Michal Mrozowski

Macromodeling of Multiport Systems Using a Fast Implementation of the Vector Fitting Method

Broadband macromodeling of large multiport systems by vector fitting can be time consuming and resource demanding when all elements of the system matrix share a common set of poles. This letter presents a robust approach which removes the sparsity of the block-structured least-squares equations by a direct application of the QR decomposition. A 60-port printed circuit board example illustrates that considerable savings in terms of computation time and memory requirements are obtained.

Simple treatment of multi-term dispersion in FDTD

Three new simple and efficient algorithms are proposed for the numerical treatment of the multi-term Debye or Lorentz dispersion in the FDTD method. The formulation is based on the auxiliary differential equation, but requires much fewer operations than the published schemes. The approach is equivalent to the best higher order recursive schemes in terms of memory and computational expense, but without the linearity assumptions.

On the mode coupling in longitudinally magnetized waveguiding structures

Propagation in waveguiding structures containing gyromagnetic material with longitudinal magnetization is analyzed in terms of the coupling between modes having different symmetry properties. It is found that the wave supported by a symmetrical structure of two coupled guides is a combination of the even and odd modes propagating in an isotropic structure. The gyromagnetic medium causes the coupling between the modes, and the energy of the wave is periodically transferred from one mode to another, which in turn results in the exchange of energy between guides. The proposed mathematical model explains the operation of novel nonreciprocal devices. >

How to Render FDTD Computations More Effective Using a Graphics Accelerator

Graphics processing units (GPUs) for years have been dedicated mostly to real time rendering. Recently leading GPU manufactures have extended their research area and decided to support also graphics computing. In this paper, we describe an impact of new GPU features on development process of an efficient finite difference time domain (FDTD) implementation.

Equivalent SPICE Circuits With Guaranteed Passivity From Nonpassive Models

In this paper, a new fast technique of passivity enforcement of a nonpassive rational model is introduced. The technique disturbs poles of the model to restore passivity in such way that the frequency response of a device being modeled is preserved. The passivity enforcement procedure is defined as an optimization routine with the gradients of the cost function evaluated using the theory of matrix perturbation. The rational model can be based either on passive (electromagnetic simulations, measurements) or nonpassive (surrogate models) data. In the second case, the proposed technique can lead to a parameterized SPICE networks. Some advanced examples are given to show the application of proposed approach in interconnect, packaging, and signal integrity analysis

A Memory Efficient and Fast Sparse Matrix Vector Product on a GPU

This paper proposes a new sparse matrix storage format which allows an e-cient implementation of a sparse matrix vector product on a Fermi Graphics Processing Unit (GPU). Unlike previous formats it has both low memory footprint and good throughput. The new format, which we call Sliced ELLR-T has been designed speciflcally for accelerating the iterative solution of a large sparse and complex-valued system of linear equations arising in computational electromagnetics. Numerical tests have shown that the performance of the new implementation reaches 69 GFLOPS in complex single precision arithmetic. Compared to the optimized six core Central Processing Unit (CPU) (Intel Xeon 5680) this performance implies a speedup by a factor of six. In terms of speed the new format is as fast as the best format published so far and at the same time it does not introduce redundant zero elements which have to be stored to ensure fast memory access. Compared to previously published solutions, signiflcantly larger problems can be handled using low cost commodity GPUs with limited amount of on-board memory.

Efficient implementation of the Cauchy method for automated CAD-model construction

An efficient and stable implementation of the multidimensional Cauchy method for creating high quality multivariate models of microwave circuits from electromagnetic (EM) simulation data is presented in this paper. The algorithm uses the total least squares method to solve the ill-conditioned interpolation problem and automatically determines the model order and distribution of support points in the parameter space in such a manner that instabilities are prevented. Numerical tests show that the method requires fewer support points to achieve similar accuracy as the rational function models published previously. The utility and accuracy of the technique is demonstrated on a filter design example involving three- and five-dimensional models.

A conductive wedge in Yee's mesh

The authors present a correction of the finite-difference (FD) methods employing Yees mesh, improving the accuracy when conductive edge corners are present in the analyzed structure. The algorithm allows the wedge to be arbitrarily located with respect to the grid points and yields significant error reduction.

Automated CAD of coupled resonator filters

A gradient-based optimization technique along with a new definition of cost function is applied to the CAD of coupled resonator filters. The topology of the structure is enforced at each step of optimization and its physical dimensions are used as optimization variables. The cost function is defined using location of zeros and poles of the filters transfer and reflection functions. Numerical tests show that with the new definition of the cost function, the optimization process converges from an arbitrarily selected starting point. This allows one to design filters even without a rough microwave synthesis which usually provides initial dimensions.

Finite Element Matrix Generation on a GPU

This paper presents an e-cient technique for fast gener- ation of sparse systems of linear equations arising in computational electromagnetics in a flnite element method using higher order ele- ments. The proposed approach employs a graphics processing unit (GPU) for both numerical integration and matrix assembly. The per- formance results obtained on a test platform consisting of a Fermi GPU (1x Tesla C2075) and a CPU (2x twelve-core Opterons), indicate that the GPU implementation of the matrix generation allows one to achieve speedups by a factor of 81 and 19 over the optimized single- and multi-threaded CPU-only implementations, respectively.