Search NASA⌕ Search

SEARCH · Search NASA

Results for “parallel 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.

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

At least 1,765 records · Page 98

Two-dimensional materials for bio-realistic neuronal computing networks

Two-dimensional (2D) van der Waals materials have found broad utility in a diverse range of applications including electronics, optoelectronics, renewable energy, and quantum information technologies. Meanwhile, exponentially growing digital data coupled with the ubiquity of artificial intelligence algorithms have generated significant interest in edge neuromorphic computing as an alternative to centralized cloud computing. The drive to incorporate neuroscience principles into computing hardware is motivated by the low power consumption, parallel processing, and reconfigurability of the human brain. The diverse library of 2D materials with atomic-level thicknesses, exceptional electrostatic tunability, and integration versatility is particularly well-suited for realizing bio-realistic synaptic and neuronal functionality. Here, we summarize past and present work in this field and outline the frontier challenges that have not yet been overcome. Here we also delineate potential solutions and suggest that the neuroscience principles of criticality and synchrony have the potential to inspire breakthrough applications of 2D materials in neuronal computing networks.

36 MATERIALS SCIENCE↗

Fast neural Poincaré maps for toroidal magnetic fields

Poincaré maps for toroidal magnetic fields are routinely employed to study gross confinement properties in devices built to contain hot plasmas. In most practical applications, evaluating a Poincaré map requires numerical integration of a magnetic field line, a process that can be slow and that cannot be easily accelerated using parallel computations. We propose a novel neural network architecture, the HénonNet, and show that it is capable of accurately learning realistic Poincaré maps from observations of a conventional field-line-following algorithm. After training, such learned Poincaré maps evaluate much faster than the field-line integration method. Moreover, the HénonNet architecture exactly reproduces the primary physics constraint imposed on field-line Poincaré maps: flux preservation. Furthermore, this structure-preserving property is the consequence of each layer in a HénonNet being a symplectic map. We demonstrate empirically that a HénonNet can learn to mock the confinement properties of a large magnetic island by using coiled hyperbolic invariant manifolds to produce a sticky chaotic region at the desired island location. This suggests a novel approach to designing magnetic fields with good confinement properties that may be more flexible than ensuring confinement using KAM tori.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Real-time dose computation: GPU-accelerated source modeling and superposition/convolution

Purpose: To accelerate dose calculation to interactive rates using highly parallel graphics processing units (GPUs). Methods: The authors have extended their prior work in GPU-accelerated superposition/convolution with a modern dual-source model and have enhanced performance. The primary source algorithm supports both focused leaf ends and asymmetric rounded leaf ends. The extra-focal algorithm uses a discretized, isotropic area source and models multileaf collimator leaf height effects. The spectral and attenuation effects of static beam modifiers were integrated into each source's spectral function. The authors introduce the concepts of arc superposition and delta superposition. Arc superposition utilizes separate angular sampling for the total energy released per unit mass (TERMA) and superposition computations to increase accuracy and performance. Delta superposition allows single beamlet changes to be computed efficiently. The authors extended their concept of multi-resolution superposition to include kernel tilting. Multi-resolution superposition approximates solid angle ray-tracing, improving performance and scalability with a minor loss in accuracy. Superposition/convolution was implemented using the inverse cumulative-cumulative kernel and exact radiological path ray-tracing. The accuracy analyses were performed using multiple kernel ray samplings, both with and without kernel tilting and multi-resolution superposition. Results: Source model performance was <9 ms (data dependent) for a high resolution (400{sup 2}) field using an NVIDIA (Santa Clara, CA) GeForce GTX 280. Computation of the physically correct multispectral TERMA attenuation was improved by a material centric approach, which increased performance by over 80%. Superposition performance was improved by {approx}24% to 0.058 and 0.94 s for 64{sup 3} and 128{sup 3} water phantoms; a speed-up of 101-144x over the highly optimized Pinnacle{sup 3} (Philips, Madison, WI) implementation. Pinnacle{sup 3} times were 8.3 and 94 s, respectively, on an AMD (Sunnyvale, CA) Opteron 254 (two cores, 2.8 GHz). Conclusions: The authors have completed a comprehensive, GPU-accelerated dose engine in order to provide a substantial performance gain over CPU based implementations. Real-time dose computation is feasible with the accuracy levels of the superposition/convolution algorithm.

