Search NASA⌕ Search

SEARCH · Search NASA

Results for “Computational 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 1,423 records · Page 79

Compiler generation and autotuning of communication-avoiding operators for geometric multigrid

This paper describes a compiler approach to introducing communication-avoiding optimizations in geometric multigrid (GMG), one of the most popular methods for solving partial differential equations. Communication-avoiding optimizations reduce vertical communication through the memory hierarchy and horizontal communication across processes or threads, usually at the expense of introducing redundant computation. We focus on applying these optimizations to the smooth operator, which successively reduces the error and accounts for the largest fraction of the GMG execution time. Our compiler technology applies both novel and known transformations to derive an implementation comparable to manually-tuned code. To make the approach portable, an underlying autotuning system explores the tradeoff between reduced communication and increased computation, as well as tradeoffs in threading schemes, to automatically identify the best implementation for a particular architecture and at each computation phase. Results show that we are able to quadruple the performance of the smooth operation on the finest grids while attaining performance within 94% of manually-tuned code. Overall we improve the overall multigrid solve time by 2.5× without sacrificing programer productivity.

97 MATHEMATICS AND COMPUTING↗

Architecture and Performance of Perlmutter’s 35 PB ClusterStor E1000 All-Flash File System

NERSC's newest system, Perlmutter, features a 35 PB all-flash Lustre file system built on HPE Cray ClusterStor E1000. We present its architecture, early performance figures, and performance considerations unique to this architecture. We demonstrate the performance of E1000 OSSes through low-level Lustre tests that achieve over 90% of the theoretical bandwidth of the SSDs at the OST and LNet levels. We also show end-to-end performance for both traditional dimensions of I/O performance (peak bulk-synchronous bandwidth) and non-optimal workloads endemic to production computing (small, incoherent I/Os at random offsets) and compare them to NERSC's previous system, Cori, to illustrate that Perlmutter achieves the performance of a burst buffer and the resilience of a scratch file system. Finally, we discuss performance considerations unique to all-flash Lustre and present ways in which users and HPC facilities can adjust their I/O patterns and operations to make optimal use of such architectures.

Lockwood, Glenn K↗

Optimizing Nuclear Physics Codes on the XT5

Scientists studying the structure and behavior of the atomic nucleus require immense high-performance computing resources to gain scientific insights. Several nuclear physics codes are capable of scaling to more than 100,000 cores on Oak Ridge National Laboratory's petaflop Cray XT5 system, Jaguar. In this paper, we present our work on optimizing codes in the nuclear physics domain.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

International Neuroscience Initiatives through the Lens of High-Performance Computing

Neuroscience initiatives aim to develop new technologies and tools to measure and manipulate neuronal circuits. In this paper, to deal with the massive amounts of data generated by these tools, the authors envision the co-location of open data repositories in standardized formats together with high-performance computing hardware utilizing open source optimized analysis codes.

60 APPLIED LIFE SCIENCES↗

Bayesian optimal experimental design for constitutive model calibration

Computational simulation is increasingly relied upon for high/consequence engineering decisions, which necessitates a high confidence in the calibration of and predictions from complex material models. However, the calibration and validation of material models is often a discrete, multi-stage process that is decoupled from material characterization activities, which means the data collected does not always align with the data that is needed. To address this issue, an integrated workflow for delivering an enhanced characterization and calibration procedure—Interlaced Characterization and Calibration (ICC)—is introduced and demonstrated. Further, this framework leverages Bayesian optimal experimental design (BOED), which creates a line of communication between model calibration needs and data collection capabilities in order to optimize the information content gathered from the experiments for model calibration. Eventually, the ICC framework will be used in quasi real-time to actively control experiments of complex specimens for the calibration of a high-fidelity material model. This work presents the critical first piece of algorithm development and a demonstration in determining the optimal load path of a cruciform specimen with simulated data. Calibration results, obtained via Bayesian inference, from the integrated ICC approach are compared to calibrations performed by choosing the load path a priori based on human intuition, as is traditionally done. The calibration results are communicated through parameter uncertainties which are propagated to the model output space (i.e. stress–strain). In these exemplar problems, data generated within the ICC framework resulted in calibrated model parameters with reduced measures of uncertainty compared to the traditional approaches.

