SEARCH · Search NASA

Results for “Unstructured algorithm”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 199 records · Page 11

All sources NASA NTRS DOE OSTI

NASA NTRS1999

Parallel Visualization of Large-Scale Aerodynamics Calculations: A Case Study on the Cray T3E

This paper reports the performance of a parallel volume rendering algorithm for visualizing a large-scale, unstructured-grid dataset produced by a three-dimensional aerodynamics simulation. This dataset, containing over 18 million tetrahedra, allows us to extend our performance results to a problem which is more than 30 times larger than the one we examined previously. This high resolution dataset also allows us to see fine, three-dimensional features in the flow field. All our tests were performed on the Silicon Graphics Inc. (SGI)/Cray T3E operated by NASA's Goddard Space Flight Center. Using 511 processors, a rendering rate of almost 9 million tetrahedra/second was achieved with a parallel overhead of 26%.

Ma, Kwan-Liu↗

NASA NTRSDate not supplied

Performance and Portability of a Linear Solver Across Emerging Architectures

A linear solver algorithm used by a large-scale unstructured-grid computational fluid dynamics application is examined for a broad range of familiar and emerging architectures. Efficient implementation of a linear solver is challenging on recent CPUs offering vector architectures. Vector loads and stores are essential to effectively utilize available memory bandwidth on CPUs, and maintaining performance across different CPUs can be difficult in the face of varying vector lengths offered by each. A similar challenge occurs on GPU architectures, where it is essential to have coalesced memory accesses to utilize memory bandwidth effectively. In this work, we demonstrate that restructuring a computation, and possibly data layout, with regard to architecture is essential to achieve optimal performance by establishing a performance benchmark for each target architecture in a low level language such as vector intrinsics or CUDA. In doing so, we demonstrate how a linear solver kernel can be mapped to Intel® Xeon™ and Xeon Phi™, Marvell® ThunderX2®, NEC® SX-Aurora™ TSUBASA Vector Engine, and NVIDIA® and AMD® GPUs. We further demonstrate that the required code restructuring can be achieved in higher level programming environments such as OpenACC, OCCA, and Intel® OneAPI™/SYCL, and that each generally results in optimal performance on the target architecture. Relative performance metrics for all implementations are shown, and subjective ratings for ease of implementation and optimization are suggested.

Programming models↗

NASA NTRS2002

Parallel Computing Strategies for Irregular Algorithms

Parallel computing promises several orders of magnitude increase in our ability to solve realistic computationally-intensive problems, but relies on their efficient mapping and execution on large-scale multiprocessor architectures. Unfortunately, many important applications are irregular and dynamic in nature, making their effective parallel implementation a daunting task. Moreover, with the proliferation of parallel architectures and programming paradigms, the typical scientist is faced with a plethora of questions that must be answered in order to obtain an acceptable parallel implementation of the solution algorithm. In this paper, we consider three representative irregular applications: unstructured remeshing, sparse matrix computations, and N-body problems, and parallelize them using various popular programming paradigms on a wide spectrum of computer platforms ranging from state-of-the-art supercomputers to PC clusters. We present the underlying problems, the solution algorithms, and the parallel implementation strategies. Smart load-balancing, partitioning, and ordering techniques are used to enhance parallel performance. Overall results demonstrate the complexity of efficiently parallelizing irregular algorithms.

Biswas, Rupak↗

NASA NTRS1988

Finite element Euler computations in three dimensions

A two-step explicit FEM solution algorithm for the three-dimensional compressible Euler and Navier-Stokes equations based on unstructured triangular and tetrahedral grids is described and demonstrated. The method represents an extension and refinement of the algorithms presented by Loehner et al. (1984 and 1985), Peraire et al. (1987), and Morgan et al. (1987). The formulation and numerical implementation are outlined; the mesh generation, data structures, and adaptive remeshing are explained; and results for a two-dimensional airfoil, a three-dimensional engine air intake, a B747 in landing configuration, and a generic fighter aircraft are presented in extensive graphics and discussed in detail.

