Performance of an embedded optical vector matrix multiplication processor architecture

An embedded architecture of optical vector matrix multiplier (OVMM) is presented. The embedded architecture is aimed at optimising the data flow of vector matrix multiplier (VMM) to promote its performance. Data dependence is discussed when the OVMM is connected to a cluster system. A simulator is built to analyse the performance according to the architecture. According to the simulation, Amdahl's law is used to analyse the hybrid opto-electronic system. It is found that the electronic part and its interaction with optical part form the bottleneck of system.

Efficient two-step digit-set-restricted modified signed-digit algorithm based on optoelectronic shared content-addressable memory

A two-step digit-set-restricted modified signed-digit (MSD) adder based on symbolic substitution is presented. In the proposed addition algorithm, carry propagation is avoided by using reference digits to restrict the intermediate MSD carry and sum digits into {(1) over bar ,0} and {0, 1}, respectively. The algorithm requires only 12 minterms to generate the final results, and no complementarity operations for nonzero outputs are involved, which simplifies the system complexity significantly. An optoelectronic shared content-addressable memory based on an incoherent correlator is used for experimental demonstration. (c) 2005 Society of Photo-Optical Instrumentation Engineers.

Efficient two-step digit-set-restricted modified signed-digit algorithm based on optoelectronic shared content-addressable memory

A two-step digit-set-restricted modified signed-digit (MSD) adder based on symbolic substitution is presented. In the proposed addition algorithm, carry propagation is avoided by using reference digits to restrict the intermediate MSD carry and sum digits into {(1) over bar ,0} and {0, 1}, respectively. The algorithm requires only 12 minterms to generate the final results, and no complementarity operations for nonzero outputs are involved, which simplifies the system complexity significantly. An optoelectronic shared content-addressable memory based on an incoherent correlator is used for experimental demonstration. (c) 2005 Society of Photo-Optical Instrumentation Engineers.

GOTOBLAS一般矩阵乘法高效实现机制的研究

对GOTOBLAS库(GOTO)的实现机制,尤其是其中的一般矩阵乘法部分的实现进行了分析.结合近年来的一些研究成果,讨论了如何高效地实现矩阵相乘操作,把存储层次对程序性能的影响提高到计算模型的高度.对比实验表明,GOTO库的性能远远高于没有考虑存储层次的一般BLAS库.证明了GOTO库性能上的优越性和将存储层次引入计算模型的必要性.

MODIFIED DIRECT TWOS-COMPLEMENT PARALLEL ARRAY MULTIPLICATION ALGORITHM FOR COMPLEX MATRIX OPERATION

A direct twos-complement parallel array multiplication algorithm is introduced and modified for digital optical numerical computation. The modified version overcomes the problems encountered in the conventional optical twos-complement algorithm. In the array, all the summands are generated in parallel, and the relevant summands having the same weights are added simultaneously without carries, resulting in the product expressed in a mixed twos-complement system. In a two-stage array, complex multiplication is possible with using four real subarrays. Furthermore, with a three-stage array architecture, complex matrix operation is straightforwardly accomplished. In the experiment, parallel two-stage array complex multiplication with liquid-crystal panels is demonstrated.

Strain Gradient Effects on Deformation Strengthening Behavior of Particle Reinforced Metal Matrix Composites

Dislocation models with considering the mismatch of elastic modulus between matrix and reinforcing particles are used to determine the effective strain gradient \ita for particle reinforced metal matrix composites (MMCp) in the present research. Based on Taylor relation and the kinetics of dislocation multiplication, glide and annihilation, a strain gradient dependent constitutive equation is developed. By using this strain gradient-dependent constitutive equation, size-dependent deformation strengthening behavior is characterized. The results demonstrate that the smaller the particle size, the more excellent in the reinforcing effect. Some comparisons with the available experimental results demonstrate that the present approach is satisfactory.

MODIFIED ENGAGEMENT METHOD FOR MATRIX OPERATION

We describe a modified engagement method for matrix operation based on a two-dimensional crossed-ring interconnection network, Our method incorporates fewer steps than that reported by Bocker et al. [Appl. Opt. 22, 804 (1983)], and its performance is found to be the most efficient (minimum steps) in comparison with other systolic and/or engagement methods for matrix operation. Thus, it may be helpful for other optical and electronic implementations of matrix operations. One compact optoelectronic integrity approach for implementing the modified engagement method is briefly described. (C) 1995 Optical Society of America

Fast calculation technique for scattering in T-matrix method

In scattering calculations using the T-matrix method, the calculation of the T-matrix involves multiplication and inversion of matrices. These two types of matrix operations are time-consuming, especially for the matrices with large size. Petrov et al. [D. Petrov, Y. Shkuratov, G. Videen, Opt. Lett. 32 (2007) 1168] proposed an optimized matrix inversion technique, which suggests the inversion of two matrices, each of which contains half the number of rows. This technique reduces time-consumption significantly. On the basis of this approach, we propose another fast calculation technique for scattering in the T-matrix method, which obtains the scattered fields through carrying out only the operations between matrices and the incident field coefficient. Numerical results show that this technique can decrease time-consumption by more than half that of the optimized matrix inversion technique by Petrov et al. (c) 2008 Elsevier B.V. All rights reserved.