60 APPLIED LIFE SCIENCES↗

Galactos: computing the anisotropic 3-point correlation function for 2 billion galaxies

The nature of dark energy and the complete theory of gravity are two central questions currently facing cosmology. A vital tool for addressing them is the 3-point correlation function (3PCF), which probes deviations from a spatially random distribution of galaxies. However, the 3PCF's formidable computational expense has prevented its application to astronomical surveys comprising millions to billions of galaxies. We present Galactos, a high-performance implementation of a novel, O( N 2 ) algorithm that uses a load-balanced k-d tree and spherical harmonic expansions to compute the anisotropic 3PCF. Our implementation is optimized for the Intel Xeon Phi architecture, exploiting SIMD parallelism, instruction and thread concurrency, and significant L1 and L2 cache reuse, reaching 39% of peak performance on a single node. Galactos scales to the full Cori system, achieving 9.8 PF (peak) and 5.06 PF (sustained) across 9636 nodes, making the 3PCF easily computable for all galaxies in the observable universe.

79 ASTRONOMY AND ASTROPHYSICS↗

Elastic Solutions to 2D Plane Strain Problems: Nonlinear Contact and Settlement Analysis for Shallow Foundations

The classical Neumann boundary value problem of an isotropic, homogeneous elastic half-plane under plane strain conditions is readdressed as the limiting case of the fully three-dimensional problem. Analytical solutions of the stress and strain tensors are obtained by taking the limit from known three-dimensional solutions. It is shown that the displacement fields for the plane strain problem are not well defined. A small number of simple expressions are developed, which provide a general solution for linearly-varying traction over arbitrary regions on the boundary. A simple, efficient, and rapidly convergent algorithm is developed which uses these solutions as analytic elements and provides a solution approach to the general boundary value problem. The method is verified against known solutions for Hertzian contact between parallel cylinders. Two numerical examples are presented for the analysis of shallow foundation systems. In the first, the boundary conditions are informed by analytical elastoplastic calculations and a strain influence analysis is performed and compared with the Schmertmann method. Subsequently, empirical laboratory contact traction distributions measured by Bauer et al., in both the normal and tangential directions are employed as boundary conditions for an analysis of the underlying stress field.

42 ENGINEERING↗

Performance-Portable Particle Advection with VTK-m

Particle advection is the fundamental kernel behind most vector field visualization methods. Yet, the efficient parallel computation of large amounts of particle traces remains challenging. This is exacerbated by the variety of hardware trends in today's HPC arena, including increasing core counts in classical CPUs, many-core designs such as the Intel Xeon Phi, and massively parallel GPUs. The dedicated optimization of a particle advection kernel for each individual target architecture is both time-consuming and error prone. In this paper, we propose a performance-portable algorithm for particle advection. Our algorithm is based on the recently introduced VTK-m system and chiefly relies on its device adapter abstraction. We demonstrate the general portability of our implementation across a wide variety of hardware. Finally, our evaluation shows that our hardware-agnostic algorithm has comparable performance to hardware-specific algorithms.

Pugmire, Dave↗

Unitary qubit lattice algorithm for three-dimensional vortex solitons in hyperbolic self-defocusing media

To obtain stable vortex structures in three-dimensional (3D) nonlinear optics, Efremidis et al. (2007) have introduced a generalized Nonlinear Schrodinger equation (NLS) in which the transverse quantum vortex components are stabilized by a longitudinal bright soliton: i.e., the elliptic operator ∇ 2 is replaced by its hyperbolic counterpart ∇ ⊥ 2 - ∂ 2 / ∂ z 2 . A new 3D mesoscopic qubit unitary lattice algorithm is developed for this generalized NLS. One introduces 2 qubits for each lattice site and entangles them with a local unitary collision operator. This entanglement is then spread throughout the lattice by nearest neighbor streaming. These interwined operators lead to an extremely well parallelized code on classical supcomputers while their unitary structure will permit encoding onto a quantum computer. Somewhat unexpectedly, the hyperbolic operator can be realized from variations in the collision operator, without introducing variations in the streaming operator. The initial line vortices are generated by Pade asymptotics. The energy constraint is conserved to 10 digit accuracy.

