High-Sigma Verification
Verifying designs with high yield requirements (e.g. >3 sigma) like memory designs is difficult. Monte Carlo analysis is inefficient, as far too many samples are required to reliably predict high-sigma values, and other scaling techniques make unrealistic assumptions about the nature of distributions.
Solido has developed a proprietary algorithm that makes high-sigma verification practical. It adaptively focuses simulations about the specifications to efficiently and accurately measure yields for high-sigma designs. This technique is 10-1000x faster than Monte Carlo analysis for measuring the robustness of high-sigma designs, but has similar accuracy.
This technology is available in the Verify High-Sigma Design statistical application.
Overall benefits:
- Quickly and accurately measure the yield of high-sigma designs
- Yield prediction, with statistical confidence intervals