42 ENGINEERING↗

Distance-to-Agreement Investigation of Tomotherapy's Bony Anatomy-Based Autoregistration and Planning Target Volume Contour-Based Optimization

Purpose: To compare Tomotherapy's megavoltage computed tomography bony anatomy autoregistration with the best achievable registration, assuming no deformation and perfect knowledge of planning target volume (PTV) location. Methods and Materials: Distance-to-agreement (DTA) of the PTV was determined by applying a rigid-body shift to the PTV region of interest of the prostate from its reference position, assuming no deformations. Planning target volume region of interest of the prostate was extracted from the patient archives. The reference position was set by the 6 degrees of freedom (dof)—x, y, z, roll, pitch, and yaw—optimization results from the previous study at this institution. The DTA and the compensating parameters were calculated by the shift of the PTV from the reference 6-dof to the 4-dof—x, y, z, and roll—optimization. In this study, the effectiveness of Tomotherapy's 4-dof bony anatomy–based autoregistration was compared with the idealized 4-dof PTV contour-based optimization. Results: The maximum DTA (maxDTA) of the bony anatomy-based autoregistration was 3.2 ± 1.9 mm, with the maximum value of 8.0 mm. The maxDTA of the contour-based optimization was 1.8 ± 1.3 mm, with the maximum value of 5.7 mm. Comparison of Pearson correlation of the compensating parameters between the 2 4-dof optimization algorithms shows that there is a small but statistically significant correlation in y and z (0.236 and 0.300, respectively), whereas there is very weak correlation in x and roll (0.062 and 0.025, respectively). Conclusions: We find that there is an average improvement of approximately 1 mm in terms of maxDTA on the PTV going from 4-dof bony anatomy-based autoregistration to the 4-dof contour-based optimization. Pearson correlation analysis of the 2 4-dof optimizations suggests that uncertainties due to deformation and inadequate resolution account for much of the compensating parameters, but pitch variation also makes a statistically significant contribution.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

Derivative discontinuity with localized Hartree-Fock potential

The localized Hartree-Fock potential has proven to be a computationally efficient alternative to the optimized effective potential, preserving the numerical accuracy of the latter and respecting the exact properties of being self-interaction free and having the correct −1/r asymptotics. In this paper we extend the localized Hartree-Fock potential to fractional particle numbers and observe that it yields derivative discontinuities in the energy as required by the exact theory. The discontinuities are numerically close to those of the computationally more demanding Hartree-Fock method. Our potential enjoys a “direct-energy” property, whereby the energy of the system is given by the sum of the single-particle eigenvalues multiplied by the corresponding occupation numbers. The discontinuities c{sub ↑} and c{sub ↓} of the spin-components of the potential at integer particle numbers N{sub ↑} and N{sub ↓} satisfy the condition c{sub ↑}N{sub ↑} + c{sub ↓}N{sub ↓} = 0. Thus, joining the family of effective potentials which support a derivative discontinuity, but being considerably easier to implement, the localized Hartree-Fock potential becomes a powerful tool in the broad area of applications in which the fundamental gap is an issue.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Reproducibility and Optimization Techniques for Evaluated Data in the Resolved Resonance Region [Slides]

This presentation discusses the experimental setup input parameters that are basic quantities for reproducibility. The goal is to increase quality of the evaluated data and decrease time needed for an evaluation. Full reproducibility should be achieved efficiently by including any type of constraints in the optimization procedure and to improve computational time to reach convergence. The goal of an automated reproducibility aligns to the goal of “certified” evaluated data, i.e., defined by a well-defined metric.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

An optimization method for chaotic turbulent flow

Evidence indicates that quantities-of-interest in some turbulent flows can be controlled despite the overall chaotic dynamics. It is typically thought that this is via relatively deterministic, larger-scale components of the turbulence. However, finding such controls, if they exist, is challenging because chaos causes sensitivity gradients to explode and the search space to become intractably non-convex. This challenge is analyzed, and a penalty method is introduced to cope with it. In the new approach, the time domain is broken into segments approximately matching the chaos time scales, so that the solution within each segment is both physical and relatively deterministic. The initial condition of each segment is included in an adjoint-based gradient optimization, which temporarily introduces artificial Δq discontinuities in the overall solution. The optimization then proceeds in stages with increasing penalization of Δq. The method is developed and illustrated for a logistic map, the Lorenz Equation, and an advection augmented Kuramoto–Sivashinsky Equation. These examples show how the Δq temporarily increases the search scale prior to the strong Δq → 0 penalization that recovers a physical solution. It is then applied to turbulent Kolmogorov flow, for which it also far outperforms a standard adjoint-based gradient search. Finally, the utility of such an optimized chaotic solution is discussed.

