Fast and Accurate High-Sigma Analysis with Worst-Case Points

Audience:
Analog Design, Analog EDA

Duration:
00:53:12

Description:
The webinar focused on IC designers’ challenge to meet stringent specifications for robustness in SRAM bit cells and Flip-Flop registers. Traditional Monte Carlo simulations face practical limitations, especially for non-normally distributed performance. MunEDA, an EDA vendor, introduced high-sigma analysis and worst-case points as a more efficient approach. Key points included the impact of parameter variation and yield on Monte Carlo techniques, the identification of worst-case points, and the introduction of the First Order Reliability Model (FORM). MunEDA’s high-sigma methods, integrated into its EDA tools, allow for scalable analysis (up to 7 sigma) with minimal simulation runs. The webinar highlighted the applicability of these methods to Machine Learning, particularly for large circuits with numerous mismatch parameters. Cases where brute-force Monte Carlo analysis is still necessary were discussed.