Non-parametric Statistical Static Timing Analysis based on Improved Parallel Monte Carlo

Although statistical methods have been widely used in static timing analysis of circuits, there are still challenges to cope with including high process variations in circuits. Due to high process variations, random variables like delay may not follow any known distributions. Therefore, the assumption of any specific distribution will yield great errors. Non-parametric methods can deal with this challenge as they do not rely on any distribution assumption. One of the most reliable types of non-parametric solutions are Monte Carlo-based methods. However, Monte Carlo-based solutions suffer from high run-time complexity. In this paper, a fast approach is proposed to improve Monte Carlo execution time for timing analysis of the circuits. Evaluations show that this method results in a faster solution in comparison to previous work.