Multi-Stage Cell Timing Modeling Using Equivalent Input Characterization Waveforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current behavioral timing models for multi-stage cells in integrated circuits are inaccurate due to their inability to account for analog effects such as crosstalk and RC long tail, especially in hierarchical designs, where they require higher margins to compensate for unmodeled effects.
Innovation Solution
The approach involves determining an equivalent input characterization waveform (EICW) that matches the output slew of a structural model, allowing it to be used with behavioral timing models to estimate the timing response of multi-stage cells, thereby improving accuracy without additional computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If structural models (CCSN) are used to model analog effects like crosstalk and RC long tail, then timing accuracy is improved, but computational intensity increases making them suitable only for gates with fewer transistors
Solution Approach 1:
The cell is divided into multiple stages, with the first stage modeled using structural models to capture analog effects accurately, while subsequent stages use behavioral models. This segmentation allows the computationally intensive structural modeling to be applied only where necessary (at the input stage where crosstalk and RC effects are most significant), while behavioral models handle the remaining stages efficiently.
Solution Approach 2:
Different modeling approaches are applied to different parts of the cell based on their specific requirements. The first stage, which is most susceptible to analog effects like crosstalk and RC long tail, receives the more accurate structural modeling treatment, while later stages use the lighter behavioral models, optimizing the balance between accuracy and computational cost locally.
2Productivity
If behavioral timing models (NLDM, CCST) are used for multi-stage cells, then computational efficiency is improved, but timing accuracy deteriorates due to inability to account for analog effects
Solution Approach 1:
The multi-stage cell is segmented such that only the critical first stage undergoes structural modeling to capture analog effects, while the remaining stages are handled by behavioral models. This allows the system to maintain computational efficiency while improving timing accuracy where it matters most.
Solution Approach 2:
The patent introduces an intermediate approach where structural modeling of the first stage produces equivalent input characterization waveforms (EICWs) that serve as mediators. These EICWs capture the analog effects and are then used as inputs to the behavioral models for subsequent stages, bridging the gap between the accuracy of structural models and the efficiency of behavioral models.
3Reliability
If behavioral models use higher margins to account for unmodeled analog effects, then timing reliability is improved, but the models become less precise for actual timing prediction
Solution Approach 1:
The patent replaces the mechanical approach of adding conservative margins to behavioral models with a more sophisticated structural modeling approach. By using structural models to explicitly capture analog effects like crosstalk and RC long tail, the system achieves timing reliability through accurate physical modeling rather than through conservative margin additions, thereby maintaining timing prediction precision.
Data Source
AI summary
An equivalent input characterization waveform (EICW) is determined for a channel-connected block (CCB) located on a boundary of a cell, for a specific waveform of interest. The EICW and the specific waveform of interest produce a same timing characteristic of the CCB, but the EICW belongs to a set of waveforms on which a behavioral timing model for the multi-stage cell is based whereas the specific waveform of interest is not so limited. A timing response of the multi-stage cell is then estimated based on applying the EICW.