3D solitons↗

A Novel Coarsening Method for Scalable and Efficient Mesh Generation

In this paper, we propose a novel mesh coarsening method called brick coarsening method. The proposed method can be used in conjunction with any graph partitioners and scales to very large meshes. This method reduces problem space by decomposing the original mesh into fixed-size blocks of nodes called bricks, layered in a similar way to conventional brick laying, and then assigning each node of the original mesh to appropriate brick. Our experiments indicate that the proposed method scales to very large meshes while allowing simple RCB partitioner to produce higher-quality partitions with significantly less edge cuts. Our results further indicate that the proposed brick-coarsening method allows more complicated partitioners like PT-Scotch to scale to very large problem size while still maintaining good partitioning performance with relatively good edge-cut metric. Graph partitioning is an important problem that has many scientific and engineering applications in such areas as VLSI design, scientific computing, and resource management. Given a graph G = (V,E), where V is the set of vertices and E is the set of edges, (k-way) graph partitioning problem is to partition the vertices of the graph (V) into k disjoint groups such that each group contains roughly equal number of vertices and the number of edges connecting vertices in different groups is minimized. Graph partitioning plays a key role in large scientific computing, especially in mesh-based computations, as it is used as a tool to minimize the volume of communication and to ensure well-balanced load across computing nodes. The impact of graph partitioning on the reduction of communication can be easily seen, for example, in different iterative methods to solve a sparse system of linear equation. Here, a graph partitioning technique is applied to the matrix, which is basically a graph in which each edge is a non-zero entry in the matrix, to allocate groups of vertices to processors in such a way that many of matrix-vector multiplication can be performed locally on each processor and hence to minimize communication. Furthermore, a good graph partitioning scheme ensures the equal amount of computation performed on each processor. Graph partitioning is a well known NP-complete problem, and thus the most commonly used graph partitioning algorithms employ some forms of heuristics. These algorithms vary in terms of their complexity, partition generation time, and the quality of partitions, and they tend to trade off these factors. A significant challenge we are currently facing at the Lawrence Livermore National Laboratory is how to partition very large meshes on massive-size distributed memory machines like IBM BlueGene/P, where scalability becomes a big issue. For example, we have found that the ParMetis, a very popular graph partitioning tool, can only scale to 16K processors. An ideal graph partitioning method on such an environment should be fast and scale to very large meshes, while producing high quality partitions. This is an extremely challenging task, as to scale to that level, the partitioning algorithm should be simple and be able to produce partitions that minimize inter-processor communications and balance the load imposed on the processors. Our goals in this work are two-fold: (1) To develop a new scalable graph partitioning method with good load balancing and communication reduction capability. (2) To study the performance of the proposed partitioning method on very large parallel machines using actual data sets and compare the performance to that of existing methods. The proposed method achieves the desired scalability by reducing the mesh size. For this, it coarsens an input mesh into a smaller size mesh by coalescing the vertices and edges of the original mesh into a set of mega-vertices and mega-edges. A new coarsening method called brick algorithm is developed in this research. In the brick algorithm, the zones in a given mesh are first grouped into fixed size blocks called bricks. These brick are then laid in a way similar to conventional brick laying technique, which reduces the number of neighboring blocks each block needs to communicate. Contributions of this research are as follows: (1) We have developed a novel method that scales to a really large problem size while producing high quality mesh partitions; (2) We measured the performance and scalability of the proposed method on a machine of massive size using a set of actual large complex data sets, where we have scaled to a mesh with 110 million zones using our method. To the best of our knowledge, this is the largest complex mesh that a partitioning method is successfully applied to; and (3) We have shown that proposed method can reduce the number of edge cuts by as much as 65%.

97 MATHEMATICS AND COMPUTING↗

A second-order distributed memory parallel fast sweeping method for the Eikonal equation

The Eikonal equation is used to calculate wave propagation and distance fields, and due to its complexity requires numerical treatment for its solution. In this work, we present a second-order distributed memory parallel fast sweeping method. The second-order solution switches on a two-point stencil when two upwind points are available, and reverts to first-order otherwise. In all examples, the second-order method improves the solution over the first-order, allowing for significant savings in memory while achieving the same accuracy. Parallelization over distributed memory saw good weak scaling with optimal convergence. The computational time for second-order was approximately 2.5 times slower than first-order, where the largest amount of mesh points ran on 144 cores (512 GB) was ≈20 billion. The savings in memory from the second-order method combined with the distributed memory algorithm result in the ability to solve problems much larger than are possible with the serial first-order method.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Development of Large-Dimension Configuration-Interaction Shell-Model Code

