Pipelined ring algorithm for matrix multiplication on transputer networks: performance analysis and estimation

A parallel matrix multiplication algorithm is presented, and studies of its performance and estimation are discussed. The algorithm is implemented on a network of transputers connected in a ring topology. An efficient scheme for partitioning the input matrices is introduced which enables overlapping computation with communication. This makes the algorithm achieve near-ideal speed-up for reasonably large matrices. Analytical expressions for the execution time of the algorithm have been derived by analysing its computation and communication characteristics. These expressions are validated by comparing the theoretical results of the performance with the experimental values obtained on a four-transputer network for both square and irregular matrices. The analytical model is also used to estimate the performance of the algorithm for a varying number of transputers and varying problem sizes. Although the algorithm is implemented on transputers, the methodology and the partitioning scheme presented in this paper are quite general and can be implemented on other processors which have the capability of overlapping computation with communication. The equations for performance prediction can also be extended to other multiprocessor systems.

Microstructure control and wear of Al2O3-SiC-(Al,Si) composites made by melt oxidation

Ceramic matrix composites of Al2O3-SiC-(Al,Si) have been fabricated by directed melt oxidation of aluminum alloys into SiC particulate preforms. The proportions of Al2O3, alloy, and porosity in the composite can be controlled by proper selection of SLC particle size and the processing temperature. The wear resistance of composites was evaluated in pin-on-disk experiments against a hard steel substrate. Minimum wear rate comparable to conventional ceramics such as ZTA is recorded for the composition containing the highest fraction of alloy, owing to the development of a thin and adherent tribofilm with a low coefficient of friction.

Recursive parametrization of quark flavour mixing matrices

We examine quark flavour mixing matrices for three and four generations using the recursive parametrization of U(n) and SU(n) matrices developed earlier. After a brief summary of the recursive parametrization, we obtain expressions for the independent rephasing invariants and also the constraints on them that arise from the requirement of mod symmetry of the flavour mixing matrix.

Density-matrix renormalization-group studies of the spin-1/2 Heisenberg system with dimerization and frustration

Using the density-matrix renormalization-group technique, we study the ground-state phase diagram and other low-energy properties of an isotropic antiferromagnetic spin-1/2 chain with both dimerization and frustration, i.e., an alternation delta of the nearest-neighbor exchanges and a next-nearest-neighbor exchange J(2). For delta = 0, the system is gapless for J(2) < J(2c) and has a gap for J(2) > J(2c) where J(2c) is about 0.241. For J(2) = J(2c) the gap above the ground state grows as delta to the power 0.667 +/- 0.001. In the J(2)-delta plane, there is a disorder line 2J(2) + delta = 1. To the left of this line, the peak in the static structure factor S(q) is at q(max) = pi (Neel phase), while to the right of the line, q(max) decreases from pi to pi/2 as J(2) is increased to large values (spiral phase). For delta = 1, the system is equivalent to two coupled chains as on a ladder and it is gapped for all values of the interchain coupling.

The Gaussian Toeplitz matrix

An analytical expression for the LL(T) decomposition for the Gaussian Toeplitz matrix with elements T(ij) = [1/(2-pi)1/2-sigma] exp[-(i - j)2/2-sigma-2] is derived. An exact expression for the determinant and bounds on the eigenvalues follows. An analytical expression for the inverse T-1 is also derived.

Some Algorithms for Eigensubspace Estimation

An important tool in signal processing is the use of eigenvalue and singular value decompositions for extracting information from time-series/sensor array data. These tools are used in the so-called subspace methods that underlie solutions to the harmonic retrieval problem in time series and the directions-of-arrival (DOA) estimation problem in array processing. The subspace methods require the knowledge of eigenvectors of the underlying covariance matrix to estimate the parameters of interest. Eigenstructure estimation in signal processing has two important classes: (i) estimating the eigenstructure of the given covariance matrix and (ii) updating the eigenstructure estimates given the current estimate and new data. In this paper, we survey some algorithms for both these classes useful for harmonic retrieval and DOA estimation problems. We begin by surveying key results in the literature and then describe, in some detail, energy function minimization approaches that underlie a class of feedback neural networks. Our approaches estimate some or all of the eigenvectors corresponding to the repeated minimum eigenvalue and also multiple orthogonal eigenvectors corresponding to the ordered eigenvalues of the covariance matrix. Our presentation includes some supporting analysis and simulation results. We may point out here that eigensubspace estimation is a vast area and all aspects of this cannot be fully covered in a single paper. (C) 1995 Academic Press, Inc.

