Linux 5.17 RISC-V Allows Rebooting Without Needing Special Driver, HiFive Unmatched Improvements

[microcode]

Yeah, good luck getting access to those development mailing lists or proposing an extension if you're not a major industry player. LibreSoC (formerly LibreRISC-V) found out the hard way. OpenPOWER by comparison has been *far* more open and welcoming. At this point RISC-V is nothing but an industry boys-only club for companies that want to avoid arm licences.

You know, an "old boys' club" is better than a single old boy, which is the status of OpenPOWER right now: IBM is the only company that has manufactured OpenPOWER (PowerISA 3.0/3.1) chips at any scale as long as the "foundation" has existed. The OpenPOWER Foundation, as it stands, is primarily a way for third parties to market secondary services (GPU interconnects, switches, etc.) to IBM customers. Apart from Microwatt, which was created by an IBM employee, LibreSoC is possibly the only open third party implementation, and it is soft IP for FPGAs, nowhere near manufacture (and it's unclear that anyone would want to manufacture it).

I'm not here to knock OpenPOWER, but comparing bona fides between OpenPOWER and RISC-V in terms of community success, it's clear that RISC-V has a much more productive community with a lot more to offer, if you ever want to see a chip not produced by a singular vendor. From the real outcomes, OpenPOWER's "community" around cores is not much different from the MIPS "community" at this point.

[Developer12]

You know, an "old boys' club" is better than a single old boy, which is the status of OpenPOWER right now: IBM is the only company that has manufactured OpenPOWER (PowerISA 3.0/3.1) chips at any scale as long as the "foundation" has existed. The OpenPOWER Foundation, as it stands, is primarily a way for third parties to market secondary services (GPU interconnects, switches, etc.) to IBM customers. Apart from Microwatt, which was created by an IBM employee, LibreSoC is possibly the only open third party implementation, and it is soft IP for FPGAs, nowhere near manufacture (and it's unclear that anyone would want to manufacture it).

I'm not here to knock OpenPOWER, but comparing bona fides between OpenPOWER and RISC-V in terms of community success, it's clear that RISC-V has a much more productive community with a lot more to offer, if you ever want to see a chip not produced by a singular vendor. From the real outcomes, OpenPOWER's "community" around cores is not much different from the MIPS "community" at this point.

The whole reason there is any popularity to RISC-V at all is because of vendors wanting to escape ARM fees. No shit the "community" is "more productive" when western digital is porting linux to random chinese chips (K210) to swap into their hard drives.

I'm not here to argue about the "productivity" of each "community" though. Part of my original point was that RISC-V was a closed society of feed-dodging industry players, after all.

My point is that OpenPOWER is considerably better managed and run. The whole reason LibreSoC moved their 180nm chip from RISC-V to POWER was because they were denied access to the forums and mailing lists to propose their extension for the "extensible" ISA. The ycouldn't even get access to the procedure without signing an NDA.

Contrast that to OpenPOWER under the linux foundation, who were very happy to work with LibreSoC and offered them a sub-mode of the ISA to use and ease the transition from the RISC-V instruction set. Then there's the far more robust Patent troll protection in OpenPOWER, which will become a real problem later when trolls that hold media codec patents come after media encoding extensions to RISC-V.

I doubt that RISC-V will be dethroned by POWER, but I'm sure it will simply take the place of cheap ARM cores with all the associated vendor blob jank and lack of cooperation.

[microcode]

Part of my original point was that RISC-V was a closed society of feed-dodging industry players, after all.

Okay, that's just false though lol. There are hundreds of RISC-V implementations of interest to industry, research, and hobbies; and there are like three PowerISA 3 implementations that are known to work; the community has produced formal models of the ISA, reference implementations of things, a considerable number of research tapeouts, and actual products that can be purchased on the open market. I am on those mailinglists, you can just join and participate.