The impact of supercomputers on IC technology development and design

Very large scale integration (VLSI) has evolved at an enormous rate, progressing from hundreds of components on an integrated circuit (IC) in the 1960's to a million components on a chip in the foreseeable future. This paper reviews some of the computer-aided design (CAD) tools that are essential for VLSI technology development and circuit design and that also require large amounts of computer resources. Specifically, we describe programs for process simulation, device simulation, and circuit simulation. This paper also reviews the impact of high-performance computing facilities on the development and use of these programs at AT & T Bell Laboratories.     

Parallel circuit simulation on supercomputers

Circuit simulation is a very time-consuming and numerically intensive application, especially when the problem size is large as in the case of VLSI circuits. To improve the performance of circuit simulators without sacrificing accuracy, a variety of parallel processing algorithms have been investigated. Research in the field of parallel circuit simulation is surveyed, and the ongoing research in this area at the University of Illinois is described. Both standard and relaxation-based approaches are considered. In particular, the forms of parallelism available within the direct method approach, used in programs such as SPICE2 and SLATE, and within the relaxation-based approaches, such as waveform relaxation, iterated timing analysis, and waveform-relaxation-Newton, are described. The specific implementation issues addressed are primarily related to general-purpose multiprocessors with a shared-memory architecture having a limited number of processors, although many of the comments apply to a number of other architectures     

Artificial intelligence applications on supercomputers

In contrast to many scientific applications that perform computations on fixed-size arrays, symbolic programs process complex data structures represented as lists, sets, and graphs that are allocated and deallocated in a variable pattern during execution. Much of the processing in symbolic computation involves following linked lists by tracing chains of pointers through scattered locations in memory. Execution speeds for single serial processors are therefore limited by the time it takes to make random retrievals of individual words of data from memory.     

Scalar finite-element analysis of optical-fiber facets

A numerically efficient scalar analysis of optical-fiber-facet problems based on the finite-element scheme is presented. By adopting the Taylor's series expansion of the characteristic matrix at the discontinuity plane, an accurate and yet numerically efficient approach is suggested for calculating the reflected and transmitted fields at discontinuities with circular symmetry. The scattering of the scalar LP/sub 01/ mode and higher order LP/sub 0m/ modes at both uncoated and coated optical-fiber facets has been analyzed, and the accuracy of the present finite-element approach is revealed through the excellent agreement of its results with those in the literature.     

D-Fiber antenna characterization using finite-element analysis

An all-fiber antenna using piezoelectric polymer coated circular core D-fiber has been characterized using finite-element analysis. The response of the D-fiber antenna was determined over a wide frequency range from 1 MHz to 2 GWz. The modeling predicts an electric field induced phase shift of 2.43×10^-6 rad/(V/m) per meter at 5 MHz. At frequencies higher than 8 MHz, the optical response is dominated by radial resonances of the D-fiber/coating composite. Using the simulation results, a minimum detectable electric field of 41 μV/m has been achieved using a 1 km length of coated D-fiber. In addition, a D-fiber antenna network intended for microcellular communications has been analyzed using shot noise limited detection. The D-fiber antenna has potential applications in areas such as electromagnetic compatibility testing and radio-over-fiber networks where it provides a convenient means of optically generating radio signals     

Two-dimensional semiconductor analysis using finite-element method

Two-dimensional simulation of semiconductor devices using a finite-element formulation is described. In the present analysis, Poisson's equation is solved by a finite-element method, based on the variational principle, and current continuity equations are solved by a method of weighted residuals. The advantage of this method is mentioned. In order to demonstrate the validity of this method, a bipolar n-p-n transistor is analyzed, considering the generation-recombination term. Not only voltage-current characteristic, but also junction capacitance and cutoff frequency are calculated. Then transistor behavior under inverse mode by using the n-type buried layer as a common emitter is discussed.     

An efficient finite-element analysis of magnetooptic channelwaveguides

A finite-element method based on the scalar-wave approximation is developed for the analysis of magnetooptic waveguides. A simple and efficient iterative method is proposed for solving a nonlinear eigenvalue equation derived from the scalar finite-element approach. To show the validity and usefulness of this method, examples are computed for magnetooptic rib-type and ridge-type waveguides. The waveguide structures which have larger nonreciprocal phase shift are discussed     

Accurate finite-element analysis of dual-mode highlyelliptical-core fibers

A scalar finite-element method is used for investigating propagation characteristics of dual-mode highly elliptical-core fibers. The dispersion property, polarization modal birefringence, and spatial modal birefringence are calculated. To improve the accuracy of solutions, isoparametric curvilinear elements are introduced. The applicability of the scalar finite-element method for highly elliptical-core fibers is assessed by comparing the results obtained with the vector finite-element method. An approximate simple approach, in which an elliptical-core is replaced by an appropriate rectangular-core, is also examined     

Quasi-static finite-element analysis of a skewed microstripcrossover

In this work, we present a quasistatic analysis of a microstrip crossover on a dielectric substrate. The microstrips are located at different planes and may cross at an arbitrary angle. Capacitances and inductances are calculated from scalar potentials. For magnetostatic formulation, the boundary conditions for scalar potential are introduced by means of partitioning surfaces. The use of the adaptive finite element method provides the required flexibility with respect to the analyzed geometry, optimal discretization and good efficiency     

