Speedup and scalability analysis of Master-Slave applications on large heterogeneous clusters

Although cluster environments have an enormous potential processing power, real applications that take advantage of this power remain an elusive goal. This is due, in part, to the lack of understanding about the characteristics of the applications best suited for these environments. This paper focuses on Master/Slave applications for large heterogeneous clusters. It defines application, cluster and execution models to derive an analytic expression for the execution time. It defines speedup and derives speedup bounds based on the inherent parallelism of the application and the aggregated computing power of the cluster. The paper derives an analytical expression for efficiency and uses it to define scalability of the algorithm-cluster combination based on the isoefficiency metric. Furthermore, the paper establishes necessary and sufficient conditions for an algorithm-cluster combination to be scalable which are easy to verify and use in practice. Finally, it covers the impact of network contention as the number of processors grow. (C) 2007 Elsevier B.V. All rights reserved.

Classificação e comparação de ferramentas para análise de desempenho de sistemas paralelos

Pós-graduação em Engenharia Elétrica - FEIS

Harmonic identification using parallel neural networks in single-phase systems

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Large amplitude oscillations for a class of symmetric polynomial differential systems in R³

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Load compensation in four-wire electrical power systems

This work describes a methodology for power factor control and correction of the unbalanced currents in four-wire electric circuits. The methodology is based on the insertion of two compensation networks, one wye-grounded neutral and another in delta, in parallel to the load. The mathematical development has been proposed in previous work [3]. In this paper, however, the methodology was adapted to accept different power factors for the system to be compensated. on the other hand, the determination of the compensation susceptances is based on the instantaneous values of the load currents. The results are obtained using the MatLab - Simulink environment.

Robust fault diagnosis in power distribution systems based on fuzzy ARTMAP neural network-aided evidence theory

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Critical temperature for first-order phase transitions in confined systems

We consider the Euclidean D-dimensional -lambda vertical bar phi vertical bar(4)+eta vertical bar rho vertical bar(6) (lambda,eta > 0) model with d (d <= D) compactified dimensions. Introducing temperature by means of the Ginzburg-Landau prescription in the mass term of the Hamiltonian, this model can be interpreted as describing a first-order phase transition for a system in a region of the D-dimensional space, limited by d pairs of parallel planes, orthogonal to the coordinates axis x(1), x(2),..., x(d). The planes in each pair are separated by distances L-1, L-2, ... , L-d. We obtain an expression for the transition temperature as a function of the size of the system, T-c({L-i}), i = 1, 2, ..., d. For D = 3 we particularize this formula, taking L-1 = L-2 = ... = L-d = L for the physically interesting cases d = 1 (a film), d = 2 (an infinitely long wire having a square cross-section), and for d = 3 (a cube). For completeness, the corresponding formulas for second-order transitions are also presented. Comparison with experimental data for superconducting films and wires shows qualitative agreement with our theoretical expressions.

Artificial immune systems for classification of petroleum well drilling operations

This paper presents two approaches of Artificial Immune System for Pattern Recognition (CLONALG and Parallel AIRS2) to classify automatically the well drilling operation stages. The classification is carried out through the analysis of some mud-logging parameters. In order to validate the performance of AIS techniques, the results were compared with others classification methods: neural network, support vector machine and lazy learning.

Electronic ballast for multiple fluorescent lamp systems

An electronic ballast for multiple tubular fluorescent lamp systems is presented. The proposed structure has a high value for the power factor, a dimming capability, and soft switching of the semiconductor devices operated at high frequencies. A zero-current switching pulse width modulated SEPIC converter is used as the rectifying stage and it is controlled using the instantaneous average input current technique. The inverting stage consists of classical resonant half-bridge converter with series-resonant parallel-loaded filters. The dimming control technique is based on varying the switching frequency and monitoring the phase shift of the current drained by the filters and lamps in order to establish a closed loop control. Experimental results are presented that validate the theoretical analysis.