Peraire, Jaime↗

NASA NTRS2008

Rotor Airloads Prediction Using Unstructured Meshes and Loose CFD/CSD Coupling

The FUN3D unsteady Reynolds-averaged Navier-Stokes solver for unstructured grids has been modified to allow prediction of trimmed rotorcraft airloads. The trim of the rotorcraft and the aeroelastic deformation of the rotor blades are accounted for via loose coupling with the CAMRAD II rotorcraft computational structural dynamics code. The set of codes is used to analyze the HART-II Baseline, Minimum Noise and Minimum Vibration test conditions. The loose coupling approach is found to be stable and convergent for the cases considered. Comparison of the resulting airloads and structural deformations with experimentally measured data is presented. The effect of grid resolution and temporal accuracy is examined. Rotorcraft airloads prediction presents a very substantial challenge for Computational Fluid Dynamics (CFD). Not only must the unsteady nature of the flow be accurately modeled, but since most rotorcraft blades are not structurally stiff, an accurate simulation must account for the blade structural dynamics. In addition, trim of the rotorcraft to desired thrust and moment targets depends on both aerodynamic loads and structural deformation, and vice versa. Further, interaction of the fuselage with the rotor flow field can be important, so that relative motion between the blades and the fuselage must be accommodated. Thus a complete simulation requires coupled aerodynamics, structures and trim, with the ability to model geometrically complex configurations. NASA has recently initiated a Subsonic Rotary Wing (SRW) Project under the overall Fundamental Aeronautics Program. Within the context of SRW are efforts aimed at furthering the state of the art of high-fidelity rotorcraft flow simulations, using both structured and unstructured meshes. Structured-mesh solvers have an advantage in computation speed, but even though remarkably complex configurations may be accommodated using the overset grid approach, generation of complex structured-mesh systems can require months to set up. As a result, many rotorcraft simulations using structured-grid CFD neglect the fuselage. On the other hand, unstructured-mesh solvers are easily able to handle complex geometries, but suffer from slower execution speed. However, advances in both computer hardware and CFD algorithms have made previously state-of-the-art computations routine for unstructured-mesh solvers, so that rotorcraft simulations using unstructured grids are now viable. The aim of the present work is to develop a first principles rotorcraft simulation tool based on an unstructured CFD solver.

Biedron, Robert T.↗

NASA NTRS1990

An implicit finite element method for high speed flows

A fast algorithm is presented for constructing continuous lines, consisting of element sides, on general unstructured two-dimensional triangular meshes. The lines must pass through each node of the mesh once and only once. The discussion focuses on the use of these lines in a relaxation method for the solution of the equation system arising from an implicit algorithm for the solution of two-dimensional Euler and Navier-Stokes equations on general unstructured grids. It is also shown that the method can be used for solving three-dimensional Navier-Stokes equations on a grid composed of both structured and unstructured regions.

Hassan, O.↗

NASA NTRS1991

An efficient advancing front algorithm for Delaunay triangulation

There has been some recent interest in fluid dynamics calculations on unstructured meshes. One method of unstructured mesh generation involves Delaunay triangulation. This method has certain advantages but it can be expensive to implement. Furthermore, there can be problems with crossing grid lines near boundaries. A method shown here avoids many of the robustness and efficiency problems previously associated with Delaunay triangulation. As an added feature, a simple algorithm is shown which allows removal of diagonal edges from cells that are nearly rectangular. This can result in significant savings in the cost per iteration of a flow solver using this grid.

Merriam, Marshal L.↗

NASA NTRS1988

Discretization formulas for unstructured grids

The Galerkin weighted residual technique using linear triangular weight functions is employed to develop finite difference formula in cartesian coordinates for the Laplacian operator, first derivative operators and the function for unstructured triangular grids. The weighted residual coefficients associated with the weak formulation of the Laplacian operator are shown to agree with the Taylor series approach on a global average. In addition, a simple algorithm is presented to determine the Voronoi (finite difference) area of an unstructured grid.

