|
Chiplet and disaggregated designs are increasingly prevalent in commercial offerings from edge to server. However, the associated design complexity poses a significant challenge for today’s design tools, flows and methodologies, especially if we aim to achieve highly optimized design at scale.
Augmented Intelligence, combining human and machine intelligence, offers a transformative solution to support multi-level and multi-domain design optimization. By strategically assigning high-level decision-making to humans and iterative computations to Artificial Intelligence, this methodology can scale the number of custom-optimized designs with the same number of resources, achieving high-quality competitive products with reduced time-to-market.
Through collaborations at Intel and with partners, we have developed and deployed Augmented Intelligence solutions spanning silicon to system design and hardware to software design, achieving design efficiency gains exceeding 90% in critical areas. This talk will present practical examples and key learnings from several years of developing Augmented Intelligence solutions for end-to-end design.
| 