The project period was devoted to several developments in the technical capabilities of the BIGSTIC large-dimension configuration-interaction shell-model code, written in Fortran 90. The specific computational goals for the project period were: (1) store Lanczos vectors on core in RAM to minimize I/O; (2) rewrite reorthogonalization with Lanczos vectors stored in core, consult with personnel at LLNL, LBL, ORNL, Iowa State University to maximize performance; (3) restrict creation of N-body jumps to those needed by an individual node; and (4) distribute 3-body interaction over many cores. Significant progress was made towards these goals, especially (1) and (2), although in the process they discovered intermediate tasks that had to be accomplished first. The achievements were as follows - I put into place structures and algorithms to facility fragmenting very large-dimension Lanczos intermediate vectors. Only by fragmenting the vectors can we carry out (1) and (2). In addition, I reorganized the action of the Hamiltonian matrix and created a new division of operations for MPI. Based upon earlier work, I made plans of a revised algorithm for distribution of work with MPI, with a particular eye towards breaking up the Lanczos vectors. I introduce a new derived type (opbundles) which collects the parameters for the Hamiltonian, and rewrote the application routines to use it. It has been validated and verified. I made progress towards revised MPI parallelization. Using the opbundles, I was able to compute a distribution of work over compute nodes, which should be very efficient. This new distribution is easier to derive and more efficient, in principle, than the old distribution. Furthermore, it should make applications with fragmented Lanczos vectors easier. Implementation is still in progress.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Using architecture information and real-time resource state to reduce power consumption and communication costs in parallel applications.

As computer systems grow in both size and complexity, the need for applications and run-time systems to adjust to their dynamic environment also grows. The goal of the RAAMP LDRD was to combine static architecture information and real-time system state with algorithms to conserve power, reduce communication costs, and avoid network contention. We devel- oped new data collection and aggregation tools to extract static hardware information (e.g., node/core hierarchy, network routing) as well as real-time performance data (e.g., CPU uti- lization, power consumption, memory bandwidth saturation, percentage of used bandwidth, number of network stalls). We created application interfaces that allowed this data to be used easily by algorithms. Finally, we demonstrated the benefit of integrating system and application information for two use cases. The first used real-time power consumption and memory bandwidth saturation data to throttle concurrency to save power without increasing application execution time. The second used static or real-time network traffic information to reduce or avoid network congestion by remapping MPI tasks to allocated processors. Results from our work are summarized in this report; more details are available in our publications [2, 6, 14, 16, 22, 29, 38, 44, 51, 54].

Brandt, James M.↗

PAKAL: A THREE-DIMENSIONAL MODEL TO SOLVE THE RADIATIVE TRANSFER EQUATION

We present a new numerical model called 'Pakal' intended to solve the radiative transfer equation in a three-dimensional (3D) geometry, using the approximation for a locally plane-parallel atmosphere. Pakal uses pre-calculated radial profiles of density and temperature (based on hydrostatic, hydrodynamic, or MHD models) to compute the emission from 3D source structures with high spatial resolution. Then, Pakal solves the radiative transfer equation in a set of (3D) ray paths, going from the source to the observer. Pakal uses a new algorithm to compute the radiative transfer equation by using an intelligent system consisting of three structures: a cellular automaton; an expert system; and a program coordinator. The code outputs can be either two-dimensional maps or one-dimensional profiles, which reproduce the observations with high accuracy, giving detailed physical information about the environment where the radiation was generated and/or transmitted. We present the model applied to a 3D solar radial geometry, assuming a locally plane-parallel atmosphere, and thermal free-free radio emission from hydrogen-helium gas in thermodynamic equilibrium. We also present the convergence test of the code. We computed the synthetic spectrum of the centimetric-millimetric solar emission and found better agreement with observations (up to 10{sup 4} K at 20 GHz) than previous models reported in the literature. The stability and convergence test show the high accuracy of the code. Finally, Pakal can improve the integration time by up to an order of magnitude compared against linear integration codes.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Spatial and rotational quality assurance of 6DOF patient tracking systems

