5
Statistical Corner Creation
After running Monte Carlo simulations, you can analyze the yield and identify the specifications for which the results need improvement. You can then create statistical corners to use them in further analysis and design optimization. Statistical corners contain the simulation settings needed to recreate the statistical variable values for a specific condition. Statistical corners apply to a particular measurement or specification.
Virtuoso Variation Option provides the following advanced methods to create statistical corners:
-
K-Sigma Corners
A corner created by the K-Sigma Corners method is based on modeling and extrapolation from the Probability Density Function (PDF) of the output. This corner is not one of the samples generated by the simulator and this method is typically faster than the Worst Samples method.
The K-Sigma Corners method is strongly recommended over the Worst Samples method when the number of statistical parameters is large (> 1000). -
Worst Samples
A corner created by the Worst Samples method is from a sample generated by the simulator. It represents the condition that allows the yield verification sign-off to pass. If the design meets specifications at the corner from the Worst Samples method, yield verification is likely to pass with the same target yield and probability. The total number of samples considered by the Worst Samples method represent the sign-off condition for the given target yield and probability.
The Worst Samples method is recommended when high accuracy is needed and the number of statistical parameters is not large (< 1000).
Related Topics
Return to top