Reproduce “Bayesian Optimization for Macro Placement” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Changyong Oh, Roberto Bondesan, Dana Kianfar, Rehan Ahmed, Rishubh Khurana, Payal Agarwal, Romain Lepert, Mysore Sriram and 1 more
The authors use Bayesian optimization to place macros, fitting a Gaussian-process model to sequence-pair layouts and selecting diverse candidates for parallel evaluation. Across five MCNC benchmarks, it yields lower half-perimeter wire length than simulated annealing at equal evaluation counts. On ami33, results are 78,359 ± 1,271 versus 97,691 ± 1,592. The test uses a simplified objective without standard-cell placement.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.