Applying Machine Learning Methods to Signal Integrity Analysis
- 11 December 2024
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Abstract
Voltage spikes caused by signal switching in adjacent circuits can lead to false switching and failures in integrated circuits. To identify such risks, signal integrity (SI) analysis is performed by extracting parasitic RC networks, followed by SPICE simulation or application of analytical methods. However, this process is time-consuming. This research proposes using machine learning methods to rapidly and accurately predict the results of SI analysis without the need for computationally expensive modeling. It accelerates SI analysis a lot with low accuracy losses. Also, this paper introduces new reduction model for victim-aggressor pair that is more accurate than previously used.Keywords
This publication has 9 references indexed in Scilit:
- Fast Analysis of Static IR Drop Effect Based on Machine Learning MethodsProceedings of the Institute for System Programming of the RAS, 2023
- Application of Neural Networks for Image Segmentation in the Prob-lem of Fast Global RoutingProceedings of the Institute for System Programming of the RAS, 2023
- Prospect of Analyzing Integrated Circuits Based on Dataset with Synthesis Results2022
- DRiLLS: Deep Reinforcement Learning for Logic Synthesis2020
- XGBoost2016
- Profile-driven Program Synthesis For Evaluation Of System Power Dissipation2005
- Accurate crosstalk noise modeling for early signal integrity analysisIEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2003
- System-level power optimization: techniques and tools2003
- Evaluation of architecture-level power estimation for CMOS RISC processors1995