In this paper, a novel hybrid algorithm for Very Large Scale Integration ce
ll placement is presented. The algorithm consists of two stages. The first
stage uses distributed space exploration to produce an acceptable intermedi
ate placement. The second stage uses a parallel fine grain algorithm to loc
ally explore subspace of the intermediate placement. The resulting placemen
t quality of this strategy is equivalent to the standard for comparison in
academia and industry-TimberWolf. However, our algorithm has an expected sp
eedup improvement of at least two orders of magnitude over TimberWolf.