Baumeister, Kenneth J.↗

NASA NTRS1988

Discretization formulas for unstructured grids

Baumeister, Kenneth J.↗

DOE OSTI2019

Compressing unstructured mesh data from simulations using machine learning

The amount of data output from a computer simulation has grown to terabytes and petabytes as increasingly complex simulations are being run on massively parallel systems. As we approach exaflop computing in the next decade, it is expected that the I/O subsystem will not be able to write out these large volumes of data. In this paper, we explore the use of machine learning to compress the data before it is written out. Despite the computational constraints that limit us to using very simple learning algorithms, our results show that machine learning is a viable option for compressing unstructured data. Furthermore, we demonstrate that by simply using a better sampling algorithm to generate the training set, we can obtain more accurate results compared to random sampling, but at no extra cost. Further, by carefully selecting and incorporating points with high prediction error, we can improve reconstruction accuracy without sacrificing the compression rate.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

PUMIPic: A mesh-based approach to unstructured mesh Particle-In-Cell on GPUs

Unstructured mesh particle-in-cell, PIC, simulations executing on the current and next generation of massively parallel systems require new methods for both the mesh and particles to achieve performance and scalability on GPUs. The traditional approach to implementing PIC simulations defines data structures and algorithms in terms of particles with a full copy of the unstructured mesh on every process. To effectively scale the unstructured mesh and particles, mesh-based PIC uses the unstructured mesh as the predominant data structure with the particles stored in terms of the mesh entities. Here, this paper details the PUMIPic library, a framework for developing efficient and performance-portable mesh-based PIC simulations on GPU systems. A pseudo physics simulation based on a five-dimensional gyro-kinetic code for modeling plasma physics is used to examine the performance of PUMIPic. Scaling studies of the unstructured mesh partition and number of particles are performed up to 4096 nodes of the Summit system at Oak Ridge National Laboratory. The studies show that mesh-based PIC can utilize a partitioned mesh and maintain scaling up to system limitations.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Parallel Randomized Tucker Decomposition Algorithms

The Tucker tensor decomposition is a natural extension of the singular value decomposition (SVD) to multiway data. Here, we propose to accelerate Tucker tensor decomposition algorithms by using randomization and parallelization. We present two algorithms that scale to large data and many processors, significantly reduce both computation and communication cost compared to previous deterministic and randomized approaches, and obtain nearly the same approximation errors. The key idea in our algorithms is to perform randomized sketches with Kronecker-structured random matrices, which reduces computation compared to unstructured matrices and can be implemented using a fundamental tensor computational kernel. We provide probabilistic error analysis of our algorithms and implement a new parallel algorithm for the structured randomized sketch. Our experimental results demonstrate that our combination of randomization and parallelization achieves accurate Tucker decompositions much faster than alternative approaches. We observe up to a 16X speedup over the fastest deterministic parallel implementation on 3D simulation data.

Tucker decompositions↗

NASA NTRS1991

Three dimensional unstructured multigrid for the Euler equations

The three dimensional Euler equations are solved on unstructured tetrahedral meshes using a multigrid strategy. The driving algorithm consists of an explicit vertex-based finite element scheme, which employs an edge-based data structure to assemble the residuals. The multigrid approach employs a sequence of independently generated coarse and fine meshes to accelerate the convergence to steady-state of the fine grid solution. Variables, residuals and corrections are passed back and forth between the various grids of the sequence using linear interpolation. The addresses and weights for interpolation are determined in a preprocessing stage using linear interpolation. The addresses and weights for interpolation are determined in a preprocessing stage using an efficient graph traversal algorithm. The preprocessing operation is shown to require a negligible fraction of the CPU time required by the overall solution procedure, while gains in overall solution efficiencies greater than an order of magnitude are demonstrated on meshes containing up to 350,000 vertices. Solutions using globally regenerated fine meshes as well as adaptively refined meshes are given.

Mavriplis, D. J.↗

NASA NTRS1991

Three dimensional unstructured multigrid for the Euler equations

The three-dimensional Euler equations are solved on unstructured tetrahedral meshes using a multigrid strategy. The driving algorithm consists of an explicit vertex-based finite-element scheme, which employs an edge-based data-structure to assemble the residuals. The multigrid approach employs a sequence of independently generated coarse and fine meshes to accelerate the convergence to steady-state of the fine grid solution. Variables, residuals and corrections are passed back and forth between the various grids of the sequence using linear interpolation. The addresses and weights for interpolation are determined in a preprocessing stage using an efficient graph traversal algorithm. The preprocessing operation is shown to require a negligible fraction of the CPU time required by the overall solution procedure, while gains in overall solution efficiencies greater than an order of magnitude are demonstrated on meshes containing up to 350,000 vertices. Solutions using globally regenerated fine meshes as well as adaptively refined meshes are given.