A two-dimensional finite-element analysis of reverberation chambers

A two-dimensional (2-D) analysis of reverberation chambers is performed at cutoff. The structure considered is lossless and corresponds to an infinite quality factor (Q) chamber. The concept of frequency stirring is used to generate field data for a discrete set of modes and the resulting statistics are analyzed. The field statistics are examined for TE and TM modes. This analysis yields statistics similar to the expected reverberation chamber statistics for the fields. Mechanical stirring is also examined and a connection to the peak-frequency deviation is presented     

Holey fiber analysis through the finite-element method

A holey fiber (HF), having very complex hole geometry, is studied by means of a numerical simulator for modal analysis based on the finite-element method (FEM). Polarization and dispersion properties as well as the full vector field distribution of the fundamental mode are investigated. The obtained numerical results show a good agreement with experimental ones reported in literature     

Three-dimensional edge-based finite-element analysis for discretebodies of revolution

An extension to three-dimensional (3-D) edge-based finite-element analysis for modeling electrically large fan-like bodies as discrete bodies of revolution is given. By exploiting the overlapping symmetries between a fan-like body and a modal expansion of the electromagnetic fields, only one lobe of the problem need be solved by the finite-element method without introducing approximations. This computational scaling makes possible the solution of electrically large structures much more efficiently. However, a periodic phase-boundary condition (PBC) must be applied to the faces of the mesh describing a single slice of the body and this condition must be enforced on both the electric and magnetic fields for a robust solution. Details on the implementation of the PBCs are given along with results which validate the overall technique     

Approximate scalar finite-element analysis of anisotropic optical waveguides

An approximate scalar finite-element program for the analysis of anisotropic optical waveguides with a diagonal permittivity tensor is described. The accuracy of the method has been checked by calculating the eigenmodes of two-dimensional, anisotropic asymmetric slab waveguides. The results obtained for a channel waveguide embedded in LiNbO3 agree well with the results of the earlier vectorial finite-element method.     

Full vectorial finite-element analysis of sharp optical waveguidecorners

The transmission efficiency of optical waves through a guided-wave structure, incorporating a sharp corner, is investigated by using rigorous numerical approaches based on the finite-element method. To show the versatility of the proposed numerical approaches, the modal power transmission coefficient of two rib waveguides is calculated, and the results are compared with those obtained by using other approaches     

New recovery error indicator for adaptive finite-element analysis on waveguiding structures

The adaptive finite-element method (FEM) is an iterative variant of the FEM where, in a first step, an initial mesh with few and low-order elements is generated, the corresponding algebraic problem is solved and the error in the solution is estimated in order to add degrees of freedom in those regions of the domain with the biggest error estimation. This process is repeated until an ending condition is reached. The two basic stages in this method are the error indication and the mesh enrichment. In this paper, within the analysis of waveguiding structures, a new error indicator based on the curl recovery is described. In addition, an overview on refinement techniques is presented, and the h-refinement employed in this study is briefly described. Results obtained with the curl-recovery indicator are discussed and compared with the classical nonadaptive FEM and two previously developed error indicators: the residual and flux continuity indicators.     

Automated finite-element analysis for deformable registration of prostate images

Two major factors preventing the routine clinical use of finite-element analysis for image registration are: 1) the substantial labor required to construct a finite-element model for an individual patient's anatomy and 2) the difficulty of determining an appropriate set of finite-element boundary conditions. This paper addresses these issues by presenting algorithms that automatically generate a high quality hexahedral finite-element mesh and automatically calculate boundary conditions for an imaged patient. Medial shape models called m-reps are used to facilitate these tasks and reduce the effort required to apply finite-element analysis to image registration. Encouraging results are presented for the registration of CT image pairs which exhibit deformation caused by pressure from an endorectal imaging probe and deformation due to swelling.     

Simple and efficient finite-element analysis of microwave andoptical waveguides

A simple and efficient finite-element method for the analysis of microwave and optical waveguiding problems is formulated using three components of the electric or magnetic field. In order to eliminate spurious solutions, edge elements are introduced. In the edge element approach the nodal parameters are not limited to the magnetic field as in the conventional three-component formulation for the dielectric waveguiding problem. An eigenvalue equation that involves only the edge variables in the transversal plane and can provide a direct solution for the propagation constant is derived. To show the validity and usefulness of this approach, computed results are illustrated for microstrip transmission lines and dielectric waveguides     

Coupling matrices for high-order finite-element analysis of acoustic-wave propagation

Finite-element coupling matrices, independent of element geometry and material properties, are presented for the high-order finite-element analysis of acoustic-wave propagation in homogeneous isotropic media. Some new results for ridge guides are used to illustrate the formulation.     

The nonunitarity of finite-element beam propagation algorithms

We examine the intrinsic nonunitarity of one-way finite-element propagation techniques in order to clarify the results of previous authors.     

科学家提出制造光子晶体管新理论

丹麦哥本哈根大学和美国哈佛大学的研究人员近日提出了一种制造光子晶体管(Photon-transistors)的新理论.