Search NASA⌕ Search

SEARCH · Search NASA

Results for “linear solver”

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 649 records · Page 36

Mixed Precision Numerical Linear Algebra (Final Report)

The objective of this subcontract was to identify opportunities for the use of mixed precision within iterative solvers and to explore these opportunities both theoretically and experimentally. All quarterly milestones were achieved. We summarize the achievements each quarter in this report.

97 MATHEMATICS AND COMPUTING↗

Mixed Precision Numerical Linear Algebra (Final Report)

The objective of this subcontract was to identify opportunities for the use of mixed precision within iterative solvers and to explore these opportunities both theoretically and experimentally. All quarterly milestones were achieved. Quarterly achievements are summarized in sections.

97 MATHEMATICS AND COMPUTING↗

Mixed Precision Numerical Linear Algebra (Final Report)

The objective of this subcontract was to identify opportunities for the use of mixed precision within iterative solvers and to explore these opportunities both theoretically and experimentally. All quarterly milestones were achieved. We summarize the achievements each quarter in sections below.

97 MATHEMATICS AND COMPUTING↗

symPACK: A GPU-Capable Fan-Out Sparse Cholesky Solver

Sparse symmetric positive definite systems of equations are ubiquitous in scientific workloads and applications. Parallel sparse Cholesky factorization is the method of choice for solving such linear systems. Therefore, the development of parallel sparse Cholesky codes that can efficiently run on today’s large-scale heterogeneous distributed-memory platforms is of vital importance. Modern supercomputers offer nodes that contain a mix of CPUs and GPUs. To fully utilize the computing power of these nodes, scientific codes must be adapted to offload expensive computations to GPUs. We present symPACK, a GPU-capable parallel sparse Cholesky solver that uses one-sided communication primitives and remote procedure calls provided by the UPC++ library. We also utilize the UPC++ "memory kinds" feature to enable efficient communication of GPU-resident data. We show that on a number of large problems, symPACK outperforms comparable state-of-the-art GPU-capable Cholesky factorization codes by up to 14x on the NERSC Perlmutter supercomputer.

Bellavita, Julian↗

Deploy threading in Nalu solver stack

The goal of the ExaWind project is to enable predictive simulations of wind farms composed of many MW-scale turbines situated in complex terrain. Predictive simulations will require computational fluid dynamics (CFD) simulations for which the mesh resolves the geometry of the turbines, and captures the rotation and large deflections of blades. Whereas such simulations for a single turbine are arguably petascale class, multi-turbine wind farm simulations will require exascale-class resources. The primary code in the ExaWind project is Nalu, which is an unstructured-grid solver for the acousticallyincompressible Navier-Stokes equations, and mass continuity is maintained through pressure projection. The model consists of the mass-continuity Poisson-type equation for pressure and a momentum equation for the velocity. For such modeling approaches, simulation times are dominated by linear-system setup and solution for the continuity and momentum systems. For the ExaWind challenge problem, the moving meshes greatly affect overall solver costs as re-initialization of matrices and re-computation of preconditioners is required at every time step In this Milestone, we examine the effect of threading on the solver stack performance against flat-MPI results obtained from previous milestones using Haswell performance data full-turbine simulations. Whereas the momentum equations are solved only with the Trilinos solvers, we investigate two algebraic-multigrid preconditioners for the continuity equations: Trilinos/Muelu and HYPRE/BoomerAMG. These two packages embody smoothed-aggregation and classical Ruge-Stiiben AMG methods, respectively. In our FY18 Q2 report, we described our efforts to improve setup and solve of the continuity equations under flat-MPI parallelism. While significant improvement was demonstrated in the solve phase, setup times remained larger than expected. Starting with the optimized settings described in the Q2 report, we explore here simulation performance where OpenMP threading is employed in the solver stack. For Trilinos, threading is acheived through the Kokkos abstraction where, whereas HYPRE/BoomerAMG employs straight OpenMP. We examined results for our mid-resolution baseline turbine simulation configuration (229M DOF). Simulations on 2048 Haswell cores explored the effect of decreasing the number of MPI ranks while increasing the number of threads. Both HYPRE and Trilinos exhibited similar overal solution times, and both showed dramatic increases in simulation time in the shift from MPI ranks to OpenMP threads. This increase is attributed to the large amount of work per MPI rank starting at the single-thread configuration. Decreasing MPI ranks, while increasing threads, may be increasing simulation time due to thread synchronization and start-up overhead contributing to the latency and serial time in the model. These result showed that an MPI+OpenMP parallel decomposition will be more effective as the amount per MPI rank computation per MPI rank decreases and the communication latency increases. This idea was demonstrated in a strong scaling study of our low-resolution baseline model (29M DOF) with the Trilinos-HYPRE configuration. While MPI-only results showed scaling improvement out to about 1536 cores, engaging threading carried scaling improvements out to 4128 cores — roughly 7000 DOF per core. This is an important result as improved strong scaling is needed for simulations to be executed over sufficiently long simulated durations (i.e., for many timesteps). In addition to threading work described above, the team examined solver-performance improvements by exploring communication-overhead in the HYPRE-GMRES implementation through a communicationoptimal- GMRE algorithm (CO-GMRES), and offloading compute-intensive solver actions to GPUs. To those ends, a HYPRE mini-app was allow us to easily test different solver approaches and HYPRE parameter settings without running the entire Nalu code. With GPU acceleration on the Summitdev supercomputer, a 20x speedup was achieved for the overall preconditioner and solver execution time for the mini-app. A study on Haswell processors showed that CO-GMRES provides benefits as one increases MPI ranks.

