Send in your ideas. Deadline December 1, 2024
logo
Resources
Source code :
https://github.com/olofk/edalize
Documentation :
https://edalize.rtfd.io
Talk
Matrix
Stay up to date
RSS or Atom feed
Grant
Theme fund: NGI0 PET
Period: 2020-12 — 2022-10
More projects like this
Hardware
Software engineering

Edalize ASIC backend

Create open hardware silicon with a fully free software toolchain

Affordable Open Source ASIC development and custom silicon has been a long-standing goal in the community. This will unlock innovation that has previously only been possible for the largest tech companies, allowing for the creation of deployable, trusted Open Source based hardware.

Step by step, this goal has come closer in the last few years as individuals, companies and academic institutions have filled in the missing pieces. Today we have a fully open source end-to-end flow for building open source ASIC - but the effort of on-boarding existing designs remains high. This project aims to provide an easy way to onboard existing gateware and full designs to an open source ASIC flow by creating a FuseSoC backend that targets this toolchain. This will enable a smoother transition from projects already running on FPGAs to also be targeting ASIC flows. It will also allow easier switching between different open source ASIC flows at the point when there are several alternatives to choose from.

In addition to the backend itself, a reference design containing SERV, the world’s smallest RISC-V CPU, will be run through the flow and committed to actual silicon. This will provide a way to guarantee a working flow and provide a simple but usable reference for everyone else looking to onboard their designs. Enabling and demonstrating this path will allow a fully trustworthy path for the fabrication of system-on-a-chip ICs, with no proprietary or closed tools as part of the flow and hence completely inspectable at all stages. This paves the road for other more complex FuseSoC-based open source silicon projects such as OpenTitan and SweRVolf.

Why does this actually matter to end users?

When you go to a store to buy a laptop or mobile phone, you may see different brands on the outside but choice in terms of what is inside the box (in particular the most expensive component, the processor technology) is pretty much limited to the same core technologies and large vendors that have been in the market for decades. This has a much bigger effect on the users than just the hefty price tag of the hardware, because the technologies at that level impact all other technologies and insecurity at that level break security across the board.

In the field of software, open source has already become the default option in the market for any new setup. In hardware, the situation is different. Users - even very big users such as governments - have very little control over the actual hardware security of the technology they critically depend on every day. Security experts continue to uncover major security issues, and users are rightly concerned about the security of their private data as well as the continuity of their operations. But in a locked-down market there is little anyone can do, because the lack of alternatives. European companies are locked out of the possibility to contribute solutions and start new businesses that can change the status quo.

To break through this standstill, developer communities are working hard to deliver open, trustworthy and accessible alternative computer hardware that anyone can use, study, modify and distribute, just like they can with open source software. This project aims to make it easier to onboard existing open hardware components and designs into the work flow, connecting the test phase to actual end design.

Run by FOSSi Foundation

Logo NLnet: abstract logo of four people seen from above Logo NGI Zero: letterlogo shaped like a tag

This project was funded through the NGI0 PET Fund, a fund established by NLnet with financial support from the European Commission's Next Generation Internet programme, under the aegis of DG Communications Networks, Content and Technology under grant agreement No 825310.