2D-ThermAl: Physics-Informed Framework for Thermal Analysis of Circuits using Generative AI

2025-11-30

AI Summary

To address the high computational cost of traditional thermal simulation methods (e.g., finite element method) in early-stage IC design—which impedes rapid design iteration—this paper proposes a physics-informed generative AI framework for end-to-end prediction of full-chip transient and steady-state temperature fields from input activity profiles. We introduce a hybrid U-Net architecture incorporating positional encoding and a Boltzmann-based physical regularization term to ensure accuracy and thermodynamic consistency across a wide temperature range (25–95°C). The model is trained on a multi-scale heat-source dataset generated via COMSOL, under physics-constrained supervision. Evaluated on large-scale circuits, it achieves a root-mean-square error of 0.71°C and a full-scale error below 2.2%, while accelerating inference by 200× over FEM. The method enables efficient hot-spot localization and interpretable thermal pattern learning.

Technology Category

• Machine Learning: Hardware-aware ML
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Thermal analysis is increasingly critical in modern integrated circuits, where non-uniform power dissipation and high transistor densities can cause rapid temperature spikes and reliability concerns. Traditional methods, such as FEM-based simulations offer high accuracy but computationally prohibitive for early-stage design, often requiring multiple iterative redesign cycles to resolve late-stage thermal failures. To address these challenges, we propose 'ThermAl', a physics-informed generative AI framework which effectively identifies heat sources and estimates full-chip transient and steady-state thermal distributions directly from input activity profiles. ThermAl employs a hybrid U-Net architecture enhanced with positional encoding and a Boltzmann regularizer to maintain physical fidelity. Our model is trained on an extensive dataset of heat dissipation maps, ranging from simple logic gates (e.g., inverters, NAND, XOR) to complex designs, generated via COMSOL. Experimental results demonstrate that ThermAl delivers precise temperature mappings for large circuits, with a root mean squared error (RMSE) of only 0.71°C, and outperforms conventional FEM tools by running up to ~200 times faster. We analyze performance across diverse layouts and workloads, and discuss its applicability to large-scale EDA workflows. While thermal reliability assessments often extend beyond 85°C for post-layout signoff, our focus here is on early-stage hotspot detection and thermal pattern learning. To ensure generalization beyond the nominal operating range 25-55°C, we additionally performed cross-validation on an extended dataset spanning 25-95°C maintaining a high accuracy (<2.2% full-scale RMSE) even under elevated temperature conditions representative of peak power and stress scenarios.

Problem

• Addresses thermal analysis challenges in modern integrated circuits.
• Overcomes computational limitations of traditional FEM-based simulations.
• Provides early-stage hotspot detection and thermal pattern learning.

Innovation

• Physics-informed generative AI framework for thermal analysis
• Hybrid U-Net with positional encoding and Boltzmann regularizer
• Rapid full-chip temperature estimation from activity profiles