An optimal parallel algorithm for sparse matrix bandwidth reduction on a linear array

A simple and efficient algorithm for the bandwidth reduction of sparse symmetric matrices is proposed. It involves column-row permutations and is well-suited to map onto the linear array topology of the SIMD architectures. The efficiency of the algorithm is compared with the other existing algorithms. The interconnectivity and the memory requirement of the linear array are discussed and the complexity of its layout area is derived. The parallel version of the algorithm mapped onto the linear array is then introduced and is explained with the help of an example. The optimality of the parallel algorithm is proved by deriving the time complexities of the algorithm on a single processor and the linear array.

Dynamic growth of tensile cracks by ductile and brittle fracture mechanisms in a viscoplastic material

In this paper, a finite element analysis of steady-state dynamic crack growth under Mode I, plane strain, small-scale yielding conditions is performed in a rate dependent plastic material characterized by the over-stress model. The main objective of the paper is to obtain theoretically the dependence of dynamic fracture toughness on crack speed. Crack propagation due to a ductile (micro-void) mechanism or a brittle (cleavage) mechanism, as well as transition from one mode to another are considered. The conversion from ductile to brittle has been observed experimentally but has received very little attention using analytical methods. Local fracture criteria based on strains and stresses are used to describe ductile and brittle fracture mechanisms. The results obtained in this paper are in general agreement with micro-structural observations of mode conversion during fracture initiation. Finally, the particular roles played by material rate sensitivity and inertia are examined in some detail.

Molecular dynamics simulations on parallel and antiparallel C.G*G triplexes

Molecular dynamics (MD) studies have been carried out on the Hoogsteen hydrogen bonded parallel and the reverse Hoogsteen hydrogen banded antiparallel C.G*G triplexes. Earlier, the molecular mechanics studies had shown that the parallel structure was energetically more favourable than the antiparallel structure. To characterize the structural stability of the two triplexes and to investigate whether the antiparallel structure can transit to an energetically more favourable structure, due to the local fluctuations in the structure during the MD simulation, the two structures were subjected to 200ps of constant temperature vacuum MD simulations at 300K. Initially no constraints were applied to the structures and it was observed that for the antiparallel tripler, the structure showed a large root mean square deviation from the starting structure within the first 12ps and the N4-H41-O6 hydrogen bond in the WC duplex got distorted due to a high propeller twist and a moderate increase in the opening angle in the basepairs. Starting from an initial value of 30 degrees, helical twist of the average structure from this simulation had a value of 36 degrees, while the parallel structure stabilized at a twist of 33 degrees. In spite of the hydrogen bond distortions in the antiparallel tripler, it was energetically comparable to the parallel tripler. To examine the structural characteristics of an undistorted structure, another MD simulation was performed on the antiparallel tripler by constraining all the hydrogen bonds. This structure stabilized at an average twist of 33 degrees. In the course of the dynamics though the energy of the molecule - compared to the initial structure - improved, it did not become comparable to the parallel structure. Energy minimization studies performed in the presence of explicit water and counterions also showed the two structures to be equally favourable energetically Together these results indicate that the parallel C.G*G tripler with Hoogsteen hydrogen bonds also represents a stereochemically and energetically favourable structure for this class of triplexes.

A Parallel Progressive Refinement Image Rendering Algorithm on a Scalable Multithreaded VLSI Processor Array

In this paper we develop a multithreaded VLSI processor linear array architecture to render complex environments based on the radiosity approach. The processing elements are identical and multithreaded. They work in Single Program Multiple Data (SPMD) mode. A new algorithm to do the radiosity computations based on the progressive refinement approach[2] is proposed. Simulation results indicate that the architecture is latency tolerant and scalable. It is shown that a linear array of 128 uni-threaded processing elements sustains a throughput close to 0.4 million patches/sec.

Mechanisms of variability and predictability of the tropical coupled ocean-atmosphere system

A conceptual model is proposed to explain the observed aperiodicity in the short term climate fluctuations of the tropical coupled ocean-atmosphere system. This is based on the evidence presented here that the tropical coupled ocean-atmosphere system sustains a low frequency inter-annual mode and a host of higher frequency intra-seasonal unstable modes. At long wavelengths, the low frequency mode is dominant while at short wavelengths, the high frequency modes are dominant resulting in the co-existence of a long wave low frequency mode with some short wave intra-seasonal modes in the tropical coupled system. It is argued that due to its long wavelength, the low frequency mode would behave like a linear oscillator while the higher frequency short wave modes would be nonlinear. The conceptual model envisages that an interaction between the low frequency linear oscillator and the high frequency nonlinear oscillations results in the observed aperiodicity of the tropical coupled system. This is illustrated by representing the higher frequency intra-seasonal oscillations by a nonlinear low order model which is then coupled to a linear oscillator with a periodicity of four years. The physical mechanism resulting in the aperiodicity in the low frequency oscillations and implications of these results on the predictability of the coupled system are discussed.