A transfer matrix approach to conductance in quantum waveguides

A transfer matrix approach is presented for the study of electron conduction in an arbitrarily shaped cavity structure embedded in a quantum wire. Using the boundary conditions for wave functions, the transfer matrix at an interface with a discontinuous potential boundary is obtained for the first time. The total transfer matrix is calculated by multiplication of the transfer matrix for each segment of the structure as well as numerical integration of coupled second-order differential equations. The proposed method is applied to the evaluation of the conductance and the electron probability density in several typical cavity structures. The effect of the geometrical features on the electron transmission is discussed in detail. In the numerical calculations, the method is found to be more efficient than most of the other methods in the literature and the results are found to be in excellent agreement with those obtained by the recursive Green's function method.

Size-dependent inelastic behavior of particle-reinforced metal-matrix composites

The strengthening behavior of particle-reinforced metal-matrix composites (MMCp) is primarily attributed to the dislocation strengthening effect and the load-transfer effect. To account for these two effects in a unified way, a new hybrid approach is developed in this paper by incorporating the geometrically necessary dislocation strengthening effect into the incremental micromechanical scheme. By making use of this hybrid approach, the particle-size-dependent inelastic deformation behavior of MMCp is given. Some comparisons with the available experimental results demonstrate that the present approach is satisfactory.

An Amino Acid Substitution Matrix for Protein Conformation Identification

Amino acid substitution matrices play an essential role in protein sequence alignment, a fundamental task in bioinformatics. Most widely used matrices, such as PAM matrices derived from homologous sequences and BLOSUM matrices derived from aligned segments of PROSITE, did not integrate conformation information in their construction. There are a few structure-based matrices, which are derived from limited data of structure alignment. Using databases PDB_SELECT and DSSP, we create a database of sequence-conformation blocks which explicitly represent sequence-structure relationship. Members in a block are identical in conformation and are highly similar in sequence. From this block database, we derive a conformation-specific amino acid substitution matrix CBSM60. The matrix shows an improved performance in conformational segment search and homolog detection.

Flow Stress of Biomorphous Metal-Matrix Composites

For metal-matrix composites (MMCs), interfacial debonding between the ductile matrix and the reinforcing hard inclusions is an important failure mode. A fundamental approach to improving the properties of MMCs is to optimize their microstructure to achieve maximum strength and toughness. Here, we investigate the flow stress of a MMC with a nanoscale microstructure similar to that of bone. Such a 'biomorphous' MMC would be made of staggered hard and slender nanoparticles embedded in a ductile matrix. We show that the large aspect ratio and the nanometer size of inclusions in the biomorphous MMC lead to significantly improved properties with increased tolerance of interfacial damage. In this case, the partially debonded inclusions continue to carry mechanical load transferred via longitudinal shearing of the matrix material between neighboring inclusions. The larger the inclusion aspect ratio, the larger is the flow stress and work hardening rate for the composite. Increasing the volume concentration of inclusion also makes the biomorphous MMC more tolerant of interfacial damage.

Material response and failure mechanism of unidirectional metal matrix composites under impulsive shear loading

The material response and failure mechanism of unidirectional metal matrix composite under impulsive shear loading are investigated in this paper. Both experimental and analytical studies were performed. The shear strength of unidirectional C-f/A356.0 composite and A356.0 aluminum alloy at high strain rate were measured with a modified split Hopkinson torsional bar technique. The results indicated that the carbon fibers did not improve the shear strength of aluminum matrix if the fiber orientation aligned with the shear loading axis. The microscopic inspection of the fractured surface showed a multi-scale zigzag feature which implied a complicated shear failure mechanism in the composite. In addition to testing, the micromechanical stress field in the composite was analyzed by the generalized Eshelby equivalent method (GEEM). The influence of cracking in matrix on the micromechanical stress field was investigated as well. The results showed that the stress distribution in the composite is quite nonhomogeneous and very high shear stress concentrations are found in some regions in the matrix. The high shear stress concentration in the matrix induces tensile cracking at 45 degrees to the shear direction. This in turn aggravates the stress concentration at the fiber/matrix interface and finally leads to a catastrophic failure in the composite. From the correlation between the analysis and experimental results, the shear failure mechanism of unidirectional C-f/A356.0 composite can be elucidated qualitatively.

Enhanced Plasticity of Zr-Based Bulk Metallic Glass Matrix Composite with Ductile Reinforcement

A composite material containing uniformly distributed micrometer-sized Nb particles in a Zr-based amorphous matrix was prepared by suction cast. The resulting material exhibits high fractured strength over 1550 MPa and enhanced plastic strain of about 29.7% before failure in uniaxial compression test at room temperature. Studies of the serrations on the stress-strain curves and the shear bands on the fractured samples reveal that the amplitude of the stress drop of each serration step corresponds to the extent of the propagation of a single shear band through the materials. The composite exhibits more serration steps and smaller amplitude of stress drop due to the pinning of shear band propagation by ductile Nb particles.

Formation, Thermal Stability and Deformation Behavior of Graphite-Flakes Reinforced Cu-based Bulk Metallic Glass Matrix Composites

Graphite-flake reinforced Cu47Ti34Zr11 Ni-8 bulk metallic glass matrix composite was fabricated by water-cooled copper mould cast. Most of the graphite flakes still keep unreacted and distribute uniformly in the amorphous matrix except that some reactive wetting occurs by the formation of TiC particles around the flakes. It reveals that the presence of graphite flakes does not affect the onset of the glass transition temperature, crystallization reaction and liquidus of the metallic glass. The resulting material shows obvious serrated flow and higher fracture strength under room temperature compressive load, comparing with the monolithic bulk metallic glass (BMG). Three types of interaction between the shear bands and graphite flakes, namely, shear band termination, shear bands branching and new shear bands formation near the graphite flakes can be observed by quasi-static uniaxial compression test and bonded interface technique through Vickers indentation.