Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Scalable 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 829 records · Page 46

Coherent beam combining of pulsed fibre amplifiers with active phase control

Coherent beam combining of pulsed fibre lasers is a promising method for power scaling while simultaneously maintaining good beam quality. We propose and demonstrate a scalable architecture for coherent beam combining of all-fibre pulsed amplifiers with active phase control using the stochastic parallel gradient descent (SPGD) algorithm. A low-pass filter is introduced to eliminate the fluctuation of the metric function caused by pulsed lasers and to extract the exact phase noise signal. Active control is thereby based on the SPGD algorithm, resulting in stable coherent beam combining at the receiving plane even in a turbulent environment. Experimental results show that the fringe visibility of the long exposure pattern increases from 0 to 0.4, and the power encircled in the main-lobe increases by 1.6 times when the system evolves from the open-loop phase-locking scheme to the closed-loop scheme. This architecture can be easily scaled up to a higher power by increasing the number of amplifying channels and the power of a single amplifier.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

GSoFa: Scalable Sparse Symbolic LU Factorization on GPUs

Decomposing a matrix $\mathbf {A}$ into a lower matrix $\mathbf {L}$ and an upper matrix $\mathbf {U}$, which is also known as LU decomposition, is an essential operation in numerical linear algebra. For a sparse matrix, LU decomposition often introduces more nonzero entries in the $\mathbf {L}$ and $\mathbf {U}$ factors than in the original matrix. A symbolic factorization step is needed to identify the nonzero structures of $\mathbf {L}$ and $\mathbf {U}$ matrices. Attracted by the enormous potentials of the Graphics Processing Units (GPUs), an array of efforts have surged to deploy various LU factorization steps except for the symbolic factorization, to the best of our knowledge, on GPUs. This article introduces gSoFa, the first GPU-based symbolic factorization design with the following three optimizations to enable scalable LU symbolic factorization for nonsymmetric pattern sparse matrices on GPUs. First, here we introduce a novel fine-grained parallel symbolic factorization algorithm that is well suited for the Single Instruction Multiple Thread (SIMT) architecture of GPUs. Second, we tailor supernode detection into a SIMT friendly process and strive to balance the workload, minimize the communication and saturate the GPU computing resources during supernode detection. Third, we introduce a three-pronged optimization to reduce the excessive space consumption problem faced by multi-source concurrent symbolic factorization. Taken together, gSoFa achieves up to 31× speedup from 1 to 44 Summit nodes (6 to 264 GPUs) and outperforms the state-of-the-art CPU project, on average, by 5×. Notably, gSoFa also achieves up to 47 percent of the peak memory throughput of a V100 GPU in the Summit Supercomputer.

97 MATHEMATICS AND COMPUTING↗

The Generalized Spallation Model

The Generalized Spallation Model (GSM) event generator is an advanced, robust, and scalable Fortran code developed from an extended and improved version of the earlier event generator codes CEM03.03 and LAQGSM03.03. The GSM implements its algorithms within the context of cascade exciton, particle equilibration, particle evaporation, and particle disintegration models sampled via Monte-Carlo techniques. Coalescence of particles is also considered. The GSM assumes the incident particles has sufficient energy to satisfy the assumptions utilized in the GSM. The GSM calculates reactions induced by hadrons and nuclei incident on nuclear targets at incident energies ranging from ~20 MeV up to several TeV per nucleon. Usage of GSM is recommended however for particles with incident energies exceeding ~100 MeV to ensure model-assumptions are satisfied. The GSM additionally provides its clients with sub-model libraries to allow them to model desired physics within their own physical models.

Juneau, Chase↗

In situ tuning of optomechanical crystals with nano-oxidation

Optomechanical crystals are a promising device platform for quantum transduction and sensing. Precise targeting of the optical and acoustic resonance frequencies of these devices is crucial for future advances on these fronts. However, fabrication disorder in these wavelength-scale nanoscale devices typically leads to inhomogeneous resonance frequencies. Here we achieve in situ , selective frequency tuning of optical and acoustic resonances in silicon optomechanical crystals via electric field-induced nano-oxidation using an atomic-force microscope. Our method can achieve a tuning range >2nm (0.13%) for the optical resonance wavelength in the telecom C-band, and >60MHz (1.2%) for the acoustic resonance frequency at 5 GHz. The tuning resolution of 1.1 pm for the optical wavelength and 150 kHz for the acoustic frequency allows us to spectrally align multiple optomechanical crystal resonators using a pattern generation algorithm. Our results establish a method for precise post-fabrication tuning of optomechanical crystals. This technique can enable coupled optomechanical resonator arrays, scalable resonant optomechanical circuits, and frequency matching of microwave-optical quantum transducers.

