CSF: Fixed-outline Floorplanning Based on the Conjugate Subgradient Algorithm Assisted by Q-Learning Note: This work is supported in part by the Fundamental Research Funds for the Central University (No. 104972024KFYjc0055), the National Natural Science Foundation of China (No. 92373102,12161039), t
Xinyan Meng, Huabin Cheng, Rujie Chen, Ning Xu, Yu Chen, Wei Zhang · 2025
Read CSF: Fixed-outline Floorplanning Based on the Conjugate Subgradient Algorithm Assisted by Q-Learning Note: This work is supported in part by the Fundamental Research Funds for the Central University (No. 104972024KFYjc0055), the National Natural Science Foundation of China (No. 92373102,12161039), t by Xinyan Meng, Huabin Cheng, Rujie Chen, Ning Xu, Yu Chen, Wei Zhang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-03796.txt.
Introduction
Chip designers must fit blocks on a fixed outline while keeping wire lengths short. Xinyan Meng and co-authors propose a 2025 algorithm for that layout task, built on a slope-based method guided by machine learning.
Older gradient methods can stall at a merely local best layout. The new approach uses a model that allows sharp corners, plus a learning agent that adjusts step size as the search runs. That balance lets the search roam widely yet settle well. On standard test circuits, the method produced legal layouts faster than graph-based rivals and matched top tools for hard blocks. Faster, tighter layouts can cut chip cost and design time.
Catalog record
| CREATOR | Xinyan Meng, Huabin Cheng, Rujie Chen, Ning Xu, Yu Chen, Wei Zhang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 52 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-03796
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
CSF: Fixed-outline Floorplanning Based on the Conjugate Subgradient Algorithm Assisted by Q-Learning Note: This work is supported in part by the Fundamental Research Funds for the Central University (No. 104972024KFYjc0055), the National Natural Science Foundation of China (No. 92373102,12161039), the Natural Science Foundation of Jiangxi Province (No. 20224BAB201011), and the Foundation of Wuhan Institute of Technology (No. K2024045). Xinyan Meng, Huabin Cheng, Rujie Chen, Ning Xu, Yu Chen, Wei Zhang Abstract The state-of-the-art researches indicate that analytic algorithms are promising in handling complex floorplanning scenarios. However, it is challenging to generate compact floorplans with excellent wirelength optimization effect due to the local convergence of gradient-based optimization algorithms designed for constructed smooth optimization models. Accordingly, we propose to construct a nonsmooth analytic floorplanning model addressed by the conjugate subgradient algorithm (CSA), which is accelerated by a population-based scheme adaptively regulating the stepsize with the assistance of Q-learning. In this way, the proposed CSA assisted by Q-learning (CSAQ) can strike a good balance on exploration and exploitation. Experimental results on the MCNC and GSRC benchmarks demonstrate that the proposed fixed-outline floorplanning algorithm based on CSAQ (CSF) not only address global floorplanning effectively, but also get legal floorplans more efficiently than the constraint graph-based legalization algorithm as well as its improved variants. It is also demonstrated that the CSF is competitive to the state-of-the-art algorithms on floorplanning scenarios only containing hard modules. Journal: Elsevier † † highlights: The nonsmooth wirelength and overlapping metrics are directly optimized by conjugate subgradient algorithm (CSA). Since the CSA searches the solution space by both the deterministic gradient-based search and the stochastic exploration, it is much more likely to get the optimal floorplan by flexibly regulating its stepsize. The CSA is implemented in a population-based search pattern equipped with a dynamical stepsize regulated based on Q-learning, which contributes to its fast convergence to the optimal floorplan. Two improved variants are proposed for the constraint graph-based legalization algorithm. They take less time to get legal floorplans with shorter wirelength. We propose a CSA-based fixed-outline floorplanning (CSF) algorithm, where both global floorplanning and legalization are implemented by the CSA assisted by Q-learning (CSAQ). It can strike a balance between time complexity and floorplan quality, and demonstrates superior performance to the compared algorithms. 