Mavriplis, D. J.↗

NASA NTRS1992

A nearly-monotone genuinely multidimensional scheme for the Euler equations

Recent progress made in the development of a genuinely multidimensional finite-volume method for the Euler equations is reported. In this method, wave information is derived from a reconstruction procedure applied to the data on a triangle. The algorithm is therefore best suited to application on an unstructured grid. A flux formula which provides nearly-monotone solutions is also presented. The new method is tested on three simple 2D flows, and the results show that dominant-wave resolution is high, and that strong-wave transitions are nearly monotone.

Parpia, Ijaz H.↗

NASA NTRS1994

Aerodynamic shape optimization of a HSCT type configuration with improved surface definition

Two distinct parametrization procedures of generating free-form surfaces to represent aerospace vehicles are presented. The first procedure is the representation using spline functions such as nonuniform rational b-splines (NURBS) and the second is a novel (geometrical) parametrization using solutions to a suitably chosen partial differential equation. The main idea is to develop a surface which is more versatile and can be used in an optimization process. Unstructured volume grid is generated by an advancing front algorithm and solutions obtained using an Euler solver. Grid sensitivity with respect to surface design parameters and aerodynamic sensitivity coefficients based on potential flow is obtained using an automatic differentiator precompiler software tool. Aerodynamic shape optimization of a complete aircraft with twenty four design variables is performed. High speed civil transport aircraft (HSCT) configurations are targeted to demonstrate the process.

Thomas, Almuttil M.↗

NASA NTRS2019

Advanced Astrophysics Discovery Technology in the Era of Data Driven Astronomy

Astrophysics is at the threshold of a new epoch in which increasinglycomplex, heterogeneous datasets will challenge our existing information infrastructure and traditional approaches to analysis. The rapid advancement of graphics processing units, compact field programmable gate arrays and dedicated artificial intelligence accelerator chips is now permitting the use of scientific methods, processes and algorithms to extract knowledge and insights from structured and unstructured data in ways never before seen. Miniaturization of spacecraft architectures and supporting infrastructure is opening new observing strategies and new discovery spaces for science. The community is just beginning to awaken to these imminent challenges as evidenced by their relative lack of emphasis in the New Worlds, New Horizons ASTRO2010 decadal survey, in the ExoPAG Science Analysis Group 11 report andin the formulation of the WFIRST Data Challenge. We suggest that the Astrophysics Science Division (ASD), which has clearly recognized this new epoch of rapidly evolving information technology, could be more affirmative in its approach. We offer a modest structural solution.

Barry, Richard K.↗

DOE OSTI2021

An efficient reconstruction algorithm for diffusion on triangular grids using the nodal discontinuous Galerkin method

High-energy-density (HED) hydrodynamics studies such as those relevant to inertial confinement fusion and astrophysics require highly disparate densities, temperatures, viscosities, and other diffusion parameters over relatively short spatial scales. This presents a challenge for high-order accurate methods to effectively resolve the hydrodynamics at these scales, particularly in the presence of highly disparate diffusion. A significant volume of engineering and physics applications use an unstructured discontinuous Galerkin (DG) method developed based on the finite element mesh generation and algorithmic framework. This work discusses the application of an affine reconstructed nodal DG method for unstructured grids of triangles. Solving the diffusion terms in the DG method is non-trivial due to the solution representations being piecewise continuous. Hence, the diffusive flux is not defined on the interface of elements. The proposed numerical approach reconstructs a smooth solution in a parallelogram that is enclosed by the quadrilateral formed by two adjacent triangle elements. The interface between these two triangles is the diagonal of the enclosed parallelogram. Similar to triangles, the mapping of parallelograms from a physical domain to a reference domain is an affine mapping, which is necessary for an accurate and efficient implementation of the numerical algorithm. Thus, all computations can still be performed on the reference domain, which promotes efficiency in computation and storage. This reconstruction does not make assumptions on choice of polynomial basis. Reconstructed DG algorithms have previously been developed for modal implementations of the convection–diffusion equations. However, to the best of the authors’ knowledge, this is the first practical guideline that has been proposed for applying the reconstructed algorithm on a nodal discontinuous Galerkin method with a focus on accuracy and efficiency. As a result, the algorithm is demonstrated on a number of benchmark cases as well as a challenging substantive problem in HED hydrodynamics with highly disparate diffusion parameters.

Computational efficiency↗