Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Parallel algorithms.”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

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

At least 199 records · Page 11

A non-hybrid method for the PDF equations of turbulent flows on unstructured grids

In probability density function (PDF) methods of turbulent flows, the joint PDF of several flow variables is computed by numerically integrating a system of stochastic differential equations for Lagrangian particles. A set of parallel algorithms is proposed to provide an efficient solution of the PDF transport equation modeling the joint PDF of turbulent velocity, frequency and concentration of a passive scalar in geometrically complex configurations. In the vicinity of walls the flow is resolved by an elliptic relaxation technique down to the viscous sublayer, explicitly modeling the high anisotropy and inhomogeneity of the low-Reynolds-number wall region without damping or wall-functions. An unstructured Eulerian grid is employed to extract Eulerian statistics, to solve for quantities represented at fixed locations of the domain (i.e., the mean pressure and the elliptic relaxation tensor) and to track particles. All three aspects regarding the grid make use of the finite element method employing the simplest linear shapefunctions. To model the small-scale mixing of the transported scalar, the interaction by exchange with the conditional mean (IECM) model is adopted. An adaptive algorithm to compute the velocity-conditioned scalar mean is proposed that homogenizes the statistical error over the sample space with no assumption on the shape of the underlying velocity PDF. Compared to other hybrid particle-in-cell approaches for the PDF equations, the current methodology is consistent without the need for consistency conditions. The algorithm is tested by computing the dispersion of passive scalars released from concentrated sources in two different turbulent flows: the fully developed turbulent channel flow and a street canyon (or cavity) flow. Algorithmic details on estimating conditional and unconditional statistics, particle tracking and particle-number control are presented in detail. Relevant aspects of performance and parallelism on cache-based shared memory machines are discussed.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

ATDM LANL FleCSI: Topology and Execution Framework

FleCSI is a compile-time configurable C++ framework designed to support multi-physics application development. As such, FleCSI attempts to provide a very general set of infrastructure design patterns that can be specialized and extended to suit the needs of a broad variety of solver and data requirements. This means that FleCSI is potentially useful to many different ECP projects. Current support includes multidimensional mesh topology, mesh geometry, and mesh adjacency information, n-dimensional hashed-tree data structures, graph partitioning interfaces, and dependency closures (to identify data dependencies between distributed-memory address spaces). FleCSI introduces a functional programming model with control, execution, and data abstractions that are consistent with state-of-the-art task-based runtimes such as Legion and Charm++. The model also provides support for fine-grained, data-parallel execution with backend support for runtimes such as OpenMP and C++17. The FleCSI abstraction layer provides the developer with insulation from the underlying runtimes, while allowing support for multiple runtime systems, including conventional models like asynchronous MPI. The intent is to give developers a concrete set of user-friendly programming tools that can be used now, while allowing flexibility in choosing runtime implementations and optimizations that can be applied to architectures and runtimes that arise in the future. This project is essential to the ECP Ristra Next-Generation Code project, part of ASC ATDM, because it provides a hierarchically parallel programming model that is consistent with the design of modern system architectures, but which allows for the straightforward expression of algorithmic parallelism in a portably performant manner.

97 MATHEMATICS AND COMPUTING↗

Fast Detection of Material Deformation through Structural Dissimilarity

Designing materials that are resistant to extreme temperatures and brittleness relies on assessing structural dynamics of samples. Algorithms are critically important to characterize material deformation under stress conditions. Here, we report on our design of coarse-grain parallel algorithms for image quality assessment based on structural information and on crack detection of gigabyte-scale experimental datasets. We show how key steps can be decomposed into distinct processing flows, one based on structural similarity (SSIM) quality measure, and another on spectral content. These algorithms act upon image blocks that fit into memory, and can execute independently. We discuss the scientific relevance of the problem, key developments, and decomposition of complementary tasks into separate executions. We show how to apply SSIM to detect material degradation, and illustrate how this metric can be allied to spectral analysis for structure probing, while using tiled multi-resolution pyramids stored in HDF5 chunked multi-dimensional arrays. Results show that the proposed experimental data representation supports an average compression rate of 10X, and data compression scales linearly with the data size. We also illustrate how to correlate SSIM to crack formation, and how to use our numerical schemes to enable fast detection of deformation from 3D datasets evolving in time.

