Recommendations and roadmap for the development of open-source silicon in the EU

Date: November 3 2023

Here we publish a list of recommendations for the European Commission and a roadmap for the development of open-source silicon in the EU.

The full document can be downloaded here.

Below we have copied only the executive summary.

Executive summary

After a brief introduction which defines the necessary terminology and introduces the political background, in chapter 4 we argue that open-source Electronic Design Automation (EDA) tools and open-source silicon are essential instruments to achieve many of the goals set by the Chips Act. This chapter does not provide any recommendations yet.

In chapter 5 we analyse the “Design Platform” foreseen by the Chips Act in the light of the feedback obtained by interrogating multiple European SMEs involved in chip design. Potential problems were identified with the foreseen cloud-based infrastructures. These are related with security, privacy, the too large spectrum of tools, forced upgrades, increased control by EDA vendors, and increased risk of discovery of patent infringement. To mitigate these problems we recommend to support, besides cloud installations, also local EDA installations, and we recommend to support open-source EDA flows besides the commercial flows.

In chapter 6 we analyse the role of standards and standards-setting bodies in the context of open-source. In particular we highlight how open-source development has needs which are substantially different from the mainstream industrial approach to standardization. We highlight in particular a set of necessary conditions that, in our experience, standards must fulfil in order to be adopted by the open-source community.

In chapter 7 we discuss academia. We argue that academia can and should play a significant role in the development of open-source EDA tools and open-source silicon. For fostering open-source development in universities we recommend that the metrics to evaluate academics should include open-source projects aside to publications, citations, etc. Next, we highlight how there are two classes of academics, which are both essential: developers of EDA tools and users of EDA tools. Given the near complete disappearance of the former, we recommend that a new generation of professors is hired to develop open-source EDA tools and to revive the corresponding knowledge in Europe. In this chapter we finally highlight how people who have not been exposed to open-source solutions often don’t appreciate its potential therefore creating a cultural bias. In conclusion, also because of other conflicts of interest, we recommend to introduce new and independent personnel in academia.

In chapter 8 we present an open letter about ecological sustainability. The signatories of this letter recommend: 1. more sober technology, 2. the “6Rs” (refurbish, reuse, repair, reliability, reduce, recycle) for electronic devices, 3. external and independent auditors for Life Cycle Assessments (LCAs), 4. encouraging world-wide regulations to limit the environmental impact in the ICT sector.

In chapter 9 we discuss patent threats and possible upcoming problems for open-source development. Unfortunately we have no consolidated recommendations yet.

In chapter 10 we briefly discuss possible implications of Artificial Intelligence on chip design. We warn that the advent of AI might produce an increased silicon-technology gap between owners of AI and the others. We recommend to put in place mechanisms to prevent a further power unbalance between large and small actors. A possible mechanism consists of guaranteeing a fully open (i.e. down to silicon) development of AI. In chapter 11 we discuss the Cyber Resilience Act (CRA) and we recommend that: 1. the concept of open-silicon is added to the CRA, and 2. open-silicon is recognized as a key ingredient to achieve some of the hardware cybersecurity goals.

In chapter 12 we finally present a roadmap for open-source silicon development. First we make a list of open-silicon chips which can be realised immediately or in the near future and highlight their impact. We then recommend to rapidly finance projects similar (in scope and management) to the DARPA OpenRoad project for open-source EDA development. This is our strongest and most important recommendation. Next, we list all political handles that policy can operate to foster open-silicon development. Finally we present a recommended timeline for the different activities and we conclude.

Acknowledgements

This document was prepared with help from many people working in university, small organizations and SMEs. Starting from the end of September 2023, it has been reviewed by about twenty people. We are very grateful to all of them for their inputs and feedback.

Funding and disclaimer

This work is funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or of the European Commission. Neither the European Union nor the European Commission can be held responsible for them.

Project name: “Go IT!” ID number: 101070660.