Toward Optimal Legalization for Mixed-Cell-Height Circuit Designs

Chen, Jianli; Zhu, Ziran; Wang, Zihao; Chang, Yao‐Wen

doi:10.1145/3061639.3062330

Cited by 45 publications

(20 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Meanwhile, Chow et al [16] propose a new algorithm that enumerates valid insertion points for cells and chooses the one that results in minimum displacement. Chen et al [3] introduce a new model for the problem as a Linear Complementary Problem by relaxing some of the problem constraints. As a result, they manage to legalize the whole circuit simultaneously, achieving a low perturbation.…”

Section: Related Workmentioning

confidence: 99%

“…Circuit fabrication is enabled by physical synthesis, which is responsible for placing and routing the cells' layouts on a 2-D surface [1]. In the physical synthesis flow the global placement step finds an initial solution for the circuit while minimizing interconnection wirelength [2] and a few other constraints such as pin accessibility [3]. However, to be able to find solutions for circuits with millions of gates within acceptable runtime, global placers disregard some layout constraints such as cell overlaps and cell alignment to rows and sites.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Enhancing Multi-Threaded Legalization Through $k$-d Tree Circuit Partitioning

Fabre

Güntzel

Pilla

et al. 2018

2018 31st Symposium on Integrated Circuits and Systems Design (SBCCI)

View full text Add to dashboard Cite

In the physical synthesis of integrated circuits the legalization step may move all circuit cells to fix overlaps and misalignments. While doing so, it should cause the smallest perturbation possible to the solution found by previous optimization steps to preserve placement quality. Legalization techniques must handle circuits with millions of cells within acceptable runtimes, besides facing other issues such as mixed-cell-height and fence regions. In this work we propose a k-d tree data structure to partition the circuit, thus removing data dependency. Then, legalization is sped up through both input size reduction and parallel execution. As a use case we employed a modified version of the classic legalization algorithm Abacus. Our solution achieved a maximum speedup of 35 times over a sequential version of Abacus for the circuits of the ICCAD2015 CAD contest. It also provided up to 10% reduction on the average cell displacement.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Enhancing Multi-Threaded Legalization Through $k$-d Tree Circuit Partitioning

Fabre

Güntzel

Pilla

et al. 2018

2018 31st Symposium on Integrated Circuits and Systems Design (SBCCI)

View full text Add to dashboard Cite

show abstract

“…After selecting the best ML model, it was retrained with the same parameters using the Table 3 shows the results of the ML models on the validation set for three different maximum displacement thresholds (∆): 5, 10, and 15 rows. We evaluated the models by their accuracy, precision, recall and F-score 3 . We selected the best results obtained for each model.…”

Section: Methodsmentioning

confidence: 99%

“…Instead of handling each cell at a time, the works from [10] and [3] propose algorithms that simultaneously legalize multiple cells. The authors of [10] propose an Integer Linear Programming (ILP) model to solve the multi-row legalization problem.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

How Deep Learning Can Drive Physical Synthesis Towards More Predictable Legalization

Netto

Fabre

Fontana

et al. 2019

Proceedings of the 2019 International Symposium on Physical Design

View full text Add to dashboard Cite

Machine learning has been used to improve the predictability of different physical design problems, such as timing, clock tree synthesis and routing, but not for legalization. Predicting the outcome of legalization can be helpful to guide incremental placement and circuit partitioning, speeding up those algorithms. In this work we extract histograms of features and snapshots of the circuit from several regions in a way that the model can be trained independently from region size. Then, we evaluate how traditional and convolutional deep learning models use this set of features to predict the quality of a legalization algorithm without having to executing it. When evaluating the models with holdout cross validation, the best model achieves an accuracy of 80% and an F-score of at least 0.7. Finally, we used the best model to prune partitions with large displacement in a circuit partitioning strategy. Experimental results in circuits (with up to millions of cells) showed that the pruning strategy improved the maximum displacement of the legalized solution by 5% to 94%. In addition, using the machine learning model avoided from 22% to 99% of the calls to the legalization algorithm, which speeds up the pruning process by up to 3×.

show abstract