Ushizima, Daniela↗

The Kokkos OpenMPTarget Backend: Implementation and Lessons Learned

As the supercomputing landscape diversifies, solutions such as Kokkos to write vendor agnostic applications and libraries have risen in popularity. Kokkos provides a programming model designed for performance portability, which allows developers to write a single source implementation that can run efficiently on various architectures. At its heart, Kokkos maps parallel algorithms to architecture and vendor specific backends written in lower level programming models such as CUDA and HIP. Another approach to writing vendor agnostic parallel code is using OpenMP’s directives based approach, which lets developers annotate code to express parallelism. It is implemented at the compiler level and is supported by all major high performance computing vendors, as well as the primary Open Source toolchains GNU and LLVM. Since its inception, Kokkos has used OpenMP to parallelize on CPU architectures. In this paper, we explore leveraging OpenMP for a GPU backend and discuss the challenges we encountered when mapping the Kokkos APIs and semantics to OpenMP target constructs. As an exemplar workload we chose a simple conjugate gradient solver for sparse matrices. We find that performance on NVIDIA and AMD GPUs varies widely based on details of the implementation strategy and the chosen compiler. Furthermore, the performance of the OpenMP implementations decreases with increasing complexity of the investigated algorithms.

Gayatri, Rahulkumar↗

Programmability of the HPCS Languages: A Case Study with a Quantum Chemistry Kernel

As high-end computer systems present users with rapidly increasing numbers of processors, possibly also incorporating attached co-processors, programmers are increasingly challenged to express the necessary levels of concurrency with the dominant parallel programming model, Fortran+MPI+OpenMP (or minor variations). In this paper, we examine the languages developed under the DARPA High-Productivity Computing Systems (HPCS) program (Chapel, Fortress, and X10) as representatives of a different parallel programming model which might be more effective on coming high-performance systems. The application used in this study is the Hartree-Fock method from quantum chemistry, which combines access to distributed data with a task-parallel algorithm and is characterized by significant irregularity in the computational tasks. We present several different implementation strategies for load balancing of the task parallel computation, as well as distributed array operations, in each of the three languages. We conclude that the HPCS languages provide a wide variety of mechanisms for expressing parallelism, which can be combined at multiple levels, making them quite expressive for this problem.

Shet, Aniruddha G↗

Solvers for $\mathcal{O} (N)$ Electronic Structure in the Strong Scaling Limit

Here we present a hybrid OpenMP/Charm\tt++ framework for solving the $\mathcal{O} (N)$ self-consistent-field eigenvalue problem with parallelism in the strong scaling regime, $P\gg{N}$, where $P$ is the number of cores, and $N$ is a measure of system size, i.e., the number of matrix rows/columns, basis functions, atoms, molecules, etc. This result is achieved with a nested approach to spectral projection and the sparse approximate matrix multiply [Bock and Challacombe, SIAM J. Sci. Comput., 35 (2013), pp. C72--C98], and involves a recursive, task-parallel algorithm, often employed by generalized $N$-Body solvers, to occlusion and culling of negligible products in the case of matrices with decay. Lastly, employing classic technologies associated with generalized $N$-Body solvers, including overdecomposition, recursive task parallelism, orderings that preserve locality, and persistence-based load balancing, we obtain scaling beyond hundreds of cores per molecule for small water clusters ([H${}_2$O]${}_N$, $N \in \{ 30, 90, 150 \}$, $P/N \approx \{ 819, 273, 164 \}$) and find support for an increasingly strong scalability with increasing system size $N$.