17 WIND ENERGY↗

Graph Modeling for Quadratic Assignment Problems Associated with the Hypercube

In the paper we consider the quadratic assignment problem arising from channel coding in communications where one coefficient matrix is the adjacency matrix of a hypercube in a finite dimensional space. By using the geometric structure of the hypercube, we first show that there exist at least n different optimal solutions to the underlying QAPs. Moreover, the inherent symmetries in the associated hypercube allow us to obtain partial information regarding the optimal solutions and thus shrink the search space and improve all the existing QAP solvers for the underlying QAPs.Secondly, we use graph modeling technique to derive a new integer linear program (ILP) models for the underlying QAPs. The new ILP model has n(n-1) binary variables and O(n{sup 3} log(n)) linear constraints. This yields the smallest known number of binary variables for the ILP reformulation of QAPs. Various relaxations of the new ILP model are obtained based on the graphical characterization of the hypercube, and the lower bounds provided by the LP relaxations of the new model are analyzed and compared with what provided by several classical LP relaxations of QAPs in the literature.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

The ANACONDA algorithm for deformable image registration in radiotherapy

Purpose: The purpose of this work was to describe a versatile algorithm for deformable image registration with applications in radiotherapy and to validate it on thoracic 4DCT data as well as CT/cone beam CT (CBCT) data. Methods: ANAtomically CONstrained Deformation Algorithm (ANACONDA) combines image information (i.e., intensities) with anatomical information as provided by contoured image sets. The registration problem is formulated as a nonlinear optimization problem and solved with an in-house developed solver, tailored to this problem. The objective function, which is minimized during optimization, is a linear combination of four nonlinear terms: 1. image similarity term; 2. grid regularization term, which aims at keeping the deformed image grid smooth and invertible; 3. a shape based regularization term which works to keep the deformation anatomically reasonable when regions of interest are present in the reference image; and 4. a penalty term which is added to the optimization problem when controlling structures are used, aimed at deforming the selected structure in the reference image to the corresponding structure in the target image. Results: To validate ANACONDA, the authors have used 16 publically available thoracic 4DCT data sets for which target registration errors from several algorithms have been reported in the literature. On average for the 16 data sets, the target registration error is 1.17 ± 0.87 mm, Dice similarity coefficient is 0.98 for the two lungs, and image similarity, measured by the correlation coefficient, is 0.95. The authors have also validated ANACONDA using two pelvic cases and one head and neck case with planning CT and daily acquired CBCT. Each image has been contoured by a physician (radiation oncologist) or experienced radiation therapist. The results are an improvement with respect to rigid registration. However, for the head and neck case, the sample set is too small to show statistical significance. Conclusions: ANACONDA performs well in comparison with other algorithms. By including CT/CBCT data in the validation, the various aspects of the algorithm such as its ability to handle different modalities, large deformations, and air pockets are shown.

