A Comparative Study on RTN Deconvolution of Richardson-Lucy and Proposed Partitioned Means for Analyzing SRAM Fail-Bit Prediction Accuracy

Download Now
Provided by: International Association of Computer Science & Information Technology (IACSIT)
Topic: Storage
Format: PDF
In this paper the authors present a comparative analysis of the Random Telegraph Noise (RTN) de-convolution accuracy between the Richardson-Lucy (R-L) algorithm and the proposed Partitioned Forward-problem based De-ConVolution means (PFDCV). Unlike the R-L based de-convolution, the proposed technique successfully solves the issue of noise amplification thanks to eliminating any operations of differential and division. This effectiveness has been demonstrated for the first time with applying it to a real analysis for the effects of the RTN on the overall SRAM margin variations.
Download Now

Find By Topic