Hatipoglu, Utku (ORCID:0000000205119196)↗

ParaText : scalable solutions for processing and searching very large document collections : final LDRD report.

This report is a summary of the accomplishments of the 'Scalable Solutions for Processing and Searching Very Large Document Collections' LDRD, which ran from FY08 through FY10. Our goal was to investigate scalable text analysis; specifically, methods for information retrieval and visualization that could scale to extremely large document collections. Towards that end, we designed, implemented, and demonstrated a scalable framework for text analysis - ParaText - as a major project deliverable. Further, we demonstrated the benefits of using visual analysis in text analysis algorithm development, improved performance of heterogeneous ensemble models in data classification problems, and the advantages of information theoretic methods in user analysis and interpretation in cross language information retrieval. The project involved 5 members of the technical staff and 3 summer interns (including one who worked two summers). It resulted in a total of 14 publications, 3 new software libraries (2 open source and 1 internal to Sandia), several new end-user software applications, and over 20 presentations. Several follow-on projects have already begun or will start in FY11, with additional projects currently in proposal.

99 GENERAL AND MISCELLANEOUS↗

Field Work Proposal ERKJ358: Black-box training for scientific machine learning models (Final Report)

The overarching goal of this project is to develop a scalable black-box training capability for scientific machine learning (SciML) problems that are non-trainable with existing automatic differentiation (AD)-based algorithms. AD assumes that a loss function can be decomposed into a sequence of elementary operations whose derivatives are known. This assumption is violated when the loss function includes a black-box physical model (e.g., a legacy simulator). The current strategy, converting a black-box simulator to an AD-enabled code via differential programming, is inflexible and time-, labor-consuming. Thus, black-box optimization is a main workhorse for training SciML models, e.g., in scientific reinforcement learning, hyper-parameter fine tuning, designing SciML models with adversarial robustness, etc.

97 MATHEMATICS AND COMPUTING↗

Genetic algorithm optimization of a chemical kinetic mechanism for propane at engine relevant conditions

Propane has demonstrated significant potential for reductions in greenhouse gas and pollutant emissions in medium- and heavy-duty engine applications, but further improvements require accurate, compact, and scalable chemical kinetic mechanisms to design the next generation of propane fueled engines, particularly at the boosted operating conditions necessary to meet the power density demand of medium- and heavy-duty applications. In this work, six key chemical reactions were identified in a reduced mechanism with 70 species and 352 reactions through a sensitivity analysis performed at conditions typical of thermodynamic trajectories observed in a high compression ratio, long stroke engine operated on propane from throttled to boosted operating conditions. While the original mechanism was validated against rapid compression machine (RCM) data, it was found to overpredict experimental autoignition tendencies in 2-zone, 0-D SI engine simulations performed in Chemkin Pro. Subsequently, a genetic algorithm approach was used to optimize the six reaction rate parameters within established uncertainty bounds by performing RCM simulations and comparing to two independent sets of literature ignition delay times for propane, thus generating two new kinetic mechanisms. The first optimization achieved a mean absolute percent error (MPE) reduction in 2nd stage ignition delay of 61.4% in seven generations, while the second optimization utilized a newer experimental RCM dataset, and achieved MPE reduction of 56.7% in seven generations, and further marginal improvement to 57.8% reduction in 34 generations. Finally, the two mechanisms were then evaluated again in the 2-zone 0-D SI engine model in Chemkin Pro comparing typical mean and knocking cycle trajectories, and it was found that the second optimized mechanism provided better prediction of knock onset at the representative conditions evaluated in this work, particularly for higher load operating conditions.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

RE-INTEGRATE EMT Simulation Tool: Input Data Processing Layer for Bulk Power System

This paper introduces an advanced input data processing layer for EMT simulations of large-scale bulk power systems. The paper proposes two versions of the RE-INTEGRATE EMT simulation tool, RE-INTEGRATE Gen-0 and RE-INTEGRATE Gen-1, which are developed to enhance simulation generalizability, scalability, and accuracy. The framework leverages a generic class design for components to incorporate linear equations, which are generated by discretizing the Differential-Algebraic Equations (DAEs) that represent the dynamics of the components. In addition, the framework employs a parsing algorithm that parses a power system’s raw and dyr files to generate a connectivity graph which is then traversed to form the overall system’s dynamics. The proposed input data processing layer is used to simulate the IEEE 39-bus test system. The obtained results demonstrate the framework’s capability to achieve simulation scalability and accuracy. Further, the results indicate that EMT simulations performed using the proposed automations can effectively handle complex grid configurations.

Mishra, Rahul [ORNL] (ORCID:0000000328205932)↗

Factors impacting performance of multithreaded sparse triangular solve.