60 APPLIED LIFE SCIENCES↗

GentenMPI: Distributed Memory Sparse Tensor Decomposition

GentenMPl is a toolkit of sparse canonical polyadic (CP) tensor decomposition algorithms that is designed to run effectively on distributed-memory high-performance computers. Its use of distributed-memory parallelism enables it to efficiently decompose tensors that are too large for a single compute node's memory. GentenMPl leverages Sandia's decades-long investment in the Trilinos solver framework for much of its parallel-computation capability. Trilinos contains numerical algorithms and linear algebra classes that have been optimized for parallel simulation of complex physical phenomena. This work applies these tools to the data science problem of sparse tensor decomposition. In this report, we describe the use of Trilinos in GentenMPl, extensions needed for sparse tensor decomposition, and implementations of the CP-ALS (CP via alternating least squares) and GCP-SGD (generalized CP via stochastic gradient descent) sparse tensor decomposition algorithms. We show that GentenMPl can decompose sparse tensors of extreme size, e.g., a 12.6-terabyte tensor on 8192 computer cores. We demonstrate that the Trilinos backbone provides good strong and weak scaling of the tensor decomposition algorithms.

97 MATHEMATICS AND COMPUTING↗

Optimal Model-Based Fault Estimation and Correction for Particle Accelerators and Industrial Plants Using Combined Support Vector Machines and First Principles Models

Timely estimation of deviations from optimal performance in complex systems and the ability to identify corrective measures in response to the estimated parameter deviations has been the subject of extensive research over the past four decades. The implications in terms of lost revenue from costly industrial processes, operation of large-scale public works projects and the volume of the published literature on this topic clearly indicates the significance of the problem. Applications range from manufacturing industries (integrated circuits, automotive, etc.), to large-scale chemical plants, pharmaceutical production, power distribution grids, and avionics. In this project we investigated a new framework for building parsimonious models that are suited for diagnosis and fault estimation of complex technical systems. We used Support Vector Machines (SVMs) to model potentially time-varying parameters of a First-Principles (FP) description of the process. The combined SVM & FP model was built (i.e. model parameters were trained) using constrained optimization techniques. We used the trained models to estimate faults affecting simulated beam lifetime. In the case where a large number of process inputs are required for model-based fault estimation, the proposed framework performs an optimal nonlinear principal component analysis of the large-scale input space, and creates a lower dimension feature space in which fault estimation results can be effectively presented to the operation personnel. To fulfill the main technical objectives of the Phase I research, our Phase I efforts have focused on: (1) SVM Training in a Combined Model Structure - We developed the software for the constrained training of the SVMs in a combined model structure, and successfully modeled the parameters of a first-principles model for beam lifetime with support vectors. (2) Higher-order Fidelity of the Combined Model - We used constrained training to ensure that the output of the SVM (i.e. the parameters of the beam lifetime model) are physically meaningful. (3) Numerical Efficiency of the Training - We investigated the numerical efficiency of the SVM training. More specifically, for the primal formulation of the training, we have developed a problem formulation that avoids the linear increase in the number of the constraints as a function of the number of data points. (4) Flexibility of Software Architecture - The software framework for the training of the support vector machines was designed to enable experimentation with different solvers. We experimented with two commonly used nonlinear solvers for our simulations. The primary application of interest for this project has been the sustained optimal operation of particle accelerators at the Stanford Linear Accelerator Center (SLAC). Particle storage rings are used for a variety of applications ranging from 'colliding beam' systems for high-energy physics research to highly collimated x-ray generators for synchrotron radiation science. Linear accelerators are also used for collider research such as International Linear Collider (ILC), as well as for free electron lasers, such as the Linear Coherent Light Source (LCLS) at SLAC. One common theme in the operation of storage rings and linear accelerators is the need to precisely control the particle beams over long periods of time with minimum beam loss and stable, yet challenging, beam parameters. We strongly believe that beyond applications in particle accelerators, the high fidelity and cost benefits of a combined model-based fault estimation/correction system will attract customers from a wide variety of commercial and scientific industries. Even though the acquisition of Pavilion Technologies, Inc. by Rockwell Automation Inc. in 2007 has altered the small business status of the Pavilion and it no longer qualifies for a Phase II funding, our findings in the course of the Phase I research have convinced us that further research will render a workable model-based fault estimation and correction for particle accelerators and industrial plants feasible.

