction techniques (e.g., tensor hypercontraction). Recent efforts have encapsulated these core operations and provided language-agnostic interfaces. Finally, this greatly increases the accessibility and flexibility of TeraChem as a platform to develop new electronic structure methods on GPUs and provides clear optimization targets for emerging parallel computing architectures.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

First experience of vectorizing electromagnetic physics models for detector simulation

The recent emergence of hardware architectures characterized by many-core or accelerated processors has opened new opportunities for concurrent programming models taking advantage of both SIMD and SIMT architectures. The GeantV vector prototype for detector simulations has been designed to exploit both the vector capability of mainstream CPUs and multi-threading capabilities of coprocessors including NVidia GPUs and Intel Xeon Phi. The characteristics of these architectures are very different in terms of the vectorization depth, parallelization needed to achieve optimal performance or memory access latency and speed. An additional challenge is to avoid the code duplication often inherent to supporting heterogeneous platforms. In this paper we present the first experience of vectorizing electromagnetic physics models developed for the GeantV project.

Amadio, G.↗