TL;DR – I’m developing an Open Hardawre platform based around RISC-V to replace the Wintel PC and am going to document my trials and tribulations here. This is also a bit of a recruitment drive for systems engineers/enthusiasts/users to help out with the project, link to the Github repo is at
This is a long one, so go make yourself a tea or coffee and strap in.
It cannot have gone unnoticed by many that the PC as a platform is under siege. For the past few years, development on RISC-V and ARM based SoCs has continued at a blinding pace, while improvements in the x86 world have somewhat stagnated. It is also easy to see how vendors of such chips have been using this opportunity to either develop or reinforce existing closed-loop ecosystems. One need only look at Apple’s recent release of its M1-based Macs for an example of this.
Thus, I have come to the conclusion that the PC as we know will be gone by the end of the decade. Rather than simply lie down and let the likes of Apple push us into a hellish dystopia of closed platforms, planned obsolescence and soldered-on-board mass storage, I believe that it is in the industry’s best interests if a free, open and fair standard platform were to be developed.
It would surprise many younger folks to learn that there once was a time when users could replace their P5 Pentiums with an AMD K6 or Cyrix 6x86 without changing anything else about the system. These days of course are long gone, giving way to the mess of chipset and socket changes we’re more familiar with today.
While the PC itself may be heading into its sunset years, this does not mean that open computing is dead. Far from it, in fact. I have noticed an ever-increasing interest from both the general populace and industry players in open-source hardware and ISAs like Power and RISC-V. Unfortunately, development seems to be quite stagnant in terms of developing products the average consumer would actually want to use.
SiFive’s RISC-V boards for example are little more than glorified prototyping boards, with next to no extensibility or consumer-friendly features. On the other end of the spectrum we have RaptorCS and their line of open-source Power ISA machines. While these are definitely a step in the right direction, software support for Power ISA is still very limited, and despite being RISC in name, the ISA itself is extremely complex and not well suited for simple/efficient consumer devices. Not only this, but Raptor’s boards (as well as IBM’s POWER chips) are lacking in the kinds of I/O that consumers would be interested in, focusing instead on more server-like implementations.
In my opinion, this fragmentation of the industry is just as bad as the closed-loop ARM ecosystems of the mobile space. At the end of the day, development work is still divided and consumers are still forced into a platform catered to a niche.
In my mind, the only way forward for the client compute industry is for an open, flexible and extensible platform to be adopted industry-wide. A platform that closely emulates the PC philosophy, but simplifies and modernises the technical underpinnings to prepare for the future of computing post-PC. A platform that defines a common set of hardware and standards to allow both software and hardware vendors the freedom to develop products for a standard lowest common denominator, and to reach a broad market. A platform that gives the user total and utter freedom over their machine.
And since no one else seems to be bothered trying to solve the post-PC problem, I thought I’d take a stab at. And so I did. I didn’t get very far on my own, though.
I started work on the Common RISC Platform last year by myself, as many of these sorts of things often start, thinking it would be a cakewalk. And it is at this point that the experienced systems engineers and computer architects reading have started laughing. The more I worked on the project, the more reading I had to do, and with every piece of literature I read, I found my reading list would grow by about 10 pieces. I’m not an idiot, but I’m not Lieutenant Commander Data, either. It’s simply impossible for one man to learn and store the entire sum of human knowledge on computer systems design. But I’m not giving up on this little dream
I intend to use this blog not only as a soapbox where I journal my escapades trying to develop the platform, but also as a call to arms. An invitation for any like-minded hardware developers, enthusiasts or even users who think they would be interested in making something like this a reality to check out the Github repo where I’ve committed some of my initial thoughts and design goals.
We are at a crossroads in the industry, and I strongly believe that a user-led revolt against the status quo is what’s needed to demonstrate that the current trend towards proprietary ecosystems is not acceptable to users, nor will it be beneficial in the long-term to the vendors’ bottom lines.
Thanks for indulging.
Jim