Mechanical and electrical driving field induced high-frequency dielectric anomalies in ferroelectric systems

Polycrystalline or single-crystal ferroelectric materials present dielectric dispersion in the frequency range 100 MHz-1 GHz that has been attributed to a dispersive ( relaxation-like) mechanism as well as a resonant mechanism. Particularly in 'normal' ferroelectric materials, a dielectric response that is indistinguishable from dispersion or a resonance has been reported. Nevertheless, the reported results are not conclusive enough to distinguish each mechanism clearly. A detailed study of the dielectric dispersion phenomenon has been carried out in PbTiO3-based ferroelectric ceramics, with the composition Pb1-xLaxTiO3 (x = 0.15), over a wide range of temperatures and frequencies, including microwave frequencies. The dielectric response of La-modified lead titanate ferroelectric ceramics, in 'virgin' and poled states, has been investigated in the temperature and frequency ranges 300-450 K and 1 kHz-2 GHz, respectively. The results revealed that the frequency dependence of the dielectric anomalies, depending on the measuring direction with respect to the orientation of the macroscopic polarization, may be described as a general mechanism related to an 'over-damped' resonant process. Applying either a uniaxial stress along the measurement field direction or a poling electric field parallel and/or perpendicular to the measuring direction, a resonant response of the real and imaginary components of the dielectric constant is observed, in contrast to the dispersion behavior obtained in the absence of the stress, for the 'virgin' samples. Both results, resonance and/or dispersion, can be explained by considering a common mechanism involving a resonant response (damped and/or over-damped) which is strongly affected by a ferroelastic-ferroelectric coupling, contributing to the low-field dielectric constant.

Methodology for load compensation in four-wire electrical power systems based on instantaneous sampling

This work describes a methodology for power factor control and correction of the unbalanced currents in four-wire electric circuits. The methodology is based on the insertion of two compensation networks, one wye-grounded neutral and other in delta, in parallel to the load. The mathematical development has been proposed in previous work [3]. In this paper, however, the determination of the compensation susceptances is based on the instantaneous values of load currents. The results are obtained using the MatLab-Simulink enviroment

Parallel simulated annealing applied to long term transmission network expansion planning

The simulated annealing optimization technique has been successfully applied to a number of electrical engineering problems, including transmission system expansion planning. The method is general in the sense that it does not assume any particular property of the problem being solved, such as linearity or convexity. Moreover, it has the ability to provide solutions arbitrarily close to an optimum (i.e. it is asymptotically convergent) as the cooling process slows down. The drawback of the approach is the computational burden: finding optimal solutions may be extremely expensive in some cases. This paper presents a Parallel Simulated Annealing, PSA, algorithm for solving the long term transmission network expansion planning problem. A strategy that does not affect the basic convergence properties of the Sequential Simulated Annealing algorithm have been implementeded and tested. The paper investigates the conditions under which the parallel algorithm is most efficient. The parallel implementations have been tested on three example networks: a small 6-bus network, and two complex real-life networks. Excellent results are reported in the test section of the paper: in addition to reductions in computing times, the Parallel Simulated Annealing algorithm proposed in the paper has shown significant improvements in solution quality for the largest of the test networks.

Field profiles of bulk plasmon polariton modes in layered systems containing a metamaterial

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Unfolding of plasmon-polariton modes in one-dimensional layered systems containing anisotropic left-handed materials

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Eigenvalue analyses of two parallel lines using a single real transformation matrix

For a typical non-symmetrical system with two parallel three phase transmission lines, modal transformation is applied using some examples of single real transformation matrices. These examples are applied searching an adequate single real transformation matrix to two parallel three phase transmission line systems. The analyses are started with the eigenvector and eigenvalue studies, using Clarke's transformation or linear combinations of Clarke's elements. The Z C and parameters are analyzed for the case that presents the smallest errors between the exact eigenvalues and the single real transformation matrix application results. The single real transformation determined for this case is based on Clarke's matrix and its main characteristic is the use of a unique homopolar reference. So, the homopolar mode becomes a connector mode between the two three-phase circuits of the analyzed system. ©2005 IEEE.