Decentralized Parallel Operation of Inverters Sharing Unbalanced and Nonlinear Loads

In this paper, a wireless control strategy for parallel operation of three-phase four-wire inverters is proposed. A generalized situation is considered where the inverters are of unequal power ratings and the loads are nonlinear and unbalanced in nature. The proposed control algorithm exploits the potential of sinusoidal domain proportional+multiresonant controller ( in the inner voltage regulation loop) to make the system suitable for nonlinear and unbalanced loads with a simple and generalized structure of virtual output-impedance loop. The decentralized operation is achieved by using three-phase P/Q droop characteristics. The overall control algorithm helps to limit the harmonic contents and the degree of unbalance in the output-voltage waveform and to achieve excellent power-sharing accuracy in spite of mismatch in the inverter output impedances. Moreover, a synchronized turn on with consequent change over to the droop mode is applied for the new incoming unit in order to limit the circulating current completely. The simulation and experimental results from-1 kVA and -0.5 kVA paralleled units validate the effectiveness of the scheme.

Modelling and analysis of the variance in parallelism in parallel computations

In this paper, we introduce an analytical technique based on queueing networks and Petri nets for making a performance analysis of dataflow computations when executed on the Manchester machine. This technique is also applicable for the analysis of parallel computations on multiprocessors. We characterize the parallelism in dataflow computations through a four-parameter characterization, namely, the minimum parallelism, the maximum parallelism, the average parallelism and the variance in parallelism. We observe through detailed investigation of our analytical models that the average parallelism is a good characterization of the dataflow computations only as long as the variance in parallelism is small. However, significant difference in performance measures will result when the variance in parallelism is comparable to or higher than the average parallelism.

Texture evolution and operative mechanisms during large-strain deformation of nanocrystalline nickel

The large-strain deformation of nanocrystalline nickel was investigated at room temperature and cryogenic (liquid N-2) temperature. Deformation mechanisms ranging from grain boundary sliding to slip, operate due to a wide distribution of grain sizes. These mechanisms leave their finger print in the deformation texture evolution during rolling of nanocrystalline nickel. The occurrence and severance of different mechanisms is understood by a thorough characterization of the deformed samples using X-ray diffraction, X-ray texture measurements, electron back-scattered diffraction and transmission electron microscopy. Crystal plasticity-based viscoplastic self-consistent simulations were used to further substantiate the experimental observations. Thus, a comprehensive understanding of deformation behavior of nanocrystalline nickel, which is characterized by simultaneous operation of dislocation-dominated and grain boundary-mediated mechanisms, has been developed.

Toward Effective Scalar Hardware for Highly Vectorizable Applications

The performance of a program will ultimately be limited by its serial (scalar) portion, as pointed out by Amdahl′s Law. Reported studies thus far of instruction-level parallelism have mixed data-parallel program portions with scalar program portions, often leading to contradictory and controversial results. We report an instruction-level behavioral characterization of scalar code containing minimal data-parallelism, extracted from highly vectorized programs of the PERFECT benchmark suite running on a Cray Y-MP system. We classify scalar basic blocks according to their instruction mix, characterize the data dependencies seen in each class, and, as a first step, measure the maximum intrablock instruction-level parallelism available. We observe skewed rather than balanced instruction distributions in scalar code and in individual basic block classes of scalar code; nonuniform distribution of parallelism across instruction classes; and, as expected, limited available intrablock parallelism. We identify frequently occurring data-dependence patterns and discuss new instructions to reduce latency. Toward effective scalar hardware, we study latency-pipelining trade-offs and restricted multiple instruction issue mechanisms.

Parallel Tempering Simulations of Liquid-Phase Adsorption of n-Alkane Mixtures in Zeolite LTA-5A

Adsorption of n-alkane mixtures in the zeolite LTA-5A under liquid-phase conditions has been studied using grand canonical Monte Carlo (GCMC) simulations combined with parallel tempering. Normal GCMC techniques fail for some of these systems due to the preference of linear molecules to coil within a single cage in the zeolite. The narrow zeolite windows severerly restrict interactions of the molecules, making it difficult to simulate cooperative rearrangements necessary to explore configuration space. Because of these reasons, normal GCMC simulations results show poor reproducibility in some cases. These problems were overcome with parallel tempering techniques. Even with parallel tempering, these are very challenging systems for molecular simulation. Similar problems may arise for other zeolites such as CHA, AFX, ERI, KFI, and RHO having cages connected by narrow windows. The simulations capture the complex selectivity behavior observed in experiments such as selectivity inversion and azeotrope formation.

Ball wear mechanisms and control in the grinding of sulfide and phosphate ores

Marked ball grinding tests were carried out in the laboratory using high carbon low alloy steel (cast and forged) and high chrome cast iron balls. Relative ball wear as a function of grinding period and milling conditions was evaluated for the different type of ball materials in the grinding of lead-zinc sulphide and phosphate ores. Results indicated that ball wear increased with time and showed a sharp increase for wet grinding over dry grinding. Ball wear under wet grinding conditions was also influenced by the gaseous atmosphere in the mill. The influence of oxygen on the corrosive wear of grinding balls was increasingly felt in case of sulphide ore grinding. The grinding ball materials could be arranged in the following order with respect to their overall wear resistance: