The AI/ML in EDA SIG explores how best to apply AI for chip design, help identify technical standards opportunities, and recommend high-value prototype projects with the goal of supplementing proprietary EDA efforts by providing standards and sharable capabilities.
One of the questions being addressed is where Si2 can contribute in Generative AI/Large Language Models – the technology that powers ChatGPT. With that in mind, Si2 was a sponsor of the inaugural IEEE Workshop on LLM Aided Design (LAD) held on June 28-29, 2024, in Almaden CA. Si2’ VP of Technology, Dr. Leigh Anne Clevenger, moderated a panel on: “Generative AI for semiconductor from architecture to fab: A pipedream? Or ready for prime time?”, with panelists from Ansys, Drexel University, IBM, Intel, Nvidia, and Siemens EDA.
The SIG also held productive discussions on how to provide a better platform for academia-industry collaboration across the AI/ML for EDA space. The EDA Schema from Drexel University and CircuitOps from Nvidia/Arizona State University were judged to be very good candidates for this purpose. The result was a decision to take the next step and fund the further development of both approaches as part of a new Si2 OpenStandards working group.
