一种快速检查底层模块电源地在上层模块连接强弱的方法
By automatically extracting and fusing multi-dimensional feature parameters through convolutional neural networks, the accuracy problem of checking the strength of power and ground connections in the underlying modules is solved, and intelligent evaluation of the power and ground connection strength of the chip is realized, thereby improving the quality and reliability of chip design.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHONGYIN MICROELECTRONICS (CHENGDU) CO LTD
- Filing Date
- 2025-11-21
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, checking the connection strength between the power ground of the lower-level module and the upper-level module relies on manual judgment, which is inefficient and highly subjective. It cannot accurately quantify the connection strength, making it difficult to identify potential risk points in chip design and affecting chip performance and reliability.
A convolutional neural network is used to construct a modular feature recognizer, which automatically extracts multi-dimensional feature parameters such as the number and density of connecting holes, the total area of the lines, and the narrowest line width. The connection strength parameters are calculated by weighted fusion to achieve automated evaluation of the connection strength level.
It significantly improves the accuracy and efficiency of chip power and ground connection strength checks, overcomes the subjectivity and inaccuracy of manual checks, and enhances chip design quality and reliability.
Smart Images
Figure CN121435906B_ABST