DESIGN AND DEVELOPMENT OF A NOVEL QUALITY ASSURANCE REGRESSION INFRASTRUCTURE

The VLSI design complexity has increased exponentially over the last few years and the transition to 7/5 nm and beyond is a big jump. There is 5 times increase in grid complexity from 16 nm to 7nm designs and the grids have more than 10 billion nodes. Design robustness, reliability, power integrity issues are a primary concern to be solved which has led to the development of a software on big data platforms which can perform multi-dimensional analysis on the design. For these and other software products, there are several release cycles in a year to meet customer requirements. With each release, build quality needs to be verified. Therefore, a modular and a robust tool to check release quality is required. The infrastructure developed in this project will regress several testcases and data mine several parameters from each run and compare it against a golden reference. The results from comparison will then be tabulated and reviewed to ensure there are no issues in product build. This paper focuses on building a quality assurance tool, to be used for the comparison of various parameters and which is adaptable in nature i.e. it can support quick and easy addition of new parameters in the future as per the requirement of any product.