97 MATHEMATICS AND COMPUTING↗

Parallel computation of multigroup reactivity coefficient using iterative method

One of the research activities to support the commercial radioisotope production program is a safety research target irradiation FPM (Fission Product Molybdenum). FPM targets form a tube made of stainless steel in which the nuclear degrees of superimposed high-enriched uranium. FPM irradiation tube is intended to obtain fission. The fission material widely used in the form of kits in the world of nuclear medicine. Irradiation FPM tube reactor core would interfere with performance. One of the disorders comes from changes in flux or reactivity. It is necessary to study a method for calculating safety terrace ongoing configuration changes during the life of the reactor, making the code faster became an absolute necessity. Neutron safety margin for the research reactor can be reused without modification to the calculation of the reactivity of the reactor, so that is an advantage of using perturbation method. The criticality and flux in multigroup diffusion model was calculate at various irradiation positions in some uranium content. This model has a complex computation. Several parallel algorithms with iterative method have been developed for the sparse and big matrix solution. The Black-Red Gauss Seidel Iteration and the power iteration parallel method can be used to solve multigroup diffusion equation system and calculated the criticality and reactivity coeficient. This research was developed code for reactivity calculation which used one of safety analysis with parallel processing. It can be done more quickly and efficiently by utilizing the parallel processing in the multicore computer. This code was applied for the safety limits calculation of irradiated targets FPM with increment Uranium.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

A Massively Parallel Implementation of the CCSD(T) Method Using the Resolution-of-the-Identity Approximation and a Hybrid Distributed/Shared Memory Parallelization Model

In this work, a parallel algorithm is described for the coupled-cluster singles and doubles method augmented with a perturbative correction for triple excitations [CCSD(T)] using the resolution-of-the-identity (RI) approximation for two-electron repulsion integrals (ERIs). The algorithm bypasses the storage of four-center ERIs by adopting an integral-direct strategy. The CCSD amplitude equations are given in a compact quasi-linear form by factorizing them in terms of amplitude-dressed three-center intermediates. A hybrid MPI/OpenMP parallelization scheme is employed, which uses the OpenMP-based shared memory model for intranode parallelization and the MPI-based distributed memory model for internode parallelization. Parallel efficiency has been optimized for all terms in the CCSD amplitude equations. Two different algorithms have been implemented for the rate-limiting terms in the CCSD amplitude equations that entail and -scaling computational costs, where N O and N V denote the number of correlated occupied and virtual orbitals, respectively. One of the algorithms assembles the four-center ERIs requiring N V 4 and N O 2 N V 2 -scaling memory costs in a distributed manner on a number of MPI ranks, while the other algorithm completely bypasses the assembling of quartic memory-scaling ERIs and thus largely reduces the memory demand. It is demonstrated that the former memory-expensive algorithm is faster on a few hundred cores, while the latter memory-economic algorithm shows a better strong scaling in the limit of a few thousand cores. The program is shown to exhibit a near-linear scaling, in particular for the compute-intensive triples correction step, on up to 8000 cores. The performance of the program is demonstrated via calculations involving molecules with 24–51 atoms and up to 1624 atomic basis functions. As the first application, the complete basis set (CBS) limit for the interaction energy of the π-stacked uracil dimer from the S66 data set has been investigated. This work reports the first calculation of the interaction energy at the CCSD(T)/aug-cc-pVQZ level without local orbital approximation. The CBS limit for the CCSD correlation contribution to the interaction energy was found to be -8.01 kcal/mol, which agrees very well with the value -7.99 kcal/mol reported by Schmitz, Hättig, and Tew [ Phys. Chem. Chem. Phys. 2014 , 16 , 22167-22178]. The CBS limit for the total interaction energy was estimated to be -9.64 kcal/mol.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Task Parallel Incomplete Cholesky Factorization using 2D Partitioned-Block Layout