Purpose: External tracking systems used for patient positioning and motion monitoring during radiotherapy are now capable of detecting both translations and rotations. In this work, the authors develop a novel technique to evaluate the 6 degree of freedom 6(DOF) (translations and rotations) performance of external motion tracking systems. The authors apply this methodology to an infrared marker tracking system and two 3D optical surface mapping systems in a common tumor 6DOF workspace. Methods: An in-house designed and built 6DOF parallel kinematics robotic motion phantom was used to perform motions with sub-millimeter and subdegree accuracy in a 6DOF workspace. An infrared marker tracking system was first used to validate a calibration algorithm which associates the motion phantom coordinate frame to the camera frame. The 6DOF positions of the mobile robotic system in this space were then tracked and recorded independently by an optical surface tracking system after a cranial phantom was rigidly fixed to the moveable platform of the robotic stage. The calibration methodology was first employed, followed by a comprehensive 6DOF trajectory evaluation, which spanned a full range of positions and orientations in a 20 × 20 × 16 mm and 5° × 5° × 5° workspace. The intended input motions were compared to the calibrated 6DOF measured points. Results: The technique found the accuracy of the infrared (IR) marker tracking system to have maximal root-mean square error (RMSE) values of 0.18, 0.25, 0.07 mm, 0.05°, 0.05°, and 0.09° in left–right (LR), superior–inferior (SI), anterior–posterior (AP), pitch, roll, and yaw, respectively, comparing the intended 6DOF position and the measured position by the IR camera. Similarly, the 6DOF RSME discrepancy for the HD optical surface tracker yielded maximal values of 0.46, 0.60, 0.54 mm, 0.06°, 0.11°, and 0.08° in LR, SI, AP, pitch, roll, and yaw, respectively, over the same 6DOF evaluative workspace. An earlier generation 3D optical surface tracking unit was observed to have worse tracking capabilities than both the IR camera unit and the newer 3D surface tracking system with maximal RMSE of 0.69, 0.74, 0.47 mm, 0.28°, 0.19°, and 0.18°, in LR, SI, AP, pitch, roll, and yaw, respectively, in the same 6DOF evaluation space. Conclusions: The proposed technique was found to be effective at evaluating the performance of 6DOF patient tracking systems. All observed optical tracking systems were found to exhibit tracking capabilities at the sub-millimeter and subdegree level within a 6DOF workspace.

60 APPLIED LIFE SCIENCES↗

DIF3D-VARIANT 12.0: Updates and New Features

The DIF3D code has been a workhorse of fast reactor analysis work at Argonne National Laboratory for over 40 years. In 1995, a transport option called VARIANT was added to DIF3D to improve the flux solutions for fast reactor problems which we term DIF3D-VARIANT today. DIF3D-VARIANT performs nodal neutron transport calculations using P N or SP N theory in Cartesian and hexagonal two- and three-dimensional geometries. The limited computing capabilities of the time restricted DIF3D-VARIANT to use at most a 6 th order spatial approximation combined with a P3 flux approximation and P1 scattering kernel for a 33 group structure on most studied reactor problems. Computer capabilities have increased steadily since 1995 and today much larger space-angle-energy approximations are possible. This manuscript serves as an update to the theory section of the original DIF3D-VARIANT manual and details more than twenty years of changes made to DIF3D to make version 12 which was released on November 1 st , 2024. The primary focus of the initial work was to extend the space-angle approximations available in DIF3D-VARIANT such that the error due to transport approximations could be better understood. This work was started and completed in 2002 and marked the official version 10. Unfortunately, those higher order approximations could not be used at that time due to the memory constraints of the BPOINTER part of DIF3D (limited to 2 GB). In version 11, completed in 2012, BPOINTER was circumvented in DIF3D-VARIANT for the largest arrays by introducing a Fortran 90 module called LMA (Large Memory Array). This seamlessly replaces all of the functionality of the BPOINTER concept, but it allows 64 bit addressing for every array such that they can be larger than 2 GB. It is now common for DIF3D-VARIANT jobs to consume 50 GB of memory on modern workstations when using high order space-angle approximations and a large number of groups. Many improvements were made to version 11 from 2012 to 2022 when work to create version 12 started. For version 12, several parts of DIF3D were updated to improve performance and thread parallelism was introduced to further reduce the runtime. Numerous minor bugs were discovered in DIF3D-VARIANT as part of the process of creating the perturbation and sensitivity code PERSENT. All of these algorithmic problems were identified in the transition from version 10 to version 11 which prevented DIF3D-VARIANT from running efficiently and reliably. Firstly, the coarse mesh rebalance scheme would routinely diverge and a study detailed in this report demonstrates how it was also typically not effective. This is not a failure of the coarse mesh rebalance methodology, but a failure of its implementation in DIF3D-VARIANT for hexagonal geometries. The fission source extrapolation algorithm was also found to be unreliable on larger group structure problems, leading to divergence in some cases and a negligible improvement in performance overall. Finally, the “Omega” acceleration applied to the partial current solver routine of DIF3D-VARIANT was found to cause DIF3D-VARIANT to converge to the wrong answer. To resolve these issues, both the coarse mesh rebalance and fission source extrapolation were permanently disabled in version 11. The Tchebychev acceleration was put in as a temporary reliable alternative but it is generally inferior to coarse mesh rebalance or coarse mesh finite difference. For the Omega acceleration, the factor was restricted to guarantee that it would not cause follow-on errors in PERSENT. Due to limited funding to support maintenance and development of DIF3D in the last 10 years, no effort was spent since to resolve the outer iteration acceleration. Except for the threading work, all of the changes discussed in this manuscript refer to changes made between version 10 and version 11. Performance comparisons are done to demonstrate the improvements from version 9 to version 12. As will be demonstrated, the updated versi

