MILU preconditioned iterative methods are useful for solving large spa
rse linear systems, which arise from the finite difference approximati
ons of three-dimensional second-order elliptic partial differential eq
uations. In these schemes, the MILU preconditioning which accelerates
the convergence is the most difficult part to vectorize on vector supe
rcomputers. Currently, the reordering techniques like the multicolor o
rdering strategy are commonly used to obtain sufficiently long vector
lengths. However, these reordering techniques deteriorate the converge
nce compared with the standard MILU preconditioner, and Gustafsson's a
cceleration does not work well for them. In this paper, a new precondi
tioning is proposed and its advantages compared with the multicolor an
d hyperplane orderings are shown through experiments on NEC vector sup
ercomputer SX-3.