97 MATHEMATICS AND COMPUTING↗

Use of DAGMan in CRAB3 to improve the splitting of CMS user jobs

CRAB3 is a workload management tool used by CMS physicists to analyze data acquired by the Compact Muon Solenoid (CMS) detector at the CERN Large Hadron Collider (LHC). Research in high energy physics often requires the analysis of large collections of files, referred to as datasets. The task is divided into jobs that are distributed among a large collection of worker nodes throughout the Worldwide LHC Computing Grid (WLCG). Splitting a large analysis task into optimally sized jobs is critical to efficient use of distributed computing resources. Jobs that are too big will have excessive runtimes and will not distribute the work across all of the available nodes. However, splitting the project into a large number of very small jobs is also inefficient, as each job creates additional overhead which increases load on infrastructure resources. Currently this splitting is done manually, using parameters provided by the user. However the resources needed for each job are difficult to predict because of frequent variations in the performance of the user code and the content of the input dataset. As a result, dividing a task into jobs by hand is difficult and often suboptimal. In this work we present a new feature called “automatic splitting” which removes the need for users to manually specify job splitting parameters. We discuss how HTCondor DAGMan can be used to build dynamic Directed Acyclic Graphs (DAGs) to optimize the performance of large CMS analysis jobs on the Grid. We use DAGMan to dynamically generate interconnected DAGs that estimate the processing time the user code will require to analyze each event. This is used to calculate an estimate of the total processing time per job, and a set of analysis jobs are run using this estimate as a specified time limit. Some jobs may not finish within the alloted time; they are terminated at the time limit, and the unfinished data is regrouped into smaller jobs and resubmitted.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

TAO Users Manual (Rev. 3.15)

The Toolkit for Advanced Optimization (TAO) focuses on the development of algorithms and software for the solution of large-scale optimization problems on high-performance architectures. Areas of interest include unconstrained and bound-constrained optimization, nonlinear least squares problems, optimization problems with partial differential equation constraints, and variational inequalities and complementarity constraints. The development of TAO was motivated by the scattered support for parallel computations and the lack of reuse of external toolkits in current optimization software. Our aim is to produce high-quality optimization software for computing environments ranging from workstations and laptops to massively parallel high-performance architectures. Our design decisions are strongly motivated by the challenges inherent in the use of large-scale distributed memory architectures and the reality of working with large, often poorly structured legacy codes for specific applications.

97 MATHEMATICS AND COMPUTING↗

Electronic structure with direct diagonalization on a D-wave quantum annealer

Quantum chemistry is regarded to be one of the first disciplines that will be revolutionized by quantum computing. Although universal quantum computers of practical scale may be years away, various approaches are currently being pursued to solve quantum chemistry problems on near-term gate-based quantum computers and quantum annealers by developing the appropriate algorithm and software base. This work implements the general Quantum Annealer Eigensolver (QAE) algorithm to solve the molecular electronic Hamiltonian eigenvalue-eigenvector problem on a D-Wave 2000Q quantum annealer. The approach is based on the matrix formulation, efficiently uses qubit resources based on a power-of-two encoding scheme and is hardware-dominant relying on only one classically optimized parameter. We demonstrate the use of D-Wave hardware for obtaining ground and excited electronic states across a variety of small molecular systems. The approach can be adapted for use by a vast majority of electronic structure methods currently implemented in conventional quantum-chemical packages. The results of this work will encourage further development of software such as qbsolv which has promising applications in emerging quantum information processing hardware and has expectation to address large and complex optimization problems intractable for classical computers.

97 MATHEMATICS AND COMPUTING↗

Quantification of Energy Cost Savings through Optimization and Control of Appliances within Smart Neighborhood Homes

Electric utilities are driving towards enabling automatic scheduling and control of the consumption pattern of appliances such as heating, ventilation, and air conditioning (HVAC) and water heater (WH) systems (e.g., through preheating and pre-cooling, etc.) within smart neighborhoods to minimize energy cost and peak load demand. Quantifying economic savings through direct comparison of the optimized energy usage profile on a specific day with the typical non-optimized usage profile on another day is not a fair comparison because energy usage highly depends on weather conditions and human behaviour especially appliance like HVAC on those days. In this paper, we propose a novel approach of identifying similar weather day pairs which can then be used to compare the energy use profiles within homes between the identified pairs. We then demonstrate how the proposed approach can be used to compute cost savings due to optimization and control of smart appliances at home and neighborhood-level within a future-focused smart neighborhood of 62 residential homes. We also demonstrate a simulation based approach to quantify cost savings and showcase our findings through customized and interactive visualizations.

Chinthavali, Supriya↗

Topology optimization of compliant mechanisms under transient thermal conditions

This work considers multi-material topology optimization of compliant mechanisms under transient thermal and quasi-static mechanical conditions wherein thermally actuated devices are optimized for different operating conditions. The materials are modeled using finite strain thermo-hyperelasticity and a two way coupling between the energy balance and equilibrium equations is investigated. The design updates are generated from the gradient-based method of moving asymptotes optimizer and the sensitivities are computed using the time dependent adjoint sensitivity analysis. Results show the impact of designing for short versus long actuation times.

42 ENGINEERING↗

Load-balanced sparse MTTKRP on GPUs

Sparse matricized tensor times Khatri-Rao product (MT- TKRP) is one of the most computationally expensive kernels in sparse tensor computations. This work focuses on optimizing the MTTKRP for floating point operations, storage, and scalability. We begin by identifying the performance bottlenecks in directly extending the state-of-the-art CSF (compressed sparse fiber) formats from CPUs to GPUs. Our detailed analysis over the recently proposed formats shows that the lower bounds on storage and flop counts can vary significantly depending on the structure of the sparse tensor. To address this, we propose a load balanced, computation and storage-efficient scheme, HYB, which combines the best of COO (coordinate), CSF and CSL (compressed slice). With these enhancements, our GPU framework significantly out- performs the current formats on both CPU and GPU platforms.

Nisa, Israt↗

Subcritical Copper-Reflected-α-phase Plutonium (SCRαP) Integral Experiment Design