Symmetrizing a Hessenberg matrix: Designs for VLSI parallel processor arrays

A symmetrizer of a nonsymmetric matrix A is the symmetric matrix X that satisfies the equation XA = A(t)X, where t indicates the transpose. A symmetrizer is useful in converting a nonsymmetric eigenvalue problem into a symmetric one which is relatively easy to solve and finds applications in stability problems in control theory and in the study of general matrices. Three designs based on VLSI parallel processor arrays are presented to compute a symmetrizer of a lower Hessenberg matrix. Their scope is discussed. The first one is the Leiserson systolic design while the remaining two, viz., the double pipe design and the fitted diagonal design are the derived versions of the first design with improved performance.

Dry sliding wear of Al alloy 2024-Al203 particle metal matrix composites

In the present investigation, Al 2024-15vol.%Al2O3 particulate (average size, 18 mu m) composites were fabricated using the liquid metallurgy route. The wear and friction characteristics of Al alloy 2024 and Al 2024-15vol.%Al2O3p, composite in the as-extruded and peak-aged conditions were studied using a pin-on-disc machine (with a steel disc as the counterface material). The worn surfaces, subsurfaces and the debris were analysed in a scanning electron microscope.The performance of the composite in the as-extruded condition is slightly inferior to that of the unreinforced alloy. However, in the T6 condition, although the wear rates of two materials are initially comparable, the unreinforced alloy seizes while the composite does not within the tested range employed. In the as-extruded condition, the presence of Al2O3 particles is not particularly beneficial as they fracture and result in extensive localized cracking and removal of material from the surface. In the peak-aged condition, however, while the unreinforced alloy exhibits severe plastic deformation and undergoes seizure, there is no significant change in the mechanism in the case of the composite. Except in the case of the peak-aged unreinforced alloy, worn surfaces of all other materials show the presence of an iron-rich layer.

Heterogeneous nucleation of nanometre-sized Bi particles embedded in a Zn matrix

It is argued that the nanometric dispersion of Bi in a Zn matrix is an ideal model system for heterogeneous nucleation experiments. The classical theory of heterogeneous nucleation with a hemispherical cap model is applied to analyse the nucleation data. It is shown that, unlike the results of earlier experiments, the derived site density for catalytic nucleation and contact angle are realistic and strongly suggest the validity of the classical theory. The surface energy between the 0001 plane of Zn and the <10(1)over bar 2> plane of Bi, which constitute the epitaxial nucleation interface, is estimated to be 39 mJ m(-2).

Intraloop and Interloop Interactions in the Folded-G Quartet Structure of Oxytricha Telomeric Sesquence

The antiparallel intramolecular G quartet structure for the 3.5 copy Oxytricha telomeric sequence d(G(4)T(4))(3)G4 has been established using a combination of spectroscopic and chemical probing methods. In the presence of Naf ions, this sequence exhibits a circular dichroism spectrum with a positive band at 295 nm and a negative band around 265 nm, characteristic of an antiparallel G quartet structure. Further, we show that d(G(4)T(4))(3)G(4) adopts an antiparallel intramolecular G quartet structure even in K+ unlike d(G(4)T(4)G(4)). KMnO4 probing experiments indicated the existence of intra and interloop interactions in the Na+ induced structure. We have found that K+ not only increases the thermal stability of,G quartet structure but also binds to the loop region and disrupts stacking and interloop interactions. Biological consequences of such cation-dependent conformational micro-heterogeneity in the loop region of G quartet structures is also discussed.

Production and properties of carbon fibre/particle reinforced cast metal matrix composites

Carbon fibres/particles can be satisfactory reinforcing material in polymer, ceramic and metal matrices. Carbon fibres/particles reinforced polymer matrix composites and ceramic matrix composites are being used extensively in critical areas of application, but carbon fibre - metal matrix composites have not reached that stage yet. This paper discusses the salient aspects of production and specific properties of carbon fibre/particle reinforced cast metal matrix composites. It is envisaged that these materials will find extensive applications where cost, weight and thermal expansion are the key factors.

Microstructure evolution during solidification of DRMMCs (discontinuously reinforced metal matrix composites): State of art

Distribution of particle reinforcements in cast composites is determined by the morphology of the solidification front. Interestingly, during solidification, the morphology of the interface is intrinsically affected by the presence of dispersed reinforcements. Thus the dispersoid distribution and length scale of matrix microstructure is a result of the interplay between these two. A proper combination of material and process parameters can be used to obtain composites with tailored microstructures. This requires the generation of a broad data base and optimization of the complete solidification process. The length scale of soldification microtructure has a large influence on the mechanical properties of the composites. This presentation addresses the concept of a particle distribution map which can help in predicting particle distribution under different solidification conditions Future research directions have also been indicated.