GPGPU implementation of growing neural gas: application to 3D scene reconstruction

Self-organising neural models have the ability to provide a good representation of the input space. In particular the Growing Neural Gas (GNG) is a suitable model because of its flexibility, rapid adaptation and excellent quality of representation. However, this type of learning is time-consuming, especially for high-dimensional input data. Since real applications often work under time constraints, it is necessary to adapt the learning process in order to complete it in a predefined time. This paper proposes a Graphics Processing Unit (GPU) parallel implementation of the GNG with Compute Unified Device Architecture (CUDA). In contrast to existing algorithms, the proposed GPU implementation allows the acceleration of the learning process keeping a good quality of representation. Comparative experiments using iterative, parallel and hybrid implementations are carried out to demonstrate the effectiveness of CUDA implementation. The results show that GNG learning with the proposed implementation achieves a speed-up of 6× compared with the single-threaded CPU implementation. GPU implementation has also been applied to a real application with time constraints: acceleration of 3D scene reconstruction for egomotion, in order to validate the proposal.

Parallel 3D Simulation of a Fault Gouge using the Lattice Solid Model

Despite the insight gained from 2-D particle models, and given that the dynamics of crustal faults occur in 3-D space, the question remains, how do the 3-D fault gouge dynamics differ from those in 2-D? Traditionally, 2-D modeling has been preferred over 3-D simulations because of the computational cost of solving 3-D problems. However, modern high performance computing architectures, combined with a parallel implementation of the Lattice Solid Model (LSM), provide the opportunity to explore 3-D fault micro-mechanics and to advance understanding of effective constitutive relations of fault gouge layers. In this paper, macroscopic friction values from 2-D and 3-D LSM simulations, performed on an SGI Altix 3700 super-cluster, are compared. Two rectangular elastic blocks of bonded particles, with a rough fault plane and separated by a region of randomly sized non-bonded gouge particles, are sheared in opposite directions by normally-loaded driving plates. The results demonstrate that the gouge particles in the 3-D models undergo significant out-of-plane motion during shear. The 3-D models also exhibit a higher mean macroscopic friction than the 2-D models for varying values of interparticle friction. 2-D LSM gouge models have previously been shown to exhibit accelerating energy release in simulated earthquake cycles, supporting the Critical Point hypothesis. The 3-D models are shown to also display accelerating energy release, and good fits of power law time-to-failure functions to the cumulative energy release are obtained.

The application of parallel processing methods to vibrational analysis

The trend in modal extraction algorithms is to use all the available frequency response functions data to obtain a global estimate of the natural frequencies, damping ratio and mode shapes. Improvements in transducer and signal processing technology allow the simultaneous measurement of many hundreds of channels of response data. The quantity of data available and the complexity of the extraction algorithms make considerable demands on the available computer power and require a powerful computer or dedicated workstation to perform satisfactorily. An alternative to waiting for faster sequential processors is to implement the algorithm in parallel, for example on a network of Transputers. Parallel architectures are a cost effective means of increasing computational power, and a larger number of response channels would simply require more processors. This thesis considers how two typical modal extraction algorithms, the Rational Fraction Polynomial method and the Ibrahim Time Domain method, may be implemented on a network of transputers. The Rational Fraction Polynomial Method is a well known and robust frequency domain 'curve fitting' algorithm. The Ibrahim Time Domain method is an efficient algorithm that 'curve fits' in the time domain. This thesis reviews the algorithms, considers the problems involved in a parallel implementation, and shows how they were implemented on a real Transputer network.

Combining algorithms to predict bacterial protein sub-cellular location:parallel versus concurrent implementations

We describe a novel and potentially important tool for candidate subunit vaccine selection through in silico reverse-vaccinology. A set of Bayesian networks able to make individual predictions for specific subcellular locations is implemented in three pipelines with different architectures: a parallel implementation with a confidence level-based decision engine and two serial implementations with a hierarchical decision structure, one initially rooted by prediction between membrane types and another rooted by soluble versus membrane prediction. The parallel pipeline outperformed the serial pipeline, but took twice as long to execute. The soluble-rooted serial pipeline outperformed the membrane-rooted predictor. Assessment using genomic test sets was more equivocal, as many more predictions are made by the parallel pipeline, yet the serial pipeline identifies 22 more of the 74 proteins of known location.

Shape discrimination using invariants defined from higher-order spectra

An approach to pattern recognition using invariant parameters based on higher-order spectra is presented. In particular, bispectral invariants are used to classify one-dimensional shapes. The bispectrum, which is translation invariant, is integrated along straight lines passing through the origin in bifrequency space. The phase of the integrated bispectrum is shown to be scale- and amplification-invariant. A minimal set of these invariants is selected as the feature vector for pattern classification. Pattern recognition using higher-order spectral invariants is fast, suited for parallel implementation, and works for signals corrupted by Gaussian noise. The classification technique is shown to distinguish two similar but different bolts given their one-dimensional profiles

Pattern recognition using invariants defined from higher order spectra - one-dimensional inputs

A new approach to pattern recognition using invariant parameters based on higher order spectra is presented. In particular, invariant parameters derived from the bispectrum are used to classify one-dimensional shapes. The bispectrum, which is translation invariant, is integrated along straight lines passing through the origin in bifrequency space. The phase of the integrated bispectrum is shown to be scale and amplification invariant, as well. A minimal set of these invariants is selected as the feature vector for pattern classification, and a minimum distance classifier using a statistical distance measure is used to classify test patterns. The classification technique is shown to distinguish two similar, but different bolts given their one-dimensional profiles. Pattern recognition using higher order spectral invariants is fast, suited for parallel implementation, and has high immunity to additive Gaussian noise. Simulation results show very high classification accuracy, even for low signal-to-noise ratios.

A new load distribution strategy for linear network with communication delays

In this paper, we propose a new load distribution strategy called `send-and-receive' for scheduling divisible loads, in a linear network of processors with communication delay. This strategy is designed to optimally utilize the network resources and thereby minimizes the processing time of entire processing load. A closed-form expression for optimal size of load fractions and processing time are derived when the processing load originates at processor located in boundary and interior of the network. A condition on processor and link speed is also derived to ensure that the processors are continuously engaged in load distributions. This paper also presents a parallel implementation of `digital watermarking problem' on a personal computer-based Pentium Linear Network (PLN) topology. Experiments are carried out to study the performance of the proposed strategy and results are compared with other strategies found in literature.

Concise row-pruning algorithm to invert a matrix

Presented here, in a vector formulation, is an O(mn2) direct concise algorithm that prunes/identifies the linearly dependent (ld) rows of an arbitrary m X n matrix A and computes its reflexive type minimum norm inverse A(mr)-, which will be the true inverse A-1 if A is nonsingular and the Moore-Penrose inverse A+ if A is full row-rank. The algorithm, without any additional computation, produces the projection operator P = (I - A(mr)- A) that provides a means to compute any of the solutions of the consistent linear equation Ax = b since the general solution may be expressed as x = A(mr)+b + Pz, where z is an arbitrary vector. The rank r of A will also be produced in the process. Some of the salient features of this algorithm are that (i) the algorithm is concise, (ii) the minimum norm least squares solution for consistent/inconsistent equations is readily computable when A is full row-rank (else, a minimum norm solution for consistent equations is obtainable), (iii) the algorithm identifies ld rows, if any, and reduces concerned computation and improves accuracy of the result, (iv) error-bounds for the inverse as well as the solution x for Ax = b are readily computable, (v) error-free computation of the inverse, solution vector, rank, and projection operator and its inherent parallel implementation are straightforward, (vi) it is suitable for vector (pipeline) machines, and (vii) the inverse produced by the algorithm can be used to solve under-/overdetermined linear systems.

Joint Channel Estimation and Data Detection in MIMO-OFDM Systems: A Sparse Bayesian Learning Approach

The impulse response of wireless channels between the N-t transmit and N-r receive antennas of a MIMO-OFDM system are group approximately sparse (ga-sparse), i.e., NtNt the channels have a small number of significant paths relative to the channel delay spread and the time-lags of the significant paths between transmit and receive antenna pairs coincide. Often, wireless channels are also group approximately cluster-sparse (gac-sparse), i.e., every ga-sparse channel consists of clusters, where a few clusters have all strong components while most clusters have all weak components. In this paper, we cast the problem of estimating the ga-sparse and gac-sparse block-fading and time-varying channels in the sparse Bayesian learning (SBL) framework and propose a bouquet of novel algorithms for pilot-based channel estimation, and joint channel estimation and data detection, in MIMO-OFDM systems. The proposed algorithms are capable of estimating the sparse wireless channels even when the measurement matrix is only partially known. Further, we employ a first-order autoregressive modeling of the temporal variation of the ga-sparse and gac-sparse channels and propose a recursive Kalman filtering and smoothing (KFS) technique for joint channel estimation, tracking, and data detection. We also propose novel, parallel-implementation based, low-complexity techniques for estimating gac-sparse channels. Monte Carlo simulations illustrate the benefit of exploiting the gac-sparse structure in the wireless channel in terms of the mean square error (MSE) and coded bit error rate (BER) performance.

Topics in randomized numerical linear algebra

This thesis studies three classes of randomized numerical linear algebra algorithms, namely: (i) randomized matrix sparsification algorithms, (ii) low-rank approximation algorithms that use randomized unitary transformations, and (iii) low-rank approximation algorithms for positive-semidefinite (PSD) matrices.

Randomized matrix sparsification algorithms set randomly chosen entries of the input matrix to zero. When the approximant is substituted for the original matrix in computations, its sparsity allows one to employ faster sparsity-exploiting algorithms. This thesis contributes bounds on the approximation error of nonuniform randomized sparsification schemes, measured in the spectral norm and two NP-hard norms that are of interest in computational graph theory and subset selection applications.

Low-rank approximations based on randomized unitary transformations have several desirable properties: they have low communication costs, are amenable to parallel implementation, and exploit the existence of fast transform algorithms. This thesis investigates the tradeoff between the accuracy and cost of generating such approximations. State-of-the-art spectral and Frobenius-norm error bounds are provided.

The last class of algorithms considered are SPSD "sketching" algorithms. Such sketches can be computed faster than approximations based on projecting onto mixtures of the columns of the matrix. The performance of several such sketching schemes is empirically evaluated using a suite of canonical matrices drawn from machine learning and data analysis applications, and a framework is developed for establishing theoretical error bounds.

In addition to studying these algorithms, this thesis extends the Matrix Laplace Transform framework to derive Chernoff and Bernstein inequalities that apply to all the eigenvalues of certain classes of random matrices. These inequalities are used to investigate the behavior of the singular values of a matrix under random sampling, and to derive convergence rates for each individual eigenvalue of a sample covariance matrix.

Paralelização em CUDA/GLSL do algoritmo SIFT para reconhecimento de íris

Neste trabalho é estudada a viabilidade de uma implementação em paralelo do algoritmo scale invariant feature transform (SIFT) para identificação de íris. Para a implementação do código foi utilizada a arquitetura para computação paralela compute unified device architecture (CUDA) e a linguagem OpenGL shading language (GLSL). O algoritmo foi testado utilizando três bases de dados de olhos e íris, o noisy visible wavelength iris image Database (UBIRIS), Michal-Libor e CASIA. Testes foram feitos para determinar o tempo de processamento para verificação da presença ou não de um indivíduo em um banco de dados, determinar a eficiência dos algoritmos de busca implementados em GLSL e CUDA e buscar valores de calibração que melhoram o posicionamento e a distribuição dos pontos-chave na região de interesse (íris) e a robustez do programa final.

Implementação paralela do algoritmo iterativo de busca do parâmetro de regularização ótimo para o funcional de Tikhonov no problema de restauração de imagens

