针对SRAM高维多失效域的良率评估方法及电子设备
By combining a multi-class logistic regression model and a spatial clustering algorithm, the problem of insufficient efficiency and accuracy in SRAM yield evaluation is solved, and efficient and accurate multi-failure domain evaluation is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING KUANWEN MICROELECTRONICS TECH CO LTD
- Filing Date
- 2026-04-03
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for SRAM yield assessment cannot simultaneously achieve high efficiency and high accuracy, especially when assessing multiple failure domains, where there are problems of low efficiency and large errors.
The failure index parameters of SRAM were processed twice using a multi-class logistic regression model, and data points were clustered in the high-dimensional space using a spatial clustering algorithm. A labeled dataset was trained to generate sample point data and evaluation point data. Finally, the yield evaluation results were determined by the multi-class logistic regression model.
It improves the efficiency and accuracy of SRAM yield assessment, reduces errors caused by boundary coupling and data imbalance, and enables fast and reliable multi-failure domain assessment.
Smart Images
Figure CN121960240B_ABST
Abstract
Citation Information
Patent Citations
Multi-start point importance sampling technology-based SRAM failure probability quick calculation method
CN107729589A
Apparatus and method for estimating an yield of a static random access memory
KR1020130085917A