Floorplanning

The input to the floorplanning problem is a set of blocks (soft or hard) and its netlist. The goal of floorplanning is to determine the location of the blocks so that the blocks do not overlap with each other. In case of soft blocks, we determine their dimensions as well. Traditional objectives include area and wirelength, and modern floorplanners address thermal hotspot, power supply noise, etc. Floorplanning problem is often solved under such design constraints as fixed outline, pre-placed blocks, alignment constraint, etc. This chapter presents sample problems related to the following works:

• Stockmeyer algorithm [Stockmeyer, 1983]

• Normalized polish expression [Wong and Liu, 1986]

• ILP-based floorplanning algorithm [Sutanthavibul et al., 1991]

• Sequence pair representation [Murata et al., 1995]

The first work determines the optimal orientation of the blocks in a given slicing floorplan. The second work presented an efficient way to represent slicing floorplans. The third work performs floorplanning for soft blocks with the integer linear programming (ILP) approach. The last work presented an effi- cient way to represent non-slicing floorplans. These works are targeting the traditional objectives such as area and wirelength.