As computational science applications grow more parallel with multi-core supercomputers having hundreds of thousands of computational cores, it will become increasingly difficult for solvers to scale. Our approach is to use hybrid MPI/threaded numerical algorithms to solve these systems in order to reduce the number of MPI tasks and increase the parallel efficiency of the algorithm. However, we need efficient threaded numerical kernels to run on the multi-core nodes in order to achieve good parallel efficiency. In this paper, we focus on improving the performance of a multithreaded triangular solver, an important kernel for preconditioning. We analyze three factors that affect the parallel performance of this threaded kernel and obtain good scalability on the multi-core nodes for a range of matrix sizes.

99 GENERAL AND MISCELLANEOUS↗

Factors impacting performance of multithreaded triangular solve.

As computational science applications grow more parallel with multi-core supercomputers having hundreds of thousands of computational cores, it will become increasingly difficult for solvers to scale. Our approach is to use hybrid MPI/threaded numerical algorithms to solve these systems in order to reduce the number of MPI tasks and increase the parallel efficiency of the algorithm. However, we need efficient threaded numerical kernels to run on the multi-core nodes in order to achieve good parallel efficiency. In this paper, we focus on improving the performance of a multithreaded triangular solver, an important kernel for preconditioning. We analyze three factors that affect the parallel performance of this threaded kernel and obtain good scalability on the multi-core nodes for a range of matrix sizes.

99 GENERAL AND MISCELLANEOUS↗

A Scalable Preconditioner for a Primal Discontinuous Petrov--Galerkin Method

In this work, we show how a scalable preconditioner for the primal discontinuous Petrov--Galerkin (DPG) method can be developed using existing algebraic multigrid (AMG) preconditioning techniques. The stability of the DPG method gives a norm equivalence which allows us to exploit existing AMG algorithms and software. We show how these algebraic preconditioners can be applied directly to a Schur complement system arising from the DPG method. One of our intermediate results shows that a generic stable decomposition implies a stable decomposition for the Schur complement. This justifies the application of algebraic solvers directly to the interface degrees of freedom. Combining such results, we obtain the first massively scalable algebraic preconditioner for the DPG system.

97 MATHEMATICS AND COMPUTING↗

Diagnosing Barren Plateaus with Tools from Quantum Optimal Control

Variational Quantum Algorithms (VQAs) have received considerable attention due to their potential for achieving near-term quantum advantage. However, more work is needed to understand their scalability. One known scaling result for VQAs is barren plateaus, where certain circumstances lead to exponentially vanishing gradients. It is common folklore that problem-inspired ansatzes avoid barren plateaus, but in fact, very little is known about their gradient scaling. In this work we employ tools from quantum optimal control to develop a framework that can diagnose the presence or absence of barren plateaus for problem-inspired ansatzes. Such ansatzes include the Quantum Alternating Operator Ansatz (QAOA), the Hamiltonian Variational Ansatz (HVA), and others. With our framework, we prove that avoiding barren plateaus for these ansatzes is not always guaranteed. Specifically, we show that the gradient scaling of the VQA depends on the degree of controllability of the system, and hence can be diagnosed through the dynamical Lie algebra $\mathfrak{g}$ obtained from the generators of the ansatz. We analyze the existence of barren plateaus in QAOA and HVA ansatzes, and we highlight the role of the input state, as different initial states can lead to the presence or absence of barren plateaus. Taken together, our results provide a framework for trainability-aware ansatz design strategies that do not come at the cost of extra quantum resources. Moreover, we prove no-go results for obtaining ground states with variational ansatzes for controllable system such as spin glasses. Our work establishes a link between the existence of barren plateaus and the scaling of the dimension of $\mathfrak{g}$.

97 MATHEMATICS AND COMPUTING↗

NO-LESS: Near optimal curtailment strategy selection for net load balancing in micro grids

The integration of renewable energy sources such as solar PVs into micro grids has increased in recent years. However, the intermittent and unpredictable nature of renewable energy generation makes supply uncertain, in turn requiring continuous net load balancing to ensure grid stability. One technique to achieve net load balancing is curtailment, however, selection of the optimal set of curtailment strategies to achieve a targeted curtailment is NP-hard. State-of-the-art curtailment selection techniques approach this problem by either developing computationally expensive optimal curtailment strategies or reduce accuracy to generate sub-optimal solutions quickly. In this work, we develop NO-LESS: a Near OptimaL CurtailmEnt Strategy Selection algorithm. NO-LESS is a novel Fully Polynomial Time Approximation Scheme (FPTAS) which determines bounded (near optimal) curtailment strategies in a small amount of time while simultaneously satisfying practical constraints on strategy switching overhead and curtailment fairness. We perform both theoretical analysis and practical evaluation to show that NO-LESS is a scalable solution for performing net load balancing through curtailment strategy selection in Micro Grids.

