Search papers, labs, and topics across Lattice.
This paper introduces a multi-stage fixed-outline floorplanning framework that integrates both non-convex and convex optimization techniques to enhance VLSI physical design. The approach begins with a quadratic placement to establish a topology-aware configuration, followed by a non-convex optimization phase using an Adam-based projected gradient method to minimize wirelength and overlaps. The final legalization phase employs a log-transformed model to effectively resolve overlaps, resulting in state-of-the-art wirelength quality improvements of at least 1% to 5% over existing methods on benchmark datasets.
Achieving a 5% reduction in wirelength compared to leading floorplanners could redefine efficiency in VLSI design.
Floorplanning is a critical early stage of VLSI physical design, and its quality directly impacts interconnect wirelength, chip performance, and downstream design efficiency. This article presents a multi-stage fixed-outline floorplanning framework that combines non-convex and convex optimization. The framework operates in three successive phases. In the initial floorplanning phase, a quadratic placement optimizes module connectivity to produce a topology-aware starting configuration with strong net clustering, albeit with significant module overlaps. In the global floorplanning phase, Adam-based projected gradient method is employed to solve non-convex optimization that minimizes wirelength and overlap. In the legalization phase, a log-transformed model exploiting horizontal and vertical constraint graphs eliminates residual overlaps under a convex formulation. Experiments on the MCNC, GSRC and HB+ benchmarks demonstrate that the framework achieves state-of-the-art wirelength quality, reducing average HPWL by at least 1\% and 5\% compared to state-of-the-art floorplanners on tested benchmarks.
