DAC-2022 #DAC59 was refreshing in several ways. First, it brought together people in the EDA and Semiconductor industries after the strict barriers imposed by the pandemic and the remote work model. Second, there was clear interest and focus amongst exhibitors, conference attendees, and panelists to chart key industry initiatives to foster open-source ecosystems and innovation. 

 

Earnest visitors and exhibitors alike made it a point to participate in these sessions to learn more about the burgeoning ecosystem around open-source EDA design tools such as #OpenROAD_EDA and the popular RISC-V processor.

 

Here are some key highlights:

Birds-of-a-Feather Session

This session was organized by OpenROAD project team members and saw enthusiastic participation from industry allies, academia, and users as an informal meeting point for anyone who would like to hear or share ideas or the latest news on:

  • the ecosystem of open-source EDA tools;
  • the use of relevant and principled benchmarking methodologies to clarify and advance the leading edge of EDA research; and
  • goals for the community, along with the current landscape, key updates, and a path forward with key steps and actions needed for the worldwide academic/CAD community.

 

Tom Spyrou, the Chief architect of OpenROAD,  kicked off the session by sharing key achievements and innovations of OpenROAD including:

  • Creation of OpenDB – a truly open source EDA database that supports other key engines with a modular architecture to scale to today’s technology nodes
  • Support for 12nm, 16nm foundry technologies
  • Continuous innovation through a rigorous regression suite integrated through state-of-the-art CI and GitHub actions to track and improve QoR and enhance user engagement and confidence

 

Google engineer Johan Euphrosin (@proppy) presented the latest enhancements to a collaborative approach to integrated chip design and test. Refer to this video link for details on this capability: https://www.zerotoasiccourse.com/post/asic-dev-in-the-cloud/

 

The MacroPlacement GitHub repo of the TILOS AI Institute is heavily using OpenROAD (RTL-MP and ORFS) in its efforts to provide an open, transparent baseline for research on deep reinforcement learning for chip placement

 

All session presentations can be found at the Github link: https://open-source-eda-birds-of-a-feather.github.io/

Cadence announcement to support foundational research

Another announcement from a major EDA provider. Cadence Design Systems announced its support for the sharing of tool scripts by academic users  to promote follow-on research and  peer review. This is a great boost to the research community, and recognizes the importance of a well-supported scientific process that prioritizes transparency and reproducibility to enable the field’s leading edge to advance more rapidly and efficiently.


Other Notable Sessions

DAC Research Panel

What is the Future for Open-Source EDA?

Link: https://vlsicad.ucsd.edu/NEWS22/dac2022_opensource_panel.pdf

Democratization of Chip Design Panel 

OpenROAD is a key enabler of the democratization of chip design as explained in this insightful panel by industry partners: https://www.embedded.com/dac-2022-panel-democratization-of-chip-design/