14 SOLAR ENERGY↗

Stochastic density functional theory

Linear-scaling implementations of density functional theory (DFT) reach their intended efficiency regime only when applied to systems having a physical size larger than the range of their Kohn–Sham density matrix (DM). This causes a problem since many types of large systems of interest have a rather broad DM range and are therefore not amenable to analysis using DFT methods. For this reason, the recently proposed stochastic DFT (sDFT), avoiding exhaustive DM evaluations, is emerging as an attractive alternative linear-scaling approach. This review develops a general formulation of sDFT in terms of a (non)orthogonal basis representation and offers an analysis of the statistical errors (SEs) involved in the calculation. Using a new Gaussian-type basis-set implementation of sDFT, applied to water clusters and silicon nanocrystals, it demonstrates and explains how the standard deviation and the bias depend on the sampling rate and the system size in various types of calculations. We also develop a basis-set embedded-fragments theory, demonstrating its utility for reducing the SEs for energy, density of states and nuclear force calculations. Finally, we discuss the algorithmic complexity of sDFT, showing it has CPU wall-time linear-scaling. Lastly, the method parallelizes well over distributed processors with good scalability and therefore may find use in the upcoming exascale computing architectures.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

DAG-Based Software Frameworks for PDEs

The task-based approach to software and parallelism is well-known and has been proposed as a potential candidate, named the silver model, for exascale software. This approach is not yet widely used in the large-scale multi-core parallel computing of complex systems of partial differential equations. After surveying task-based approaches we investigate how well the Uintah software and an extension named Wasatch fit in the task-based paradigm and how well they perform on large scale parallel computers. The conclusion is that these approaches show great promise for petascale but that considerable algorithmic challenges remain.

Directed Acyclic Graph↗

DOME: Directional medical embedding vectors from Electronic Health Records

Motivation: The increasing availability of Electronic Health Record (EHR) systems has created enormous potential for translational research. Recent developments in representation learning techniques have led to effective large-scale representations of EHR concepts along with knowledge graphs that empower downstream EHR studies. However, most existing methods require training with patient-level data, limiting their abilities to expand the training with multi-institutional EHR data. On the other hand, scalable approaches that only require summary-level data do not incorporate temporal dependencies between concepts. Methods: We introduce a DirectiOnal Medical Embedding (DOME) algorithm to encode temporally directional relationships between medical concepts, using summary-level EHR data. Specifically, DOME first aggregates patient-level EHR data into an asymmetric co-occurrence matrix. Then it computes two Positive Pointwise Mutual Information (PPMI) matrices to correspondingly encode the pairwise prior and posterior dependencies between medical concepts. Following that, a joint matrix factorization is performed on the two PPMI matrices, which results in three vectors for each concept: a semantic embedding and two directional context embeddings. They collectively provide a comprehensive depiction of the temporal relationship between EHR concepts. Results: We highlight the advantages and translational potential of DOME through three sets of validation studies. First, DOME consistently improves existing direction-agnostic embedding vectors for disease risk prediction in several diseases, for example achieving a relative gain of 5.5% in the area under the receiver operating characteristic (AUROC) for lung cancer. Second, DOME excels in directional drug-disease relationship inference by successfully differentiating between drug side effects and indications, correspondingly achieving relative AUROC gain over the state-of-the-art methods by 10.8% and 6.6%. Finally, DOME effectively constructs directional knowledge graphs, which distinguish disease risk factors from comorbidities, thereby revealing disease progression trajectories. The source codes are provided at https://github.com/celehs/Directional-EHRembedding.

60 APPLIED LIFE SCIENCES↗

Asynchronous Two-Level Checkpointing Scheme for Large-Scale Adjoints in the Spectral-Element Solver Nek5000

Adjoints are an important computational tool for large-scale sensitivity evaluation, uncertainty quantification, and derivative-based optimization. An essential component of their performance is the storage/recomputation balance in which efficient checkpointing methods play a key role. We introduce a novel asynchronous two-level adjoint checkpointing scheme for multistep numerical time discretizations targeted at large-scale numerical simulations. The checkpointing scheme combines bandwidth-limited disk checkpointing and binomial memory checkpointing. Based on assumptions about the target petascale systems, which we later demonstrate to be realistic on the IBM Blue Gene/Q system Mira, we create a model of the expected performance of our checkpointing approach and validate it using the highly scalable Navier-Stokes spectralelement solver Nek5000 on small to moderate subsystems of the Mira supercomputer. In turn, this allows us to predict optimal algorithmic choices when using all of Mira. We also demonstrate that two-level checkpointing is significantly superior to single-level checkpointing when adjoining a large number of time integration steps. To our knowledge, this is the first time two-level checkpointing had been designed, implemented, tuned, and demonstrated on fluid dynamics codes at large scale of 50k+ cores.

Adjoints↗