A Subcritical Copper-Reflected α-phase Plutonium (SCRαP) integral benchmark experiment is proposed where multiplication will be approximated using correlated neutron data from a detector system consisting of {sup 3}He tubes inside polyethylene. Measurements will be performed on various subcritical experimental configurations consisting of a weapons-grade plutonium sphere surrounded by different Cu thicknesses. In addition to the proposed base experimental configurations with Cu, additional configurations are planned with the plutonium ball nested in various thicknesses of interleaved polyethylene spherical shells mixed in with the Cu shells. The polyethylene is intended to provide fast neutron moderation and reflection, resulting in additional measurements with differing multiplication, spectra, and nuclear data sensitivity. The proposed experiments will performed at the National Criticality Experiments Research Center (NCERC). NCERC is home to a 4.5-kg α-phase stainless steel clad plutonium sphere that is referred to as the BeRP (Beryllium-Reflected Plutonium) ball due to its historical use in a beryllium-reflected critical experiment. In 2014, similar subcritical measurements were performed with the BeRP ball surrounded by nickel and tungsten. The nickel-reflected measurements were documented as a benchmark evaluation which was accepted by the International Criticality Safety Benchmark Evaluation Project (ICSBEP). Similar to past measurements, the proposed work will help both identify deficiencies and quantify uncertainties in nuclear data and validate computational methods related to neutron multiplication inference for subcritical benchmark evaluations. Monte Carlo simulations were performed for the preliminary design of a subcritical Copper-Reflected - phase Plutonium (SCRP) integral benchmark experiment. Specifically, MCNP{sup R} version 6 was used to determine the k{sub eff} and associated nuclear data sensitivities for different preliminary configurations (reflector thickness, material type). The final experimental design process will include continuous-energy nuclear data sensitivity analysis. This will help assess the potential impact of this integral measurement as it relates to international efforts to continuously improve current libraries (ENDF, JEFF, JENDL etc) that still tend to over/under-estimate the measured results of integral experiments, sometimes significantly. Deficiencies in underlying nuclear data quantities such as Nubar have been shown to have an effect on inferred values from subcritical measurements as well. The proposed experiments and subsequent computational validation will help identify such deficiencies. It will also help validate new nuclear data evaluations, including one for Cu that was recently performed in the resolved region up to 100 MeV for {sup 65}Cu and {sup 63}Cu which is expected to improve Cu-related benchmark simulation performance. The final experimental design phase will also include simulation results that will incorporate the detectors that will be used for the subcritical measurements. For the benchmark portion of the measurements, the main detectors that are being considered are portable neutron multiplicity counters made up of {sup 3}He tubes embedded in polyethylene and lined with cadmium. These detectors employ a data acquisition system that produces list-mode data (correlated information where each event is listed sequentially) that can be analyzed using the Feynman variance-to-mean method. Other detector systems including SNAP (single {sup 3}He tube embedded in polyethylene), and a high-purity germanium (HPGe) detector are being considered for supplementary measurements only and are not intended to be used for the final benchmark measurements. Decoupling the supplementary detectors from the system eliminates any contribution to the total system multiplication, and greatly simplifies the final computational benchmark model. Therefore, an optimization of the portable neutron multiplicity counter placement in the benchmark measurements will be performed to ensure that there is minimal contribution to the total system multiplication. In addition, new temperature measurement techniques and hardware will be explored. (authors)

42 ENGINEERING↗

A memory-distributed quasi-Newton solver for nonlinear programming problems with a small number of general constraints

We approach the problem of parallelizing state-of-the-art nonlinear programming optimization algorithms.Specifically, we focus on parallelizing quasi-Newton interior-point methods that use limited-memory secant Hessian approximations. Such interior-point methods are known to have better convergence properties and to be more effective on large-scale problems than gradient-based and derivative-free optimization algorithms. We target nonlinear and potentially nonconvex optimization problems with an arbitrary number of bound constraints and a small number of general equality and inequality constraints on the optimization variables. These problems occur for example in the form of optimal control, optimal design, and inverse problems governed by ordinary or partial differential equations, whenever they are expressed in a “reduced-space” optimization approach. We introduce and analyze the time and space complexity of a decomposition method for solving the quasi-Newton linear systems that leverages the fact that the quasi-Newton Hessian matrix has a small number of dense blocks that border a low-rank update of a diagonal matrix. This enables an efficient parallelization on memory-distributed computers of the iterations of the optimization algorithm, a state-of-the-art filter line-search interior-point algorithm by Wächter et. al. We illustrate the efficiency of the proposed method by solving structural topology optimization problems on up to 4608 cores on a parallel machine.

97 MATHEMATICS AND COMPUTING↗

A complementarity-based vapor-liquid equilibrium formulation for equation-oriented simulation and optimization

Vapor-Liquid Equilibrium (VLE) is a cornerstone of computer aided process engineering (CAPE). Embedded within process system models, VLE calculations are inherently procedural with non-smooth behavior that frequently require discrete decisions. Traditionally, these features resist the incorporation of VLE within efficient, large-scale equation-oriented (EO) process simulation and optimization strategies. On the other hand, recent reformulation of VLE models through the incorporation of complementarity constraints has broadened its scope to deal seamlessly with phase transitions and even supercritical excursions in process simulation and optimization. In this study, we extend these VLE complementarity models to EO frameworks where procedural thermodynamic property libraries are still required. Here, we develop an efficient, non-intrusive, and intuitive “square-flash” equation system that has been implemented within the IDAES Integrated Platform (IDAES-IP). Further, the effectiveness of this modular approach is demonstrated on case studies for non-ideal flash calculations and distillation optimization with disappearing phases and supercritical transitions.

97 MATHEMATICS AND COMPUTING↗