memory requirements of CEM methods ?

I can find the memory requirements of MOM ( N*N - N is unknown number; also run time is N*N*N) and MLFMM (Nlog(N) and run time is N*log(N)*log(N)) but I couldn't find the requirements of other methods esipecially time domain solutions like FDTD, TLM ....

What are the memory requirements of these methods ?

hello,

FDTD and TLM are time domain solvers. These are typically iterative solvers. While MoM, FEM... are frequency domain solvers involving a large Z-matrix to fill and to invert. So the memory requirement and speed are determined by the Z-matrix filling and inverting algorithms. Time domain solvers do not involve large matrices to invert. The fields are put into the solution domain using a pulse and the fields are calculated at each time step as they travel through the solution domain. Some power of the pulse is reflected towards the excitation, this allowes for S11 calculation. Time domain solver's speed is determined by time step (depence on smallest cell size). Time domain solvers have converged to a solution if the reflected power has decayed to a predetermined value (e.g. -40dB). Structures with high Q reflect more power of the impuls and converge a lot slower. The frequency domain solvers do not suffer this structure dependence. In summary, it is difficult to define the memory requirements and speed of time domain solvers.

best regards

I agree with the first reply, the time-domain solutions such as FDTD are not matrix-fill-and-solve methods. They are fundamentally different since they are a very direct numerical solution to Maxwell's equations. Most commercial time-domain tools are very memory efficient (CST-MS, Microstripes, XFDTD, etc.). For example, XFDTD uses only 30 MB RAM for 1 million grid cells. Many of them use automatic graded subgridding for fine geometry modeling, which further improves the memory efficiency.

FEM produces a sparse, banded matrix as opposed to MoM which produces a full matrix. For a code such as HFSS, the matrix contains approximately as many elements as there are tetrahedra nodes. RAM requirements of FEM tools can be pretty hefty. A scheme such as adaptive meshing can help decrease the amount of unknowns in the FEM matrix by only adding tetrahedra where they are required.

Also, running on a 64-bit OS can get around the 2 GB memory addressing limitation of 32-bit systems.