We introduce a task-parallel algorithm for sparse incomplete Cholesky factorization that utilizes a 2D sparse partitioned-block layout of a matrix. Our factorization algorithm follows the idea of algorithms-by-blocks by using the block layout. The algorithm-byblocks approach induces a task graph for the factorization. These tasks are inter-related to each other through their data dependences in the factorization algorithm. To process the tasks on various manycore architectures in a portable manner, we also present a portable tasking API that incorporates different tasking backends and device-specific features using an open-source framework for manycore platforms i.e., Kokkos. A performance evaluation is presented on both Intel Sandybridge and Xeon Phi platforms for matrices from the University of Florida sparse matrix collection to illustrate merits of the proposed task-based factorization. Experimental results demonstrate that our task-parallel implementation delivers about 26.6x speedup (geometric mean) over single-threaded incomplete Choleskyby- blocks and 19.2x speedup over serial Cholesky performance which does not carry tasking overhead using 56 threads on the Intel Xeon Phi processor for sparse matrices arising from various application problems.

2D layout↗

Interim report on fuel cycle neutronics code development.

As part of the Global Nuclear Energy Partnership (GNEP), a fast reactor simulation program was launched in April 2007 to develop a suite of modern simulation tools specifically for the analysis and design of sodium cooled fast reactors. The general goal of the new suite of codes is to reduce the uncertainties and biases in the various areas of reactor design activities by enhanced prediction capabilities. Under this fast reactor simulation program, a high-fidelity deterministic neutron transport code named UNIC is being developed. The final objective is to produce an integrated, advanced neutronics code that allows the high fidelity description of a nuclear reactor and simplifies the multi-step design process by direct coupling with thermal-hydraulics and structural mechanics calculations. Currently there are three solvers for the neutron transport code incorporated in UNIC: PN2ND, SN2ND, and MOCFE. PN2ND is based on a second-order even-parity spherical harmonics discretization of the transport equation and its primary target area of use is the existing homogenization approaches that are prevalent in reactor physics. MOCFE is based upon the method of characteristics applied to an unstructured finite element mesh and its primary target area of use is the fine grained nature of the explicit geometrical problems which is the long term goal of this project. SN2ND is based on a second-order, even-parity discrete ordinates discretization of the transport equation and its primary target area is the modeling transition region between the PN2ND and MOCFE solvers. The major development goal in fiscal year 2008 for the MOCFE solver was to include a two-dimensional capability that is scalable to hundreds of processors. The short term goal of this solver is to solve two-dimensional representations of reactor systems such that the energy and spatial self-shielding are accounted for and reliable cross sections can be generated for the homogeneous calculations. In this report we present good results for an OECD benchmark obtained using the new two-dimensional capability of the MOCFE solver. Additional work on the MOCFE solver is focused on studying the current parallelization algorithms that can be applied to both the two- and three-dimensional implementations such that they are scalable to thousands of processors. The initial research into this topic indicates that, as expected, the current parallelization scheme is not sufficiently scalable for the detailed reactor geometry that it is intended for. As a consequence, we are starting the investigative research to determine the alternatives that are applicable for massively parallel machines. The major development goal in fiscal year 2008 for the PN2ND and SN2ND solvers was to introduce parallelism by angle and energy. The motivation for this is two-fold: (1) reduce the memory burden by picking a simpler preconditioner with reduced matrix storage and (2) improve parallel performance by solving the angular subsystems of the within group equation simultaneously. The solver development in FY2007 focused on using PETSc to solve the within group equation where only spatial parallelization was utilized. Because most homogeneous problems required relatively few spatial degrees of freedom (tens of thousands) the only way to improve the parallelism was to spread the angular moment subsystems across the parallel system. While the coding has been put into place for parallelization by space, angle, and group, we have not optimized any of the solvers and therefore do not give an assessment of the achievement of this work in this report. The immediate task to be completed is to implement and validate Tchebychev acceleration of the fission source iteration algorithm (inverse power method in this work) and optimize both the PN2ND and SN2ND solvers. We further intend to extend the applicability of the UNIC code by adding a first-order discrete ordinates solver termed SN1ST. Upon completion of this work, all memory usage problems are to be identified and studied in the solvers with the intent of making the new version of an exportable production code in either FY2008 or FY2009. This report covers the status of these tasks and discusses the work yet to be completed.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Modeling UV Radiation Feedback from Massive Stars. I. Implementation of Adaptive Ray-tracing Method and Tests [Modeling UV Radiation Feedback from Massive Stars: I. Implementation of Adaptive Ray Tracing Algorithm]

