An Efficient SRAM yield Analysis Using Scaled-Sigma Adaptive Importance Sampling

Abstract

Statistical SRAM yield analysis has become a growing concern for the requirement of high integration density and reliability of SRAM under process variations. It is a challenge to estimate the SRAM failure probability efficiently and accurately because the circuit failure is a "rare-event". Existing methods are still not efficient enough to solve the problem, especially in high dimensions. In this paper, we develop a scaled-sigma adaptive importance sampling (SSAIS) which is an extension of the adaptive importance sampling. This method changes not only the location parameters but the shape parameters by searching the failure region iteratively. The 40nm SRAM cell experiment validated that our method outperforms Monte Carlo method by 1500x and is 2.3x~5.2x faster than the state-of-art methods with reasonable accuracy. Another experiment on sense amplifier shows our method achieves 1811x speedup over the Monte Carlo method and 2x~11x speedup over the other methods.