Yi-Kan Cheng

Efficient trimmed-sample Monte Carlo methodology and yield-aware design flow for analog circuits

This paper proposes efficient trimmed-sample Monte Carlo (TSMC) methodology and novel yield-aware design flow for analog circuits. This approach focuses on “trimming simulation samples” to speedup MC analysis. The best possible yield and the worst performance are provided “before” MC simulations such that designers can stop MC analysis and start improving circuits earlier. Moreover, this work can combine with variance reduction techniques or low discrepancy sequences to reduce the MC simulation cost further. Using Latin Hypercube Sampling as an example, this approach gives 29× to 54× speedup over traditional MC analysis and the yield estimation errors are all smaller than 1%. For analog system designs, the proposed flow is still efficient for high-level MC analysis, as demonstrated by a PLL system.