Here, we present an implementation of an adaptive ray-tracing (ART) module in the Athena hydrodynamics code that accurately and efficiently handles the radiative transfer involving multiple point sources on a three-dimensional Cartesian grid. We adopt a recently proposed parallel algorithm that uses nonblocking, asynchronous MPI communications to accelerate transport of rays across the computational domain. We validate our implementation through several standard test problems, including the propagation of radiation in vacuum and the expansions of various types of H II regions. Additionally, scaling tests show that the cost of a full ray trace per source remains comparable to that of the hydrodynamics update on up to $\sim {10}^{3}$ processors. To demonstrate application of our ART implementation, we perform a simulation of star cluster formation in a marginally bound, turbulent cloud, finding that its star formation efficiency is 12% when both radiation pressure forces and photoionization by UV radiation are treated. We directly compare the radiation forces computed from the ART scheme with those from the M 1 closure relation. Although the ART and M 1 schemes yield similar results on large scales, the latter is unable to resolve the radiation field accurately near individual point sources.

79 ASTRONOMY AND ASTROPHYSICS↗

Parallel physics-informed neural networks via domain decomposition

Here we develop a distributed framework for the physics-informed neural networks (PINNs) based on two recent extensions, namely conservative PINNs (cPINNs) and extended PINNs (XPINNs), which employ domain decomposition in space and in time-space, respectively. This domain decomposition endows cPINNs and XPINNs with several advantages over the vanilla PINNs, such as parallelization capacity, large representation capacity, efficient hyperparameter tuning, and is particularly effective for multi-scale and multi-physics problems. Here, we present a parallel algorithm for cPINNs and XPINNs constructed with a hybrid programming model described by MPI + X, where X ∈ {CPUs, GPUs}. The main advantage of cPINN and XPINN over the more classical data and model parallel approaches is the flexibility of optimizing all hyperparameters of each neural network separately in each subdomain. We compare the performance of distributed cPINNs and XPINNs for various forward problems, using both weak and strong scalings. Our results indicate that for space domain decomposition, cPINNs are more efficient in terms of communication cost but XPINNs provide greater flexibility as they can also handle time-domain decomposition for any differential equations, and can deal with any arbitrarily shaped complex subdomains. To this end, we also present an application of the parallel XPINN method for solving an inverse diffusion problem with variable conductivity on the United States map, using ten regions as subdomains.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Multi-node and Multi-core Performance Studies of a Monte Carlo code RMC

