Research
FlowPlace: Flow Matching for Chip Placement
arXiv:2604.23658v1 Announce Type: cross Abstract: Chip placement plays an important role in physical design. While generative models like diffusion models offer promising learning-based solutions, cur
arXiv:2604.23658v1 Announce Type: cross Abstract: Chip placement plays an important role in physical design. While generative models like diffusion models offer promising learning-based solutions, current methods have the following limitations: they use random synthetic data for pre-training, require long sampling times, and often result in overlaps due to their dependence on gradient-based solvers during the sampling process. To overcome these issues, we propose FlowPlace, which features mask-guided synthetic data generation, flow-based efficient training with flexible prior injection, and hard constraint sampling for overlap-free layouts. Experiments on OpenROAD and ICCAD 2015 benchmarks show FlowPlace achieves better PPA metrics, 10-50imes faster sampling efficiency, and zero overlaps.
Related
- From Navigation to Refinement: Revealing the Two-Stage Nature of Flow-based Diffusion Models through Oracle Velocity
- Frequency-Forcing: From Scaling-as-Time to Soft Frequency Guidance
- Real-Time Streamable Generative Speech Restoration with Flow Matching
- Attention-Guided Flow-Matching for Sparse 3D Geological Generation
Source: arXiv cs.AI | 2026-04-28