Analytical placement for rectilinear blocks

This paper proposes a fast analytical placement for rectilinear blocks. The LSI production method is also improved. As a result, there are much large number of elements on one chip. On the other hand, its turn-around-time is same as or less than those of the previous designs. To solve this difficulty, the reuse of the designed modules, that is, IPs, is promising. However, the LSI production improvement also leads larger numbers of IPs, while the shape of IP seems to be rectilinear. Thus, a fast analytical placement for rectilinear blocks is needed. In this paper, we enhance Overlap-Removable Length which has been introduced to the rectangle block placement to rectilinear block placement. We show the efficiency of the proposed method empirically.