Over the past 15 years the architecture of high-performance computing systems has become increasingly heterogeneous with individual computing cores being subdivided into nodes and even auxiliary co-processing units such as graphics processing units or the Intel many integrated core co-processors. In fact, future computer platforms is moving toward a larger number of nodes and cores with lower memory available per node. Today, several production Monte Carlo codes use some types of hybrid parallelism strategy that makes use of shared-memory parallelism within a computing node and message-passing parallelism between nodes, which enables the codes to utilize available computing resources as much as possible, both among and within computing nodes on a cluster. The so-called hybrid parallelization is also possible with Reactor Monte Carlo Code RMC by activating both MPI and OpenMP parallel algorithms together. On HPC' which typically consists of several or several-ten thousand nodes equipping 16-32 computing intra-cores per node, simulation of RMC can be carried out in the hybrid parallelization with the MPI inter-node parallelization and the OpenMP intra-node parallelization. RMC has adopted a hybrid message-passing and shared-memory approach to parallelism, which enables the code to utilize available computing resources, both among and within computing nodes on a cluster. Shared-memory parallelism is especially valuable for large-model simulations because it enables multiple compute cores within a node to share common model information such as the geometry description, material, cross section and tally data, thus reducing the memory usage per processor. The calculation on the Tianhe-2 supercomputer took 6144 cores, resulting in a rate of 173,314,000 neutrons/minute for BEAVRS benchmark. While pin-wise full core H-M benchmark burnup calculations with millions of depletion regions was selected using MPI/OpenMP parallelism, the speed of 6,144 cores reach 1,843,200 burnable regions per minute. Keeping up with the continuing trend for increases in the number of computing resources within modern HPC systems is a significant challenge for teams that develop and maintain high-performance MC radiation transport solvers. order to address this challenge, RMC has adopted a hybrid message-passing and shared-memory approach to parallelism, which enables the code to utilize available computing resources, both among and within computing nodes on a cluster. Shared-memory parallelism is especially valuable for large-model simulations because it enables multiple compute cores within a node to share common model information such as the geometry description, material and composition definitions, and cross section data, thus reducing the memory usage per processor. The parallel performance has been studied by using two complex and representative benchmarks, H-M and BEAVRS benchmark. All simulations were run on Intel Xeon E5-2692 12 core, 2.2 GHz processors at MilkyWay-2 Super Computer with up to 12,288 available computing cores. While employing MPI parallelism, excellent parallel efficiency is achieved even above 12,000 processors and RMC reaches the parallel efficiency of 96% on 12,288 cores for the pin-by-pin computation of the BEAVRS benchmark. While BEAVRS criticality benchmark was selected using MPI/OpenMP parallelism, the speed of 6,144 cores reach 173,314,000 particles per minute. While pin-wise full core H-M benchmark burnup calculations with millions of depletion regions was selected using MPI/OpenMP parallelism, the speed of 6,144 cores reach 1,843,200 burnable regions per minute. Results from these studies demonstrate that RMC, using the hybrid parallel model described in this paper, is able to handle pin-wise full core burnup calculations with millions of depletion regions and scales well through thousands of processors. (authors)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

SmartApps: Middle-ware for Adaptive Applications on Reconfigurable Platforms

There are several reasons why the performance of current distributed and heterogeneous systems is often disappointing. For example, the characteristics of the application may be input sensitive and evolve during execution causing dramatic changes in memory reference patterns, resource requirements, or degree of concurrency between different phases of the computation. Or, the system may change dynamically with nodes failing or appearing, some network links severed and other links established with different latencies and bandwidths. Another important reason for poor performance is the fairly compartmentalized approach to optimization: applications, compilers, operating systems and hardware configurations are designed and optimized in isolation and without the knowledge of instance specific information and needs of a running application. There is too little information flow across these boundaries and no global optimization is even attempted. For example, most operating systems services like paging, virtual-to-physical page mapping, I/O, or data layout in disks, provide little or no application customization. Similarly, the off-the-shelf hardware used by most commercial systems is optimized to give best average-case performance. To address this problem, we have proposed application-centric computing, or Smart Applications (SAS). In the SAS executable, the compiler embeds most run-time system services, and a performance-optimizing feedback loop that monitors the application's performance and adaptively reconfigures the application and the OS/system platform. At run-time, after incorporating the code's input and determining the system's resources and state, the SAS performs an *instance* specific optimization, which is more tractable than a global generic optimization between application, OS and system. The overriding philosophy of SAS is ``measure, compare, and adapt if beneficial.'' That is, the application will continually monitor its performance and the available resources to determine if, and by how much, performance could be improved if the application was restructured. Then, if the potential performance benefit outweighs the projected overhead costs, the application will restructure itself and the underlying system accordingly. The SAS framework includes performance monitoring and modeling components and mechanisms for performing the actual restructuring at various levels including: (i) algorithmic adaptation, (ii) run-time software optimization (e.g., input sensitivity analysis, etc.), (iii) tuning reconfigurable OS services (scheduling policy, page size, etc), and (iv) system configuration (e.g., selecting which, and how many, computational resources to use). SmartApps is being developed in the STAPL infrastructure. STAPL (the Standard Template Adaptive Parallel Library) is a framework for developing highly-optimizable, adaptable, and portable parallel and distributed applications. It consists of a relatively new and still evolving collection of generic parallel algorithms and distributed containers and a run-time system (RTS) through which the application and compiler interact with the OS and hardware.

