SimPL: An Effective Placement Algorithm

Authors:
Myung-Chul Kim;Dong-Jin Lee;Igor L. Markov
Affiliations:
Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, MI, USA;Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, MI, USA;Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, MI, USA
Venue:
IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Year:
2012

Citing 0
Cited 8

MAPLE: multilevel adaptive placement for mixed-size designs

Proceedings of the 2012 ACM international symposium on International Symposium on Physical Design
ComPLx: A Competitive Primal-dual Lagrange Optimization for Global Placement

Proceedings of the 49th Annual Design Automation Conference
Progress and challenges in VLSI placement research

Proceedings of the International Conference on Computer-Aided Design
Placement: hot or not?

Proceedings of the International Conference on Computer-Aided Design
Sub-quadratic objectives in quadratic placement

Proceedings of the Conference on Design, Automation and Test in Europe
Taming the complexity of coordinated place and route

Proceedings of the 50th Annual Design Automation Conference
Ripple 2.0: high quality routability-driven placement via global router integration

Proceedings of the 50th Annual Design Automation Conference
Cell density-driven detailed placement with displacement constraint

Proceedings of the 2014 on International symposium on physical design

Quantified Score

Hi-index	0.03

Visualization

Abstract

We propose a self-contained, flat, quadratic global placer that is simpler than existing placers and easier to integrate into timing-closure flows. It maintains lower-bound and upper-bound placements that converge to a final solution. The upper-bound placement is produced by a novel look-ahead legalization algorithm. Our placer SimPL outperforms mPL6, FastPlace3, NTUPlace3, APlace2, and Capo simultaneously in runtime and solution quality, running 7.10 times faster than mPL6 (when using a single thread) and reducing wirelength by 3% on the ISPD 2005 benchmark suite. More significant improvements are achieved on larger benchmarks. The new algorithm is amenable to parallelism, and we report empirical studies with SSE2 instructions and up to eight parallel threads.