AI ASIC floor planning (20250420)

i hadn’t been following the drama associated with this ASIC planning paper in nature. (A graph placement methodology for fast chip design | Nature) note, the nature article is behind a paywall.

this CACM article ([Reevaluating Google’s Reinforcement Learning for IC Macro Placement]https://dl.acm.org/doi/10.1145/3676845)) was a fascinating read. there’s a lot to unpack in here. but suffice it to say that that matter doesn’t appear to be closed and it looks like the google researchers have a lot of gaps to close.