Memory System Power Analysis Using Encrypted Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current memory systems face challenges in analyzing overall power characteristics, as noise from power distribution networks affects semiconductor devices, and existing methods only analyze power characteristics of power distribution networks, limiting the understanding of their impact on the entire memory system.
Innovation Solution
A method is introduced to design and fabricate memory systems by analyzing power characteristics of both the power supply circuit and the memory system using encrypted models, including a power characteristic model of the power supply circuit, board power distribution network, and semiconductor device, to accurately predict and improve power characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If only power characteristics of power distribution networks are analyzed, then the analysis process is simple, but the understanding of impact on the entire memory system is limited
Solution Approach 1:
The analysis system is segmented into multiple independent modules: power distribution network analysis module, power supply circuit analysis module, and semiconductor device analysis module. Each module analyzes specific components separately using dedicated power characteristic models, then results are integrated to achieve comprehensive system-level analysis without requiring complex monolithic analysis processes
Solution Approach 2:
Power characteristic models serve as intermediaries between different analysis modules. These models (including encrypted models for intellectual property protection) enable standardized data exchange and integration between the power distribution network, power supply circuit, and semiconductor device analyses, facilitating comprehensive system analysis while maintaining modular simplicity
2Reliability
If encrypted models are used for power supply circuit analysis, then intellectual property is protected, but model integration and analysis become more complex
Solution Approach 1:
Instead of directly integrating proprietary encrypted models, the system creates functional copies or surrogate models that replicate the essential power characteristic behaviors. These surrogate models maintain the protective encryption while enabling seamless integration with other analysis modules, thus protecting intellectual property without significantly increasing integration complexity
Solution Approach 2:
The system transforms encrypted power supply circuit models into standardized parameter formats that can be easily integrated with other modules. By changing the representation parameters while maintaining functional equivalence and encryption, the system enables smooth model integration without exposing proprietary information or increasing complexity
Data Source
AI summary
A method of designing a memory system, which includes a semiconductor device and a power supply circuit supplying power to the semiconductor device via a board power distribution network, includes analyzing power characteristics of respective components of the power supply circuit by using a power characteristic model of the power supply circuit, and analyzing power characteristics of the memory system. The power characteristic model of the power supply circuit includes an encrypted model.