An official website of the United States government
As transistor densities increase, managing thermal challenges in 3D IC designs becomes more complex. Traditional methods like finite element methods and compact thermal models (CTMs) are computationally expensive, while existing machine learning (ML) models require large datasets and a long training time. To address these challenges with the ML models, we introduce a novel ML framework that integrates with CTMs to accelerate steady-state thermal simulations without needing large datasets. Our approach achieves up to 70× speedup over state-of-the-art simulators, enabling real-time, high-resolution thermal simulations for 2D and 3D IC designs.
more »
« less
This content will become publicly available on May 27, 2026
Steady-State Temperature Prediction Based on Compact Thermal Models Using Machine Learning
With the scaling up of transistor densities, the thermal management of integrated circuits (IC) in 3D designs is becoming challenging. Conventional simulation methods, such as finite element methods, are accurate but computationally expensive. Compact thermal models (CTMs) provide an effective alternative and produce accurate thermal simulations using numerical solvers. Recent work has also designed machine learning (ML) models for predicting thermal maps. However, most of these ML models are limited by the need for a large dataset to train and a long training time for large chip designs. To overcome these challenges, we present a novel ML framework that integrates with CTMs to accelerate thermal simulations without the need for large datasets. We introduce a methodology that effectively combines the accuracy of CTMs with the efficiency of ML using a physically informed linear regression model based on the thermal conduction equation. We further introduce a window-based model reduction technique for scalability across a range of grid sizes and system architectures by reducing computational overhead without sacrificing accuracy. Unlike most of the existing ML methods for temperature prediction, our model adapts to changes in floorplans and architectures with minimum retraining. Experimental results show that our method achieves up to 70x speedup over the state-of-the-art thermal simulators and enables real-time, high-resolution thermal simulations on different IC designs from 2D to 3D.
more »
« less
- Award ID(s):
- 2131127
- PAR ID:
- 10659114
- Publisher / Repository:
- 24th IEEE Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITherm) 2025
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
With growing transistor densities, analyzing temperature in 2D and 3D integrated circuits (ICs) is becoming more complicated and critical. Finite-element solvers give accurate results, but a single transient run can take hours or even days. Compact thermal models (CTMs) shorten the temperature simulation running time using a numerical solver based on the duality between thermal and electric properties. However, CTM solvers often still take hours for small-scale chips because of iterative numerical solvers. Recent work using machine learning (ML) models creates a fast and reliable framework for predicting temperature. However, current ML models demand large input samples and hours of GPU training to reach acceptable accuracy. To overcome the challenges stated, we design an ML framework that couples with CTMs to accelerate steady-state and transient thermal analysis without large data inputs. Our framework combines principal-component analysis (PCA) with closed-form linear regression to predict the on-chip temperature directly. The linear regression weights are solved analytically, so training for a grid size of 512 × 512 finishes in under a minute with only 15–20 CTM samples. Experimental results show that our framework can achieve more than 33x and 49.6x speedup for steady-state and transient simulation of a chip with a 245.95mm^2 footprint, keeping the mean squared error below 0.1 deg C^2 .more » « less
-
Analog integrated circuit (IC) placement is a heavily manual and time-consuming task that has a significant impact on chip quality. Several recent studies apply machine learning (ML) techniques to directly predict the impact of placement on circuit performance or even guide the placement process. However, the significant diversity in analog design topologies can lead to different impacts on performance metrics (e.g., common-mode rejection ratio (CMRR) or offset voltage). Thus, it is unlikely that the same ML model structure will achieve the best performance for all designs and metrics. In addition, customizing ML models for different designs require more tremendous engineering efforts and longer development cycles. In this work, we leverage Neural Architecture Search (NAS) to automatically develop customized neural architectures for different analog circuit designs and metrics. Our proposed NAS methodology supports an unconstrained DAG-based search space containing a wide range of ML operations and topological connections. Our search strategy can efficiently explore this flexible search space and provide every design with the best-customized model to boost the model performance. We make unprejudiced comparisons with the claimed performance of the previous representative work on exactly the same dataset. After fully automated development within only 0.5 days, generated models give 3.61% superior accuracy than the prior art.more » « less
-
Understanding thermal stress evolution in metal additive manufacturing (AM) is crucial for producing high-quality components. Recent advancements in machine learning (ML) have shown great potential for modeling complex multiphysics problems in metal AM. While physics-based simulations face the challenge of high computational costs, conventional data-driven ML models require large, labeled training datasets to achieve accurate predictions. Unfortunately, generating large datasets for ML model training through time-consuming experiments or high-fidelity simulations is highly expensive in metal AM. To address these challenges, this study introduces a physics-informed neural network (PINN) framework that incorporates governing physical laws into deep neural networks (NNs) to predict temperature and thermal stress evolution during the laser metal deposition (LMD) process. The study also discusses enhanced accuracy and efficiency of the PINN model when supplemented with small simulation data. Furthermore, it highlights the PINN transferability, enabling fast predictions with a set of new process parameters using a pre-trained PINN model as an online soft sensor, significantly reducing computation time compared to physics-based numerical models while maintaining accuracy.more » « less
-
Abstract Machine learning (ML) models are used for in-situ monitoring in additive manufacturing (AM) for defect detection. However, sensitive information stored in ML models, such as part designs, is at risk of data leakage due to unauthorized access. To address this, differential privacy (DP) introduces noise into ML, outperforming cryptography, which is slow, and data anonymization, which does not guarantee privacy. While DP enhances privacy, it reduces the precision of defect detection. This paper proposes combining DP with Hyperdimensional Computing (HDC), a brain-inspired model that memorizes training sample information in a large hyperspace, to optimize real-time monitoring in AM while protecting privacy. Adding DP noise to the HDC model protects sensitive information without compromising defect detection accuracy. Our studies demonstrate the effectiveness of this approach in monitoring anomalies, such as overhangs, using high-speed melt pool data analysis. With a privacy budget set at 1, our model achieved an F-score of 94.30%, surpassing traditional models like ResNet50, DenseNet201, EfficientNet B2, and AlexNet, which have performance up to 66%. Thus, the intersection of DP and HDC promises accurate defect detection and protection of sensitive information in AM. The proposed method can also be extended to other AM processes, such as fused filament fabrication.more » « less
