I recently read a blog written by Dr. Noa Kageyama, performance psychologist and Juilliard alumnus and faculty member, titled “The Perils of Aiming Low: How Our Expectations Can Shape Our Students’ Learning & Performance.”
Based on research findings from schools and sports, Dr. Kageyama concluded that high expectations from teachers and coaches correlate positively with an individual’s learning and growth, helping improve confidence and making the most of one’s ability.
The blog resonates with me because I am a parent, always seeking ways to help my daughters reach their maximum potential. But It also reminds me of a common practice I see in the industry regarding formal verification adoption.
Often, due to a lack of understanding of formal verification’s capabilities, companies bring in formal for easy tasks that have a high degree of automation. However, this low-hanging fruit rarely offers enough rewards for the engineers to feel deep satisfaction in their work or management to see the value of formal verification. All too often, companies hoping to fully adopt formal verification quickly give up because of the low return on investment (ROI).
The problem is not in the formal technology, but in the incorrect expectations set by management, regarding goals for formal. The illustration below shows formal use case models. Automatic formal, formal apps and formal assertion verification are all useful. However, the goal of these simpler use models is bug hunting.
End-to-End formal is the only one that enables formal sign-off. Formal sign-off is achieved when formal verification completely replaces simulation at the block level, and this claim is based on some of the same reliable metrics used in simulation, such as coverage.
End-to-End formal verification enables verification engineers to use it for final sign-off.
Without the goal of formal sign-off, formal engineers do not have a sense of ownership and accountability for the quality of the Design Under Test (DUT). Without having complete responsibility for the quality of the DUT, formal engineers may not feel compelled to learn the techniques needed to tackle and overcome formal’s complexity.
Without a grasp of these advanced formal techniques, they will not grow in their careers to become experts. A career without learning and growth potential is a career few would want to pursue and take on long term. Conversely, if management sets high expectations to achieve formal sign-off, not only will formal engineers be challenged to learn advanced formal techniques and strive to become formal experts, companies will see greater ROI from the formal efforts. This is a win-win for all.
Most companies today do not know formal verification can be used for sign-off. Worse yet, far too often, we see formal licenses not being put to use. Formal expertise is way too scarce in the industry. This can be changed by simply setting expectations to aim for formal sign-off. Any other way of setting expectations is of no benefit.