43 PARTICLE ACCELERATORS↗

Deploy threading in Nalu solver stack

The goal of the ExaWind project is to enable predictive simulations of wind farms composed of many MW-scale turbines situated in complex terrain. Predictive simulations will require computational fluid dynamics (CFD) simulations for which the mesh resolves the geometry of the turbines, and captures the rotation and large deflections of blades. Whereas such simulations for a single turbine are arguably petascale class, multi-turbine wind farm simulations will require exascale-class resources. The primary code in the ExaWind project is Nalu, which is an unstructured-grid solver for the acoustically-incompressible Navier-Stokes equations, and mass continuity is maintained through pressure projection. The model consists of the mass-continuity Poisson-type equation for pressure and a momentum equation for the velocity. For such modeling approaches, simulation times are dominated by linear-system setup and solution for the continuity and momentum systems. For the ExaWind challenge problem, the moving meshes greatly affect overall solver costs as re-initialization of matrices and re-computation of preconditioners is required at every time step.

17 WIND ENERGY↗

A flexible gyro-fluid system of equations

Gyro-fluid equations are velocity space moments of the gyrokinetic equations. Special gyro-Landau-fluid closures have been developed that include the damping due to kinetic resonances by fitting to the collisionless local plasma response functions. This damping allows for accurate linear eigenmodes to be computed with a relatively low number of velocity space moments compared to the number of velocity quadrature points in gyrokinetic codes. However, none of the published gyro-Landau-fluid closure schemes considers the Onsager symmetries of the resulting quasi-linear fluxes as a constraint. Onsager symmetry guarantees that the matrix of diffusivities is positive definite, an important property for the numerical stability of a transport solver. A two-parameter real closure for improving the accuracy of low-resolution gyro-fluid equations, which preserves the Onsager symmetry and allows higher velocity space moments, is presented in this paper. The new linear gyro-fluid system (GFS) is used to extend the TGLF quasi-linear transport model so that it can compute the energy and momentum fluxes due to parallel magnetic fluctuations, completing the transport matrix. The GFS equations do not use a bounce average approximation. The GFS equations are fully electromagnetic with general flux surface magnetic geometry, pitch angle scattering for electron collisions, and subsonic equilibrium toroidal rotation. Using GFS eigenmodes in the quasi-linear TGLF model will be shown to yield a more accurate match to fluxes computed by CGYRO turbulence simulations. In conclusion, prospects for future applications of a quasi-linear theory to new plasma transport regimes and magnetic confinement devices in addition to tokamaks are opened by the flexibility of the GFS eigensolver.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Spectral partitioning with blends of eigenvectors

We report many common methods for data analysis rely on linear algebra. We provide new results connecting data analysis error to numerical accuracy in the context of spectral graph partitioning. We provide pointwise convergence guarantees so that spectral blends (linear combinations of eigenvectors) can be employed to solve data analysis problems with confidence in their accuracy. We apply this theory to an accessible model problem, the ring of cliques, by deriving the relevant eigenpairs and finding necessary and sufficient solver tolerances. Analysis of the ring of cliques provides an upper bound on eigensolver tolerances for graph partitioning problems. These results bridge the gap between linear algebra based data analysis methods and the convergence theory of iterative approximation methods. These results explain how the combinatorial structure of a problem can be recovered much faster than numerically accurate solutions to the associated linear algebra problem.

97 MATHEMATICS AND COMPUTING↗

A fast implicit solver for semiconductor models in one space dimension

Several different approaches are proposed for solving fully implicit discretizations of a simplified Boltzmann-Poisson system with a linear relaxation-type collision kernel. This system models the evolution of free electrons in semiconductor devices under a low-density assumption. At each implicit time step, the discretized system is formulated as a fixed-point problem, which can then be solved with a variety of methods. A key algorithmic component in all the approaches considered here is a recently developed sweeping algorithm for Vlasov-Poisson systems. A synthetic acceleration scheme has been implemented to accelerate the convergence of iterative solvers by using the solution to a drift-diffusion equation as a preconditioner. The performance of four iterative solvers and their accelerated variants has been compared on problems modeling semiconductor devices with various electron mean-free-path.