1 Introduction Floorplanning is a key stage in the physical design flow of very large-scale integration circuit (VLSI) [ 1 ] . With the rapid development of integrated circuit manufacturing technology, the complexity of VLSI design has increased dramatically, and modern design usually focuses on fixed-outline floorplanning (FOFP) problems [ 2 ] . Compared to layout planning problems without fixed outline, the FOFP problem is more challenging due to the additional constraint of placing a large number of modules within a given outline [ 3 ] . Additionally, the usage of intellectual property (IP) core introduces hard modules in the VLSI floorplan, making it much more challenging to optimize the fixed-outline floorplan of mixed-size modules [ 4 ] . Floorplanning algorithms generally fall into two categories: floorplanning algorithms based on combinatorial optimization model (FA-COM) and floorplanning algorithms based on analytic optimization model (FA-AOM). Coding the relative positions of modules by combinatorial data structures, FA-COM formulates floorplanning problems as combinatorial optimization problems addressed by heuristics and metaheuristics [ 5 , 6 , 7 , 8 ] . Although the combinatorial codes can be naturally converted into compact floorplans complying with the non-overlapping constraint, the combinatorial explosion contributes to their poor performance on large-scale cases. A remedy for this challenge is to implement multilevel methods, in which clustering or partitioning strategies are introduced to reduce the scale of the problem [ 3 ] . However, the multilevel strategy typically optimizes the floorplan in a local way, which in turn makes it difficult to converge to the globally optimal floorplan of large-scale VLSI designs [ 9 , 10 ] . FA-AOM address analytical floorplanning models by continuous optimization algorithms, contributing to their low complexity and fast convergence in large-scale cases [ 11 ] . A typical FA-AOM consists of two stages, the global floorplanning stage generating promising results, and the legalization stage that refines the positions of modules to eliminate constraint violations. To get a nice VLSI floorplan, the cooperative design of global floorplanning and legalization is extremely challenging due to the complicated characteristics of floorplanning models. The efficient global floorplanning is typically implemented by addressing a smoothed optimization model with a first-order gradient-based analytical algorithm [ 12 , 13 ] , however, the smoothing approximation could introduce some distortion to the characteristics of the mathematical model, and the first-order analytical algorithm usually converges to a local optimal solution. Legalization can be implemented through the refinement of constraint graphs [ 12 , 13 ] or the second-order cone programming strategy [ 14 ] , which cannot strike a good balance between time complexity and minimization of wirelength. To address the fixed-outline floorplanning problem efficiently, this paper proposes a fixed-outline floorplanning algorithm based on the conjugate subgradient algorithm (CSA), where the Q-learning is introduced to adaptively regulate the step size of CSA. The main contributions of this paper are as follows. 1. The nonsmooth wirelength and overlapping metrics are directly optimized by CSA. Since the CSA searches the solution space by both the deterministic gradient-based search and the stochastic exploration, it is much more likely to get the optimal floorplan by flexibly regulating its step size. 2. The CSA is implemented in a population-based search pattern equipped with a dynamical stepsize regulated based on Q-learning, which contributes to its fast convergence to the optimal floorplan. 3. Two improved variants are proposed for the CG-based legalization algorithm [ 12 , 13 ] . They take less time to get legal floorplans with shorter wirelength. 4. We propose a CSA-based fixed-outline floorplanning (CSF) algorithm, where both global floorplanning and legalization are implemented by the CSA assisted by Q-learning (CSAQ). It can strike a balance between time complexity and floorplan quality, and demonstrates superior performance to the compared algorithms. The remainder of this paper is organized as follows. Section 2 reviews related work. Section 3 introduces some preliminaries. Then, the proposed algorithm developed for fixed-outline floorplanning problems is presented in Sections 4 . The numerical experiment is performed in Section 5 to demonstrate the competitiveness of the proposed algorithm, and finally, this paper is concluded in Section 6 . 2 Related Work By formulating the floorplanning problem as an analytic optimization model, Zhan et al. [ 15 ] proposed an algorithm framework consisting of the global floorplanning stage and the legalization stage, demonstrating performance superior to the Parquet-4 based on sequence pair and simulated annealing [ 3 ] . Lin and Hung [ 14 ] modelled the global floorplanning as a convex optimization problem, where modules are transformed into circles, and a push-pull mode is proposed with consideration of wirelength. Then, a second-order cone programming is deployed to legalize the floorplan. By incorporating the electrostatic field model proposed in eplace [ 16 ] , a celebrated algorithm for standard cell placement of VLSI, Li et al. [ 12 ] proposed an analytic floorplanning algorithm for large-scale floorplanning cases. Then, the horizontal constraint graph (HCG) and the vertical constraint graph (VCG) were constructed to eliminate overlap