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Formalized Configurational Calculations of Parallel Mechanisms

Theoretically possible configurations of parallel mechanisms are classified, taking account of the hinge arrangement at the base and the moving platform. This classification permits the formulation of computational templates for each class of configurations. The associated computational algorithms require a minimum of initial data.

42 ENGINEERING↗

Optimizing Grain Boundary Structures with LAMMPS Using Evolutionary Algorithms

Grain boundary structure optimization is an important part of materials modeling. Current methods for grain boundary structure optimization involve inefficient, time-consuming processes that do not fully explore the interface parameter space. Evolutionary algorithms have recently been demonstrated to be effective at determining both stable and metastable grain boundary interface structures. In this work, we demonstrate the use of GBOpt, a grain boundary structure optimization software designed to use the Large-scale Atomic/Molecular Massively Parallel Simulation (LAMMPS) software to efficiently determine grain boundary structures. We demonstrate that a only a few manipulations, namely atom insertion, atom removal, and relative grain displacement, are sufficient to explore much of the grain boundary structure parameter space. The efficacy of this approach is demonstrated on an FCC Ni system, and a BCC Fe system. The computational cost is compared against the gamma-surface sampling approach to demonstrate performance improvement.

Evolutionary algorithms↗

Parallel Multivariate Spatio-Temporal Clustering of Large Ecological Datasets on Hybrid Supercomputers

A proliferation of data from vast networks of remote sensing platforms (satellites, unmanned aircraft systems (UAS), airborne etc.), observational facilities (meteorological, eddy covariance etc.), state-of-the-art sensors, and simulation models offer unprecedented opportunities for scientific discovery. Unsupervised classification is a widely applied data mining approach to derive insights from such data. However, classification of very large data sets is a complex computational problem that requires efficient numerical algorithms and implementations on high performance computing (HPC) platforms. Additionally, increasing power, space, cooling and efficiency requirements has led to the deployment of hybrid supercomputing platforms with complex architectures and memory hierarchies like the Titan system at Oak Ridge National Laboratory. The advent of such accelerated computing architectures offers new challenges and opportunities for big data analytics in general and specifically, large scale cluster analysis in our case. Although there is an existing body of work on parallel cluster analysis, those approaches do not fully meet the needs imposed by the nature and size of our large data sets. Moreover, they had scaling limitations and were mostly limited to traditional distributed memory computing platforms. We present a parallel Multivariate Spatio-Temporal Clustering (MSTC) technique based on k-means cluster analysis that can target hybrid supercomputers like Titan. We developed a hybrid MPI, CUDA and OpenACC implementation that can utilize both CPU and GPU resources on computational nodes. We describe performance results on Titan that demonstrate the scalability and efficacy of our approach in processing large ecological data sets.

Sreepathi, Sarat↗