Statistical aspects of FinFET based SRAM metrics subject to process and statistical variability

Variability in nanoscale FinFETs brings huge challenges to SRAM design. Using recently developed unified variability compact models, in this paper we investigate the impact of long-range process variation and local statistical variability on various 6T-SRAM metrics including static noise margin, bitline write margin and retention leakage in two types of cell layout. The results show that these metrics are affected by different types of variability with different sensitivities.