A new preconditioner for CGS iteration in solving large sparse nonsymmetric linear equations in semiconductor device simulation

The authors present a new and efficient precondition method (ILUV-CGS), which is an incomplete LU (ILU) factorization by value instead of a sparse nonzero pattern, in conjunction with CGS iteration for solving nonsymmetric linear equations in finite-element simulation of semiconductor devices. To demonstrate the efficiency of this method, the simulations of a group of representative devices with different numerical characteristics have been carried out using the ILUV-CGS algorithm, and these are compared with those of the existing ILU-CGS method. The comparison results are in favor of using the ILUV=CGS method.>