While open source software is taking over the world, a push for open source hardware has been quietly building.
The RISC-V Foundation has been pushing its open sourced instruction set architecture for chips based on the long-established paradigms for reduced instruction set computing. And one of its most vocal advocates is Calista Redmond, the chief executive of the RISC-V Foundation, which is working to promote its adoption.
This is a slow burn. RISC-V won’t change the world overnight, but will slowly but surely seep into every corner of the computing industry – and looking at the incompatible, closed-source mess that is the ARM world, we really need RISC-V on all those millions of embedded devices we use every day.
It’ll work on microcontrollers and very simple/low end but there is a reason ARM is successful in the middle/high end. Most of the companies that build the chips don’t design the CPUs (the exception are very big companies like Qualcomm, Samsung and Apple). The license the core design from the ARM company and then they package it with different things.
It’ll be difficult to take off because of legacy, the ones that have it easier are Apple because they completely control their ecosystem.
Um, why? At the end of the day you can’t inspect the silicon to determine how much the manufacturer modified the open source design in which ways, and can’t just grab the source code and make your own chip out of duct tape and lollipops; so it’s effectively identical to closed source for all practical purposes (except for licencing fees).
For a simple example, RISC-V is currently being used by Nvidia (for managing GPUs) in some of the most proprietary products on the planet.
So many articles have this wrong.
RISC-V is not open source.
It’s open spec. Quite a difference.
Companies are completely within their rights to take the open spec, develop proprietary addons to it, or even diverge from the open spec itself, and sell it, providing no ‘sources’ whatsoever.
As an analog to software, RISC-V is the API.
The implementation can still use that API and be completely proprietary in implementation, providing no source or credit.
.
All that being said, it is still very exciting and a shit show better for innovation than what we’ve had with other ISA’s to date.
As LKCL has pointed out, the RISC-V ecosystem is very closed and proprietary in nature, even though it is free for anyone to implement. This is probably inevitable as any such open hardware project can only really exist with industry backing, and the industry only cares about openness as a way of cutting costs (eliminating royalty fees). RISC-V is really an anti-ARM cartel, to avoid paying pennies per unit to ARM.
I think RISC-V an anti proprietary cartel to lower the bar to hardware innovation. It’s a legitimate move and ARM need to adapt or die. Latter is unlikely. Here is another company that might suffer, however their approach to supporting custom hardware seems slight years ahead.
https://0x04.net/~mwk/doc/xtensa.pdf https://ip.cadence.com/news/156/330/Tensilica-Offers-Low-Cost-Development-Tools-for-Diamond-Standard-Processors
Arm is reacting but the ticket to play still seems to be $75K.
https://www.theregister.co.uk/2019/07/16/arm_licensing_change/
Yeah, cutting costs is lowering the bar to innovate. No question. But To software guys, it won’t mean that much. Maybe a new syntax of assembly to be versed in, and a new architecture to figure out how to standardize operating systems on which will face the exact same hurdles as arm architectures.
Bill Shooter of Bul,
Well, it would mean a lot to me if the bootup environment were completely standardized and could run mainline linux (and other operating systems) out of the box. This is a major hassle with many ARM SBCs. I don’t have experience with RISCV, but assuming 1) they are able to become competitive with ARM on performance, 2) they ditch any and all proprietary drivers, and 3) they are able to standardize the bootup process, then I for one would have no problem ditching ARM for RISCV. As much as I wanted to like ARM for years, IMHO they never did a good job with respect to #2 and #3.
However, I think the future’s going to be tough for RISCV proponents. In industry, a newcomer can fail to displace defacto standards like x86 and ARM regardless of it’s merits just because of strong network effects. And while the “openness” of RISCV is certainly appealing for many of us, it’s not a foregone conclusion that it will translate into openness in end user products. Considering the way nvidia is taking advantage of RISCV’s openness for itself while providing a proprietary product to us, the end users, the manufacturers might well embrace royalty free RISCV CPU cores for their own benefit, meanwhile continuing to push out mostly proprietary products. I hope this is not the case in general, but time will tell if manufacturers give us open platforms for these “open” cpus. I want to keep an open mind, but I am concerned because the industry keeps slowly moving in the direction of more user restrictions and locks.
I did see a relatively large Risc-V presence at Embedded World last week. The ecosystem is growing fast.
However your prediction that it will be mostly used in proprietary designs is likely to come true.