The day may not be far when Intel and ARM aren’t the only ones making processors for your laptop and phone. There has been a growing trend of open source development in the tech industry in recent years. The field of computer architecture is no different. Historically, processor designs have been owned by companies like Intel and IBM with few details about the actual design being released to the general public. RISC-V is about to change this. Developed at the UC Berkeley, RISC-V is a new open instruction set architecture (ISA), which forms the language of a computer. Its main goals are to promote free open sharing of processor designs among the academic community and to become a standard open architecture for industry implementations.
A report published by UC Berkeley explains the need for free open ISAs for decreasing the barrier to entry in computer architecture research. Free and open source hardware implementations are only possible using open ISAs. As RISC-V gains popularity, it is expected that new implementations from all over academia as well as industry will start emerging.
In several academic communities, this is already happening. A research group at MIT’s Computer Science and Artificial Intelligence Laboratory has been working on an open source implementation of the RISC-V ISA, a project called Riscy Expedition. The project has been successful in creating basic versions of a processor using a programming language called Bluespec. Bluespec offers a way to write high-level, extremely modular source code, an attribute which is highly appreciated in the computer science world.
The next step in the project is to create advanced versions of the processor to achieve better performance. One of the goals is to show that hardware implementations using high-level languages such as Bluespec can perform on a par with industry standards. Riscy Expedition is not the only project out there which is creating open processors using RISC-V. UC Berkeley has their own implementation of RISC-V called Rocket and they are also conducting research in parallel with MIT.
Research in computer architecture has always been difficult. Successful ISAs like x86 and ARM, which are used in almost all computers, are patented and hence need licenses. As the UC Berkeley report points out, “Negotiations take 6-24 months and they can cost $1M-$10M, which rules out academia and others with small volumes.” However, there is no real reason for these ISAs to be proprietary except for historical or business reasons. Hopefully, open ISAs like RISC-V will lead to an increased interest in computer architecture and we will begin seeing new processors designs from other communities.