Mobilise your data to drive predictable verification

Conference:	Verification Futures 2024 (click here to see full programme)
Speaker:	Bryan Dickman, Joe Convey
Presentation Title:	Mobilise your data to drive predictable verification
Abstract:	What level of understanding and insight into the effectiveness and efficiency of your verification campaign do you have? We will show you examples of what both good and bad look like and challenge you to think about what level of data and analytics maturity you have today. Verification is a data-heavy problem and therefore verification engineers need to be data analysts, though they may not realise it. Verification is also a compute-heavy process that suffers from multiple completeness dilemmas. Engineering teams need to take many judged decisions around the question of “are we done yet?”. The consequences of getting this wrong can be costly. A leading measure of product quality is the volume of impactful bug escapes and how they impact the end user. This in turn leads to re-work costs, reputational costs and potentially missed sales costs when products are late to market, or teams are consumed with rework and product support when they should be focused on developing the next product. Mobilising data to gain insights and drive more predictable verification plans and outcomes is an imperative and a challenge to all of us; how to do a better job?
Speaker Bio:	Bryan Dickman: several decades of leading engineering teams in the development and verification of complex hardware IP products, 23 years of which was with Arm. I am particularly driven by the benefits of helping companies, their leaders and engineering teams, to understand, analyse and utilise engineering datasets. Currently working as an independent expert and thought leader, and a co-founder of Silicon Insights Limited. Joe Convey: Nearly 20 years in semiconductor and EDA. A Senior Leader helping semiconductor clients mobilise their engineering data to improve IP quality and delivery predictability using analytics-driven engineering techniques. Extensive experience with design enablement, EDA partnership and complex commercial negotiations. Currently working as an independent expert and thought leader, and co-founder of Silicon Insights Limited.
Key Points:	We present a simple method for assessing your teams' data and analytics capability. To do that you need to mobilise your engineering data so we discuss the challenges of collecting good data, and techniques that enable teams to join and correlate different data sources to reveal powerful insights. We will review some different scenarios that are typical of real projects and explain how the data points towards problem areas that can be addressed with more efficient and effective verification depending on which dimensions of quality, cost or schedule you need to drive continuous improvements on. How can better planning of verification campaigns help to make the cost, quality, and schedule more predictable? Do you understand what your current and previous campaigns look like from the dimensions of consumption, verification efficiency and verification effectiveness, and the final quality of the delivered product? Doing a better job will enable you to drive more predictable verification.