O uso de técnicas com o funcional de Tikhonov em processamento de imagens tem sido amplamente usado nos últimos anos. A ideia básica nesse processo é modificar uma imagem inicial via equação de convolução e encontrar um parâmetro que minimize esse funcional afim de obter uma aproximação da imagem original. Porém, um problema típico neste método consiste na seleção do parâmetro de regularização adequado para o compromisso entre a acurácia e a estabilidade da solução. Um método desenvolvido por pesquisadores do IPRJ e UFRJ, atuantes na área de problemas inversos, consiste em minimizar um funcional de resíduos através do parâmetro de regularização de Tikhonov. Uma estratégia que emprega a busca iterativa deste parâmetro visando obter um valor mínimo para o funcional na iteração seguinte foi adotada recentemente em um algoritmo serial de restauração. Porém, o custo computacional é um fator problema encontrado ao empregar o método iterativo de busca. Com esta abordagem, neste trabalho é feita uma implementação em linguagem C++ que emprega técnicas de computação paralela usando MPI (Message Passing Interface) para a estratégia de minimização do funcional com o método de busca iterativa, reduzindo assim, o tempo de execução requerido pelo algoritmo. Uma versão modificada do método de Jacobi é considerada em duas versões do algoritmo, uma serial e outra em paralelo. Este algoritmo é adequado para implementação paralela por não possuir dependências de dados como de Gauss-Seidel que também é mostrado a convergir. Como indicador de desempenho para avaliação do algoritmo de restauração, além das medidas tradicionais, uma nova métrica que se baseia em critérios subjetivos denominada IWMSE (Information Weighted Mean Square Error) é empregada. Essas métricas foram introduzidas no programa serial de processamento de imagens e permitem fazer a análise da restauração a cada passo de iteração. Os resultados obtidos através das duas versões possibilitou verificar a aceleração e a eficiência da implementação paralela. A método de paralelismo apresentou resultados satisfatórios em um menor tempo de processamento e com desempenho aceitável.

Implementação paralela de um código de elementos finitos em 2D para as Equações de Navier-Stokes para fluidos incompressíveis com transporte de escalares.

O estudo do fluxo de água e do transporte escalar em reservatórios hidrelétricos é importante para a determinação da qualidade da água durante as fases iniciais do enchimento e durante a vida útil do reservatório. Neste contexto, um código de elementos finitos paralelo 2D foi implementado para resolver as equações de Navier-Stokes para fluido incompressível acopladas a transporte escalar, utilizando o modelo de programação de troca de mensagens, a fim de realizar simulações em um ambiente de cluster de computadores. A discretização espacial é baseada no elemento MINI, que satisfaz as condições de Babuska-Brezzi (BB), que permite uma formulação mista estável. Todas as estruturas de dados distribuídos necessárias nas diferentes fases do código, como pré-processamento, solução e pós-processamento, foram implementadas usando a biblioteca PETSc. Os sistemas lineares resultantes foram resolvidos usando o método da projeção discreto com fatoração LU por blocos. Para aumentar o desempenho paralelo na solução dos sistemas lineares, foi empregado o método de condensação estática para resolver a velocidade intermediária nos vértices e no centróide do elemento MINI separadamente. Os resultados de desempenho do método de condensação estática com a abordagem da solução do sistema completo foram comparados. Os testes mostraram que o método de condensação estática apresenta melhor desempenho para grandes problemas, às custas de maior uso de memória. O desempenho de outras partes do código também são apresentados.

