Free and Open Source Silicon (FOSSi) are components and systems that are inside silicon devices (‘chips’). It is our core belief that building blocks that form such digital devices should be free and open.
Digital systems are nowadays created not only by industry experts, but the community of hobbyists and academics is steadily increasing. The introduction of the FPGA (a re-programmable digital hardware device) is the key enabler for this development and communities have evolved around open source ‘IP blocks’ and entire System-on-Chip.
FOSSi is a good way for hobbyists and academics to collaborate and re-use existing building blocks to build their own systems. But it is not limited to such non-commercial users. Instead, there is an increasing number of companies that use open source IP blocks and make their own development open source. Beside the publicity companies gain communities and involve people around their designs by doing so.
We hence see the three pillars of companies, hobbyists and academics for the success of FOSSi.
The starting point of a digital hardware design is a hardware description, most often a hardware description language, sometimes schematic. This is the source code of a design. This is then synthesized (compiled) into netlist that is the structural definition of the hardware design. The next major step is the Implementation and Placement & Routing of the design to the actual hardware output. In case of a chip (ASIC) this step results in the creation of masks for the silicon production, while in case of an FPGA the result is a bitstream which is used to program the FPGA.
While the concept of pre-synthesized netlists exists (mostly for closed digitial designs), the standard way of exchanging FOSSi components is on the hardware description language level. FOSSi covers everything from building blocks (for example a UART controller or a processor) or complex System-on-Chip.
Two non-functional aspects are critical to FOSSi: Clear interface definitions are required for interoperability of different building blocks and building blocks must be trusted to follow well-defined (open) protocols. Furthermore, reliable testbenches and test strategies are required, because hardware bugs are much harder to cope with as they can in the case of silicon production not be fixed.
The open source software movement is similarly old as the open source hardware movement. For example the Homebrew Computer Club in the seventies can be seen as the big bang of both movements where hardware design and the software for it were exchanged freely. Not only has an entire industry evolved from this club, but also the discussion around open source software has been a dynamic topic over the decades since this time. After periods of large companies with closed software dominating the market, open source software is widely accepted nowadays, at latest since Linux and collaborative platforms on the Internet. Open Source Hardware has had a similar history, but at much smaller dimension. But over the last decade the “maker scene” has evolved in the tradition of the Homebrew Computer Club.
Free and Open Source Silicon is slightly different in the conditions. Distributing software and compiling, modifying and testing it is very easy nowadays. While FOSSi sources can generally be distributed as easy as software source code, it is very expensive to build a chip out of it. Fortunately, it has become moderately expensive since the introduction of the FPGA and nowadays FPGA development boards are available for as low as 25 Dollars. Nevertheless, the community has been much smaller and often limited to experts as the traditional learning curve for building hardware was a bit steeper as in software programming (we are of course not claiming it is harder to design excellent hardware than to design excellent software).
The Open Source Hardware movement has seen a big growth in the last decade, mainly driven by cheap embedded system platforms like Arduino or RaspberryPi and the advances of collaborative (software) development online-platforms. While there also was an increase in FOSSi in this period, OSHW has grown much faster until now. We think that the main reason why FOSSi did not see that breakthrough until now is on the one hand the cost and difficulty of production and on the other hand the reluctance of companies from this field as it still seen critical.
Beside fostering the development of Free and Open Source Silicon we encourage the open source development of the FOSSi ecosystem. This ecosystem is mainly the Electronic Design Automation (EDA) tools, that have been traditionally expensive and required a lot of experience to use. This has improved over the last decade and the FPGA vendors provide free tools for their small FPGA devices. We encourage the development of open source EDA tools, especially for system synthesis and testing.