Hmm, Raspberry Pi Ltd. joins RISC-V group (Jan 2019). Raspberry Pi Ltd. releases Rpi5 with a unified Rpi1 I/O chip (Oct 2023) freeing them from being tied to a particular SoC family. ARM Ltd. invests in Raspberry Pi Ltd. (Nov 2023). Hmmm...... Really seems like a "here's some cash, stay ARM."
Can you imagine the marketing impact of a RISC-V RPi board after all these years of it being ARM based? Sure, the number of boards effected isn't huge, but it's the marketing impact of losing a flagship product that needs to be considered.
It's easy enough to churn out SBCs in RPi format. The hard part is the software support. Raspberry Pi, while not perfect, is miles and miles ahead of everyone else. This has become their greatest strength really. It doesn't matter much if someone releases some 20-core arm/risc-v/mips/whatever board when you have to spend weeks on getting a modern kernel and distro setup, and then it all stops working in a few months and the manufacturer never ever updates the software ever again.
> Can you imagine the marketing impact of a RISC-V RPi board
IMO, the marketing value would be fairly limited. I think the main benefit would be an impetus and platform for doing RISC-V optimizations of many common software libraries & packages.
TBH, the A76 was a best-case scenario, for the Pi 5. I expected it'd be A75, since that actually has better PPA than the A76. The thing about Raspberry Pi is that they care a lot more about price and features than performance, since they want to cater to the widest market.
Eben Upton said this point was drilled in, early on, by a highly respected tech executive who told him in no uncertain terms that the Pi's #1 value proposition is its low cost of entry. Because of this, the Pi doesn't use the latest node. So, even if ARM cut them a sweetheart deal on licensing fees, it would still affect costs to fab a bigger core on an older manufacturing node (not to mention power & heat). That's why I don't think it'd be realistic for the Pi 5 to have used anything bigger than the A76.
As for IOT, RP2040 is definitely chipping away at the ESP dominance, but the missing bit is integrated WiFi/RF. A version of RP2040 with WiFi would probably become the dominant iot chip in a couple of years.
I'm glad that ARM seems to appreciate the contributions made by the Raspberry Pi & its community to the ARM ecosystem. I think the Pi Foundation didn't really *want* to shift to RISC-V, considering how long they like to support their products and how slow they were to fully release a 64-bit distro.
My biggest hope is that ARM can provide the resources for the Pi to get its GPU driver up to snuff. Let's see desktop OpenGL support at least up to 3.3, but I'd really rather 4.3. Hardware-accelerated decode of h.264 and at least GPU-accelerated decode of AV1 should be fast-tracked!
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
11 Comments
Back to Article
dwillmore - Friday, November 3, 2023 - link
Hmm, Raspberry Pi Ltd. joins RISC-V group (Jan 2019). Raspberry Pi Ltd. releases Rpi5 with a unified Rpi1 I/O chip (Oct 2023) freeing them from being tied to a particular SoC family. ARM Ltd. invests in Raspberry Pi Ltd. (Nov 2023). Hmmm...... Really seems like a "here's some cash, stay ARM."Can you imagine the marketing impact of a RISC-V RPi board after all these years of it being ARM based? Sure, the number of boards effected isn't huge, but it's the marketing impact of losing a flagship product that needs to be considered.
Sivar - Friday, November 3, 2023 - link
This is a good observation.meacupla - Friday, November 3, 2023 - link
There's already a RPi sized Risv-V SBC.Milk-V Mars CM
shadowjk - Friday, November 3, 2023 - link
It's easy enough to churn out SBCs in RPi format. The hard part is the software support. Raspberry Pi, while not perfect, is miles and miles ahead of everyone else. This has become their greatest strength really. It doesn't matter much if someone releases some 20-core arm/risc-v/mips/whatever board when you have to spend weeks on getting a modern kernel and distro setup, and then it all stops working in a few months and the manufacturer never ever updates the software ever again.mode_13h - Thursday, November 9, 2023 - link
> Can you imagine the marketing impact of a RISC-V RPi boardIMO, the marketing value would be fairly limited. I think the main benefit would be an impetus and platform for doing RISC-V optimizations of many common software libraries & packages.
twelvebore - Saturday, November 11, 2023 - link
> Can you imagine the marketing impact of a RISC-V RPi boardI can. Tumbleweed comes to mind.
patel21 - Friday, November 3, 2023 - link
Hopefully next Pi will use more newer ARM core than the 5 year old A76 used in pi5.nandnandnand - Sunday, November 5, 2023 - link
Sure, it will use A77.mode_13h - Thursday, November 9, 2023 - link
TBH, the A76 was a best-case scenario, for the Pi 5. I expected it'd be A75, since that actually has better PPA than the A76. The thing about Raspberry Pi is that they care a lot more about price and features than performance, since they want to cater to the widest market.Eben Upton said this point was drilled in, early on, by a highly respected tech executive who told him in no uncertain terms that the Pi's #1 value proposition is its low cost of entry. Because of this, the Pi doesn't use the latest node. So, even if ARM cut them a sweetheart deal on licensing fees, it would still affect costs to fab a bigger core on an older manufacturing node (not to mention power & heat). That's why I don't think it'd be realistic for the Pi 5 to have used anything bigger than the A76.
shadowjk - Friday, November 3, 2023 - link
As for IOT, RP2040 is definitely chipping away at the ESP dominance, but the missing bit is integrated WiFi/RF. A version of RP2040 with WiFi would probably become the dominant iot chip in a couple of years.mode_13h - Thursday, November 9, 2023 - link
I'm glad that ARM seems to appreciate the contributions made by the Raspberry Pi & its community to the ARM ecosystem. I think the Pi Foundation didn't really *want* to shift to RISC-V, considering how long they like to support their products and how slow they were to fully release a 64-bit distro.My biggest hope is that ARM can provide the resources for the Pi to get its GPU driver up to snuff. Let's see desktop OpenGL support at least up to 3.3, but I'd really rather 4.3. Hardware-accelerated decode of h.264 and at least GPU-accelerated decode of AV1 should be fast-tracked!