Accelerating large-scale physical verification: Best practices with Calibre on Google Cloud

As advanced-node SoCs continue to grow in size and complexity, physical verification (PV) closure is increasingly becoming a schedule and compute bottleneck. In addition to the quality of the databases, we now also must look over the quality of the run itself. This presentation describes how a combination of practical methodology for closing large designs using Calibre and Google Cloud infrastructure was used to improve productivity and reduce turnaround time during signoff for a design in the leading-edge technology node near full reticle die size limit.

Key focus areas include hierarchical verification planning, efficient job parallelization, optimized debug workflows, and automation to streamline repeated runs. Cloud-based scaling allowed optimal access to the next-generation hardware and dynamic compute capacity during peak demand, significantly reducing queue delays and enabling faster design iteration.