In many areas of embedded, Linux is the default OS choice for a SoC these days. Sometimes Android for things with screen/UI. I expect Linux to be the default for RISC-V SoCs (not microcontrollers) also, even without any RPI involvement.
But the Board Support Packages from typical vendors tend to be as-closed-as-possible, with limited access to documentation, software updates and support for small customers. RPi comes from a different direction, much more open and accessible. So hopefully they will drag/push the vendors to be this type of open from the start in the RISC-V space.
> But the Board Support Packages from typical vendors tend to be as-closed-as-possible, with limited access to documentation, software updates and support for small customers. RPi comes from a different direction, much more open and accessible. So hopefully they will drag/push the vendors to be this type of open from the start in the RISC-V space.
Yes, I was thinking from the perspective of the end user, as in those of normal embedded devices are intentionally ignorant of it's workings... where as the audience of the raspberry pi is the exact opposite.
I know this doesn't promise anything but... the idea of a RISC-V raspberry pi! That has to be the fastest way to push Linux into the RISC-V world.
A bit curious considering how closely they have collaborated with Broadcom so far, and with Broadcom not being a RISC-V member.
Though 2 of 3 founders of https://www.lowrisc.org/ worked on Raspberry PI, that is a common connection.
I doubt they'll design their own chip, but I wouldn't be shocked to find that they joined to try to steer other partners and development into the areas they'd need to make a board comparable with a RISC-V cpu and some kind of GPU.
Need a fully open source GPU arhchitecture...
Sounds like they're just dipping their toes in the water, no public intention to build a RISC-V chip yet (unfortunately).
Makes me happy that my school ensures all EE/CE/SE students are learning RISC-V vs ARM :)
except for fully closed GPU that runs the whole thing, right? :) ARM is just a cpu core glued to the ass of the proper SoC, cant even boot itself.
It makes sense for raspberry pi go fully open and move away from the properitary arm architecture.
Whatever the 1.5% ARM royalty fee is for your org: https://www.codasip.com/2017/04/11/risc-vs-impact-on-process...
For someone like Western Digital, that's probably a huge deal. I don't think it saves the RPi folks much. I suspect they get a loss leader deal. The total volume can't be that high, but it's good press for Texas Instruments.
RPI is good press for Broadcom. Texas instrument would be BeagleBone/BeagleBoard.
Ahh, good catch, thanks.
Around 0%. Its not about cutting cost, its about finding new ways of continuing hardware effort. Broadcom has ~zero incentive screwing with outdated SoC at this point, but it might be relatively easy to sell company board on RISC-V research initiative.
Anybody knows what's the potential savings? How much cheaper can the rpi be?