Data Variability Study of Advanced 3D NAND Memory using Python

The methodology for data analysis with large variability based on Python is discussed and demonstrated. As NAND memory transitioned from 2D to 3D, a significant new source of variability emerged due to the stacking of several hundred layers of memory cells on top of each other. This 3D construction demands a novel approach to efficient analysis that allows for a more general overview without sacrificing the details of individual memory cells. The authors explore how to achieve this by identifying sources of variation and systematically parsing and manipulating the data file into smaller files. This methodology enables straightforward visualization of parametric variability in relation to test and process conditions and facilitating the creation of accessible documentation for future reference. The use of Python provides an open-source and user-friendly interface that can be further customized to enhance consistency and reusability.