ArtNet: Hierarchical Clustering-Based Artificial Netlist Generator for ML and DTCO Application

2025-10-15

AI Summary

To address the scarcity of training data and prolonged design cycles that hinder machine learning–based design technology co-optimization (DTCO) in advanced process nodes, this work proposes a hierarchical clustering–driven synthetic netlist generation method. By faithfully reconstructing critical topological features, the approach significantly enhances the realism and diversity of generated netlists, enabling broader exploration of the design space. Coupled with a target-parameter matching mechanism, the synthesized data supports both CNN-based dynamic voltage regulation (DRV) timing prediction and mini-brain circuit synthesis. Experimental results demonstrate that data augmentation improves DRV prediction F1-score by 0.16; moreover, the generated mini-brain designs achieve 97.94% PPA (power, performance, area) matching accuracy against reference implementations. This work effectively alleviates the data bottleneck in DTCO, thereby improving model generalizability and accelerating the overall design flow.

Technology Category

• Machine Learning: Hardware-aware ML
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Web Mining and Content Analysis: Web data generation and simulation
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets

Abstract

In advanced nodes, optimization of power, performance and area (PPA) has become highly complex and challenging. Machine learning (ML) and design-technology co-optimization (DTCO) provide promising mitigations, but face limitations due to a lack of diverse training data as well as long design flow turnaround times (TAT). We propose ArtNet, a novel artificial netlist generator designed to tackle these issues. Unlike previous methods, ArtNet replicates key topological characteristics, enhancing ML model generalization and supporting broader design space exploration for DTCO. By producing realistic artificial datasets that moreclosely match given target parameters, ArtNet enables more efficient PPAoptimization and exploration of flows and design enablements. In the context of CNN-based DRV prediction, ArtNet's data augmentationimproves F1 score by 0.16 compared to using only the original (real) dataset. In the DTCO context, ArtNet-generated mini-brains achieve a PPA match up to 97.94%, demonstrating close alignment with design metrics of targeted full-scale block designs.

Problem

• Addresses lack of diverse training data for machine learning applications
• Reduces long design flow turnaround times in technology optimization
• Enhances power, performance and area optimization through artificial netlists

Innovation

• Generates artificial netlists replicating topological characteristics
• Enhances ML generalization and DTCO design space exploration
• Produces realistic datasets matching target parameters for PPA optimization