The industry is all about x86, ARM, and RISC-V, and sticking with the majority certainly has its benefits. But I’d argue that PPC is already “there” in the server space, as it’s been powering high performance systems for two decades now. When Apple moved off from PPC, IBM continued. I have a Power9 system, and life is good, if you can stick to only open source software.
But yeah, an OpenPower CPU in anything other than a server is wishful thinking, and seeing MicroWatt in real silicon vs just an FPGA is also wishful thinking. I’d hope that the industry would support more diversity in this aspect, but I guess not.
Edit: there’s also OpenSparc, which I think has had less activity than OpenPower, but still has something.
Fedora has RISC-V as a kind of non-official port. I posted the news article link above to a RISC-V thread on the Fedora devel@ mailing list. It’s to discuss if it’s time to prepare to add the RISC-V CPU to Fedora’s build system as an official port. And here is a RISC-V thread on Fedora Discussion Forum as a reference.
My desktop is a HiFive Unmatched running OpenBSD and to be honest, it’s a very long way from being usable as a daily driver for most people. The performance is just not good enough for the average person. It’s fine for messing around in Xterm, or for some light tasks and Netsurf, but any modern web browser would crush it, even if there was a functioning port of them to RISCV. I do know that there are newer boards than the Unmatched, but have not tried them.
The other issue I see here is stability, both in terms of hardware and support. Even the ISA is not set in stone yet, and there’s a good chance that all the various non-standard extensions to it will cause incompatibility between different RISCV boards from different manufacturers. Plus, the BIOS on these things is not the most stable. It is very finicky, and that would be a negative experience for most people.
Despite all that, I do think this is a good chance for someone like SiFive to make a Framework compatible motherboard for people who want to tinker with it in their laptop, which will iron out the bugs until it’s ready for the general public. In fact, it’s worth reaching out to them and asking if they’d consider it. There’s no reason Framework has to be the only one building motherboards for FW laptops–that’s the whole point.
Disappointed that the mainboard features a mainland China RISC CPU with a board produced by a company from Hong Kong. There were no US/EU/JP/TW sources for either?
Unless there is a hotbed of RISC-V development most people are not aware of, the choices of who to partner with are quite slim. It is an open standard they are trying to promote as the majority of major processors in the world are proprietary.
It is more of along the lines of proof of concept and putting a development board into a known format instead of everything else that is a one-off development project for prototypes. If it helps advance into more smaller products it is a benefit to everyone.
Remember a few years ago when most of the auto industry was hampered because the chips they thought would be around forever, weren’t? I will give you one guess which company did not have as big of an issue because they modularlized their control systems (door locks, body control module, HVAC, dashboard, etc.) [Hint they make electric vehicles…]
All those little control systems in a vehicle could run easily on something with a small microprocessor, sure would be nice to not be locked into a single supplier to bring their product to market as things change over time.
That’s a feature, not a bug. The point of Risc-V is to break the cartel of “Western” IT gatekeepers like Intel, AMD and ARM. It’s not surprising that the keenest adopters are in so-called BRICS nations.
What I find sad is that replacing hardware does not go hand in hand with exciting new operating Systems. If you are just running the same tired Windows and Linux rubbish as we curently do, then your spanking new architecture does not really make any difference to the user experience, except maybe some battery life and speed. In fact there will be a difference - old faithful, proprietary programmes that people never needed to replace because they worked, now won’t work.
Surely a transition to new architectures is a great opportunity to jump to legacy-free software. Haiku, for instance, is a strong early adopter of Risc-V and currently booting on various single board computers.
I’m talking in re FW decision here, not the Chinese one. SiFive is a US company. Cortus is French.
Supporting mainland China chip manufacturer is a practical equivalent of supporting their war effort and we’ll be at war within the next 3-5 years, if not sooner.
True. It is great seeing the upsurge of interest in Plan9 and Haiku (others will no doubt name their favourites).
But a new hardware means porting over all the programmes - assuming the source is available or the original vendor is still trading.
So why not take this opportunity to entrench a more modern operating system? Haiku for instance has tools that makes it easy to import Linux programmes. And newer system Genode has tools that make it easy to port over Haiku programmes.
The major undertaking of porting all Windows or Linux and it’s software to a new hardware would not be vastly harder if this was also to a new operating system. In fact it would likely be easier because the younger systems have less cruft and certainly a smaller software catalogue.
Has there been any discussion on when the schematics and such for the RISC-V board will be available? Do we know if the processor can be replaced or resauldered with a different chip? I’m interested to see if there are any individuals looking at hacking the board itself.