积分偏移的计算几何与并行实现

The Second Round of Oil & Gas Exploration needs more precision imaging method, velocity vs. depth model and geometry description on Complicated Geological Mass. Prestack time migration on inhomogeneous media was the technical basic of velocity analysis, prestack time migration on Rugged surface, angle gather and multi-domain noise suppression. In order to realize this technique, several critical technical problems need to be solved, such as parallel computation, velocity algorithm on ununiform grid and visualization. The key problem is organic combination theories of migration and computational geometry. Based on technical problems of 3-D prestack time migration existing in inhomogeneous media and requirements from nonuniform grid, parallel process and visualization, the thesis was studied systematically on three aspects: Infrastructure of velocity varies laterally Green function traveltime computation on ununiform grid, parallel computational of kirchhoff integral migration and 3D visualization, by combining integral migration theory and Computational Geometry. The results will provide powerful technical support to the implement of prestack time migration and convenient compute infrastructure of wave number domain simulation in inhomogeneous media. The main results were obtained as follows: 1. Symbol of one way wave Lie algebra integral, phase and green function traveltime expressions were analyzed, and simple 2-D expression of Lie algebra integral symbol phase and green function traveltime in time domain were given in inhomogeneous media by using pseudo-differential operators’ exponential map and Lie group algorithm preserving geometry structure. Infrastructure calculation of five parts, including derivative, commutating operator, Lie algebra root tree, exponential map root tree and traveltime coefficients , was brought forward when calculating asymmetry traveltime equation containing lateral differential in 3-D by this method. 2. By studying the infrastructure calculation of asymmetry traveltime in 3-D based on lateral velocity differential and combining computational geometry, a method to build velocity library and interpolate on velocity library using triangulate was obtained, which fit traveltime calculate requirements of parallel time migration and velocity estimate. 3. Combining velocity library triangulate and computational geometry, a structure which was convenient to calculate differential in horizontal, commutating operator and integral in vertical was built. Furthermore, recursive algorithm, for calculating architecture on lie algebra integral and exponential map root tree (Magnus in Math), was build and asymmetry traveltime based on lateral differential algorithm was also realized. 4. Based on graph theory and computational geometry, a minimum cycle method to decompose area into polygon blocks, which can be used as topological representation of migration result was proposed, which provided a practical method to block representation and research to migration interpretation results. 5. Based on MPI library, a process of bringing parallel migration algorithm at arbitrary sequence traces into practical was realized by using asymmetry traveltime based on lateral differential calculation and Kirchhoff integral method. 6. Visualization of geological data and seismic data were studied by the tools of OpenGL and Open Inventor, based on computational geometry theory, and a 3D visualize system on seismic imaging data was designed.

波动方程辛几何算法三维叠前深度偏移及其并行实现

Oil and scientific groups have been focusing on the 3D wave equation prestack depth migration since it can solve the complex problems of the geologic structure accurately and maintain the wave information, which is propitious to lithology imaging. The symplectic method was brought up by Feng Kang firstly in 1984 and became the hotspot of numerical computation study. It will be widely applied in many scientific field of necessity because of its great virtue in scientific sense. This paper combines the Symplectic method and the 3-D wave equation prestack depth migration to bring up an effectual numerical computation method of wave field extrapolatation technique under the scientific background mentioned above. At the base of deep analysis of computation method and the performance of PC cluster, a seismic prestack depth migration flow considering the virtue of both seismic migration method and Pc cluster has formatted. The software, named 3D Wave Equation Prestack Depth Migration of Symplectic Method, which is based on the flow, has been enrolled in the National Bureau of Copyright (No. 0013767). Dagang and Daqing Oil Field have now put it into use in the field data processing. In this paper, the one way wave equation operator is decompounded into a phase shift operator and a time shift operator and the correct item with high rank Symplectic method when approaching E exponent. After reviewing eliminating alias frequency of operator, computing the maximum angle of migration and the imaging condition, we present the test result of impulse response of the Symplectic method. Taking the imaging results of the SEG/EAGE salt and overthrust models for example and seeing about the imaging ability with complex geologic structure of our software system, the paper has discussed the effect of the selection of imaging parameters and the effectuation on the migration result of the seismic wavelet and compared the 2-D and 3-D prestack depth migration result of the salt mode. We also present the test result of impulse response with the overthrust model. The imaging result of the two international models indicates that the Symplectic method of 3-D prestack depth migration accommodates great transversal velocity variation and complex geologic structure. The huge computing cost is the key obstruction that 3-D prestack depth migration wave equation cannot be adopted by oil industry. After deep analysis of prestack depth migration flow and the character of PC cluster ,the paper put forward :i)parallel algorithms in shot and frequency domain of the common shot gather 3-D wave equation prestack migration; ii)the optimized setting scheme of breakpoint in field data processing; iii)dynamic and static load balance among the nodes of the PC cluster in the 3-D prestack depth migration. It has been proven that computation periods of the 3-D prestack depth migration imaging are greatly shortened given that adopting the computing method mentioned in the paper. In addition,considering the 3-D wave equation prestack depth migration flow in complex medium and examples of the field data processing, the paper put the emphasis on: i)seismic data relative preprocessing, ii) 2.5D prestack depth migration velocity analysis, iii)3D prestack depth migration. The result of field data processing shows satisfied application ability of the flow put forward in the paper.