Article: Formal Engines Learn From Experience - FirstEDA
portfolio_page-template-default,single,single-portfolio_page,postid-23004,ajax_fade,page_not_loaded,,qode-child-theme-ver-1.0.0,qode-theme-ver-14.5,qode-theme-bridge,wpb-js-composer js-comp-ver-6.2.0,vc_responsive

Article: Formal Engines Learn From Experience

By Chris Edwards, 8th May 2019



Machine learning is coming to formal verification as vendors explore ways to make their products easier to use for non-formal specialists.


At CDNLive EMEA this week, Cadence Design Systems launched what it called its third generation of the JasperGold formal-verification suite. The product adds several mechanisms based on various forms of artificial intelligence (AI) for automating the job of picking and scheduling the engines that are deployed based on how their RTL is structured.


A few weeks earlier at a seminar organised by TV&S on machine learning in EDA in April, formal specialist OneSpin Solutions described its own work on using forms of supervised learning to improve the ability of tools to detect which engines need to be deployed on RTL and when. As an early user of the new version, STMicroelectronics said it saw an average doubling in throughput with the solver-inference feature, which determines which engines fit the design and task best.