Abstract:on-chip interconnects are very important in both integrated circuits and systems, affecting signal transmission directly. To design the interconnects with better performance, designers usually rely on heuristic methods and even rules of thumb with high labor resources. To alleviate this issue, transfer learning-based deep neural networks (TL-DNNs) are proposed to realize the inverse and fully-automatic design of the on-chip interconnects. Such a TL-DNN-based surrogate model consists of two deep neural networks (DNNs). The first DNN is employed to map the performance parameters to the geometric parameters in the equivalent circuit space, where the trained hyperparameters of the first DNN are transferred to the second DNN model. Furthermore, 300 datasets from the in-house electromagnetic (EM) parallel solver space are generated and used to train the second DNN, with 3000 datasets used for testing. Compared with the DNN-based and deep hybrid neural network (HDNN)-based inverse design methods, the proposed method not only achieves a high testing accuracy (93.67%), but also requires only 6.15 h of modeling time. In addition, the well-trained model has the ability to achieve the desired design specification within a fraction of a second, compared to the hours-days-weeks of the forward design approach. An elaborate interconnect prototype is fabricated, and its measured results fully satisfy the performance demands. It is expected that the proposed modeling scheme can be used for the design of many interconnects.

Fast impedance prediction for power distribution network using deep learning

Fast PDN Impedance Prediction Using Deep Learning

Worst-case Power Integrity Prediction Using Convolutional Neural Network.

Worst-Case Dynamic Power Distribution Network Noise Prediction Using Convolutional Neural Network

Efficient DC and AC Impedance Calculation for Arbitrary-Shape and Multilayer PDN Using Boundary Integration

Inverse Design of On-Chip Interconnect via Transfer Learning-Based Deep Neural Networks

EMI Radiation Prediction and Structure Optimization of Packages by Deep Learning

Fast Decoupling Capacitor Optimization for Power Delivery Network Based on Model and Data Fusion Method

Plane-pair PEEC method for impedance analysis of multi-slots power distribution network with experimental validation

A 3D Convolutional Network for Fast and Accurate S/Z-Parameters Prediction of Si-Interposer Power Distribution Network With Through Silicon Vias

Plane-Pair PEEC Modeling for Package Power Layer Nets with Inductance Extraction

Impedance Calculation of Grid Power Distribution Network by Using Artificial Neural Network

Equivalent Inductance Analysis and Quantification for PCB PDN Design

Emi Prediction Of Packages By Deep Neural Network

A hybrid domain decomposition and optimization method for predicting electromagnetic emissions from printed circuit boards

Decoupling Optimization for Complex PDN Structures Using Deep Reinforcement Learning

The Analysis and Experimental Investigation of Electromagnetic Characteristics on High Speed Circuit PDN with Multislots

OpeNPDN: A Neural-network-based Framework for Power Delivery Network Synthesis

A Novel Machine Learning-Based Short-Circuit Current Prediction Method for Active Distribution Networks

An Integral Equation Hybrid Method for the Impedance Calculation of the Grid Power Distribution Network with an Arbitrary Shape

An Effective and Efficient Approach for Radiated Emission Prediction Based on Amplitude-Only Near-Field Measurements