Yield prediction for architecture exploration in nanometer technology nodes:: a model and case study for memory organizations

Abstract

Process variability has a detrimental impact on the performance of memories and other system components, which can lead to parametric yield loss at the system level due to timing violations. Conventional yield models do not allow to accurately analyze this, at least not at the system level. In this paper we propose a technique to estimate this system level yield loss for a number of alternative memory organization implementations. This can aid the designer into making educated trade-offs at the architecture level between energy consumption and parametric timing yield by using memories from different available libraries with different energy/performance characteristics considering the impact of manufacturing variations. The accuracy of this technique is very high, an average error of less than 1% is reported, which enables an early exploration of the available options.