A PPA-Driven 3D-IC Partitioning Selection Framework with Surrogate Models

2026-04-20

AI Summary

This work addresses the limitation of conventional 3D-IC netlist partitioning approaches that rely on proxy objectives and struggle to effectively enhance actual PPA (Power, Performance, Area) metrics. To bridge the gap between proxy-based optimization and real-world PPA outcomes, the authors propose DOPP, a novel framework that uniquely integrates D-optimal experimental design with PPA-driven optimization. DOPP directly optimizes real PPA targets by leveraging surrogate modeling, D-optimal sampling, and parallel evaluation to efficiently identify near-optimal partitions among a small set of candidates. Evaluated on eight 3D-IC designs, DOPP achieves significant improvements—reducing congestion by 9.99%, shortening total wirelength by 7.87%, improving worst negative slack (WNS) by 7.75%, enhancing total negative slack (TNS) by 21.85%, and lowering power consumption by 1.18%—all while evaluating only a limited number of candidate partitions.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Machine Learning: Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

3D-IC netlist partitioning is commonly optimized using proxy objectives, while final PPA is treated as a costly evaluation rather than an optimization signal. This proxy-driven paradigm makes it difficult to reliably translate additional PPA evaluations into better PPA outcomes. To bridge this gap, we present DOPP (D-Optimal PPA-driven partitioning selection), an approach that bridges the gap between proxies and true PPA metrics. Across eight 3D-IC designs, our framework improves PPA over Open3DBench (average relative improvements of 9.99% congestion, 7.87% routed wirelength, 7.75% WNS, 21.85% TNS, and 1.18% power). Compared with exhaustive evaluation over the full candidate set, DOPP achieves comparable best-found PPA while evaluating only a small fraction of candidates, substantially reducing evaluation cost. By parallelizing evaluations, our method delivers these gains while maintaining wall-clock runtime comparable to traditional baselines.

Problem

• 3D-IC partitioning
• PPA optimization
• proxy objectives
• design evaluation
• surrogate models

Innovation

• PPA-driven optimization
• 3D-IC partitioning
• surrogate modeling
• D-optimal design
• design space reduction