97 MATHEMATICS AND COMPUTING↗

Reducing Communication Costs for Sparse Matrix Multiplication within Algebraic Multigrid

We consider the sequence of sparse matrix-matrix multiplications performed during the setup phase of algebraic multigrid. In particular, we show that the most commonly used parallel algorithm is often not the most communication-efficient one for all of the matrix-matrix multiplications involved. By using an alternative algorithm, we show that the communication costs are reduced (in theory and practice), and we demonstrate the performance benefit for both model (structured) and more realistic unstructured problems on large-scale distributed-memory parallel systems. Our theoretical analysis shows that we can reduce communication by a factor of up to 5.4 for a model problem, and we observe in our empirical evaluation communication reductions of factors up to 4.7 for structured problems and 3.7 for unstructured problems. These reductions in communication translate to run-time speedups of up to factors of 2.3 and 2.5, respectively.

97 MATHEMATICS AND COMPUTING↗

Fast and Accurate Support Vector Machines on Large Scale Systems

Support Vector Machines (SVM) is a supervised Machine Learning and Data Mining (MLDM) algorithm, which has become ubiquitous largely due to its high accuracy and obliviousness to dimensionality. The objective of SVM is to find an optimal boundary --- also known as hyperplane --- which separates the samples (examples in a dataset) of different classes by a maximum margin. Usually, very few samples contribute to the definition of the boundary. However, existing parallel algorithms use the entire dataset for finding the boundary, which is sub-optimal for performance reasons. In this paper, we propose a novel distributed memory algorithm to eliminate the samples which do not contribute to the boundary definition in SVM. We propose several heuristics, which range from early (aggressive) to late (conservative) elimination of the samples, such that the overall time for generating the boundary is reduced considerably. In a few cases, a sample may be eliminated (shrunk) pre-emptively --- potentially resulting in an incorrect boundary. We propose a scalable approach to synchronize the necessary data structures such that the proposed algorithm maintains its accuracy. We consider the necessary trade-offs of single/multiple synchronization using in-depth time-space complexity analysis. We implement the proposed algorithm using MPI and compare it with libsvm--- de facto sequential SVM software --- which we enhance with OpenMP for multi-core/many-core parallelism. Our proposed approach shows excellent efficiency using up to 4096 processes on several large datasets such as UCI HIGGS Boson dataset and Offending URL dataset.

Vishnu, Abhinav↗

Collective network for computer structures

A system and method for enabling high-speed, low-latency global collective communications among interconnected processing nodes. The global collective network optimally enables collective reduction operations to be performed during parallel algorithm operations executing in a computer structure having a plurality of the interconnected processing nodes. Router devices ate included that interconnect the nodes of the network via links to facilitate performance of low-latency global processing operations at nodes of the virtual network and class structures. The global collective network may be configured to provide global barrier and interrupt functionality in asynchronous or synchronized manner. When implemented in a massively-parallel supercomputing structure, the global collective network is physically and logically partitionable according to needs of a processing algorithm.

97 MATHEMATICS AND COMPUTING↗

Collective network for computer structures

A system and method for enabling high-speed, low-latency global collective communications among interconnected processing nodes. The global collective network optimally enables collective reduction operations to be performed during parallel algorithm operations executing in a computer structure having a plurality of the interconnected processing nodes. Router devices are included that interconnect the nodes of the network via links to facilitate performance of low-latency global processing operations at nodes of the virtual network and class structures. The global collective network may be configured to provide global barrier and interrupt functionality in asynchronous or synchronized manner. When implemented in a massively-parallel supercomputing structure, the global collective network is physically and logically partitionable according to needs of a processing algorithm.

Blumrich, Matthias A.↗