97 MATHEMATICS AND COMPUTING↗

Algebraic multigrid for directed graph Laplacian linear systems (NS‐LAMG)

Summary We propose nonsymmetric lean algebraic multigrid (NS‐LAMG), a new algebraic multigrid algorithm for directed graph Laplacian systems that combines ideas from undirected graph Laplacian multigrid solvers and multigrid algorithms for Markov chain stationary distribution systems. Low‐degree elimination, proposed in LAMG for undirected graphs, is generalized to directed graphs and is a key component of NS‐LAMG. In the setup phase, we propose a simple stationary‐aggregation multigrid algorithms for Markov chain stationary distribution systems solver enhanced by low‐degree elimination to find the right null‐space vector that is used for the intergrid transfer operators. Numerical results show that low‐degree elimination improves performance and that NS‐LAMG outperforms generalized minimal residual method with restart and stable bi‐conjugate gradient method for real‐world, directed graph Laplacian linear systems.

Fox, Alyson↗

Parallel Interior Point Solver - Simple Branch & Bound

The purpose of this software package is to solve large-scale block-angular mixed-integer (linear) programs specified in SMPS format; the major application of this software package is to solving large-scale stochastic mixed-integer programs, which are block-angular.

Oxberry, Geoffrey↗

Eigenvalue Solver for Fluid and Kinetic Plasma Models in Arbitrary Magnetic Topology

ArbiTER (Arbitrary Topology Equation Reader) is a new code for solving linear eigenvalue problems arising from a broad range of physics and geometry models. The primary application area envisioned is boundary plasma physics in magnetic confinement devices; however ArbiTER should be applicable to other science and engineering fields as well. The code permits a variable numbers of dimensions, making possible application to both fluid and kinetic models. Additionally, the use of specialized equation and topology parsers permits a high degree of flexibility in specifying the physics and geometry.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Parallel Interior Point Solver - Simple Branch & Bound

The purpose of this software package is to solve large-scale block-angular mixed-integer (linear) programs specified in SMPS format; the major application of this software package is to solving large-scale stochastic mixed-integer programs, which are block-angular.

Oxberry, Geoffrey↗

Nalu-Wind and OpenFAST: A high-fidelity modeling and simulation environment for wind energy. Milestone ECP-Q2-FY19

The goal of the ExaWind project is to enable predictive simulations of wind farms comprised of many megawatt-scale turbines situated in complex terrain. Predictive simulations will require computational fluid dynamics (CFD) simulations for which the mesh resolves the geometry of the turbines and captures the rotation and large deflections of blades. Whereas such simulations for a single turbine are arguably petascale class, multi-turbine wind farm simulations will require exascale-class resources. The primary physics codes in the ExaWind project are Nalu-Wind, which is an unstructured-grid solver for the acoustically incompressible Navier-Stokes equations, and OpenFAST, which is a whole-turbine simulation code. The Nalu-Wind model consists of the mass-continuity Poisson-type equation for pressure and a momentum equation for the velocity. For such modeling approaches, simulation times are dominated by linear-system setup and solution for the continuity and momentum systems. For the ExaWind challenge problem, the moving meshes greatly affect overall solver costs as reinitialization of matrices and recomputation of preconditioners is required at every time step. This milestone represents the culmination of several parallel development activities towards the goal of establishing a full-physics simulation capability for modeling wind turbines operating in turbulent atmospheric inflow conditions. The demonstration simulation performed in this milestone is the first step towards the "ground truth" simulation and includes the following components: neutral atmospheric boundary layer inflow conditions generated using a precursor simulation, a hybrid RANS/LES simulation of the wall-resolved turbine geometry, hybridization of the turbulence equations using a blending function approach to transition from the atmospheric scales to the blade boundary layer scales near the turbine, fluid-structure interaction (FSI) that accounts for the complete set of blade deformations (bending, twisting and pitch motion, yaw and tower displacements) by coupling to a comprehensive turbine dynamics code (OpenFAST). The use of overset mesh methodology for the simulations in this milestone presents a significant deviation from the previous efforts where a sliding mesh approach was employed to model the rotation of the turbine blades. The choice of overset meshes was motivated by the need to handle arbitrarily large deformations of the blade and to allow for blade pitching in the presence of a controller and the ease of mesh generation compared to the sliding mesh approach. FSI and the new timestep algorithm used in the simulations were developed in partnership with the A2e High-Fidelity Modeling project. The individual physics components were verified and validated (V%V) through extensive code-to-code comparisons and with experiments where possible. The detailed V&V efforts provide confidence in the final simulation where these physics models were combined together even though no detailed experimental data is available to perform validation of the final configuration. Taken together, this milestone successfully demonstrated the most advanced simulation to date that has been performed with Nalu-Wind.

17 WIND ENERGY↗