Hacker News Reader

Comment by t43562 | Hacker News Reader

The Raspberry Pi Foundation's offices are in Cambridge where ARM is and there are ex Broadcom engineers (I think?) working there so just culturally I can see why ARM is a natural connection. I think this "lets make everything RISCV" meme (like "lets rewrite everything in rust") is going to get a cold evaluation as it doesn't originate there and the cold evalution will be that ARM is a big stable player that isn't going to disappear suddenly or radically change focus, that the cost of its licenses isn't material to RPi and that RPi's success is not endangered by sticking with ARM ... as things stand today.

RP1 and software support are their strengths and I think they won't take risks with instruction set changes or even the vendor of their particular ARM chip.

So I think this strategic investment is probably equivalent to a marketing cost - it won't stop RPi doing RISCV because they wouldn't do it anyhow. It will just help Rpi reach more people.

View on HN

Interesting, I didn't realize the connections were so deep. It does make sense.

Do you think RPi might be hurt by not adopting RISC-V if it takes off? If some of their competitors offer it, could one of them dethrone RPi?

Not the person you asked, but I have opinions:

It'd depend entirely on the level of support a competitor offers, right? I mean that's why RPi is still so popular despite there having been a number of ARM competitors offering more power per dollar for like a decade now. The strong software support led to a strong community/overall ecosystem, which has led to a strong install base, which further improved the community and ecosystem.

It's a combination of first mover advantage with continued strong software support that keeps them so prominent as-is. I don't see why a different ISA would fundamentally change that formula.

Out of curiosity, who are the competitors who are competitive in power per dollar?

I don’t follow the space closely but the times I’ve checked out other offerings, they’re not competitive in price/$ or it won’t have wifi or other features to be comparable.

I actually started looking at older used desktops- which are fine for some uses and not so much for others.

Check out some of the RK3588-based SBCs. The Orange Pi 5 Plus is a fantastic alternative to the Raspberry Pi 5 if you're okay with a bit of fiddling as mainline kernel support is still a work in progress.

I just got a Pi Zero 2 W and it seems like I won't be able to connect it to a WPA3 WiFi network (standard's only been out, oh, five years or so!) without recompiling the kernel, so YMMV over in Pi-land too.

Using a Pi as an access point with Hostapd has also been buggy and crashy for years now. There have been multiple driver updates that just seem to make it worse.

I’ve been doing development on a Rock 5B and it’s really powerful for a ~$200 SBC. I’ve been using as my aarch64 CI machine and test system as my company starts to add ARM support to our software stack.

Also, OBS runs great on it, although it’s not taking advantage of the on-board MPP media accelerator (H.264/265 and other codecs). Apparently GStreamer has support for it though and some folks have experimental FFMPEG forks adding support.

I am not usually a fan of recommending youtube videos for people seeking information, but the goto I use for keeping up with SBCs competitive to Pi is the 'explaining computers' channel. It is oriented towards the lower end of audience sophistication but gives a quick breakdown of the features along with practical uses and actual setup and configuration of boards instead of just a rundown of specs.

This is exactly it, and I find it surprising no-one else seems to realise how much good software support makes an offering incredibly attractive compared to the alternatives.

No, I think their ability to offer anything comes from the use of hardware that is already "old" and made in great numbers. Then there's no chance of them getting stuck with a chip that disappears as some small RISCV company decides to stop producing it or that at best becomes a total dead end.

Offering RISC-V isn't a benefit to anyone - who really cares? I would guess a tiny number of people. Only if the performance is amazing will it matter. If that performance isn't accessible via software then bad luck too. ARM is not likely to have much trouble keeping up. To beat RPi someone would probably have to accept very marginal returns - I doubt that RPi Ltd is raking in silly amounts of money. It started from a charitable foundation and has to be barely profitable but I think it has no competitors because its situation is unattractive.

I think given they’ve been so allied to Broadcom and the ARM culture in Cambridge since the start, RISCV is never going to happen.

Other companies will make RISCV Pi alternatives in the same format, but the RPi Foundation won’t go there.

Until one of the competitors catches up to RPi's software support and stability and community, they have nothing to be fearful of. Not yet anyway.

The RPi shortage led me (and my coworkers) to examining the field of competitors, and all of them were buggier, less stable and badly documented in comparison, sadly. Perhaps that's changed since?

>I think this "lets make everything RISCV" meme (like "lets rewrite everything in rust") is going to get a cold evaluation

Speaking from a geopolitical standpoint as a westerner, I would very much prefer if the US (x86; Intel and AMD) and UK (ARM) continue to be the forefront of microprocessor tech. RISC-V for all its virtues is ultimately primarily beneficial to China at this point, which is an entity we really do not want leading the world in the coming decades and century.

Let me jump on this riscv bandwagon, thanks for the testimonial

There are people who has no choice. But there are also people who choose totalitarianism. Cannot even mourn a communist leader …

We are doomed to seek changes and happiness, those who know what is life in a totalitarian regime would wish to live differently, those enslaved by a corporativist dystopia being herd day after day until the end of days, wish to live differently, mine was more a joke about people thinking that it would change something for them, regardless who's in power

> ARM is a big stable player that isn't going to disappear suddenly or radically change focus

You've picked the wrong company to state that about. It was bought by SoftBank for way too much money, and they've been trying to make good on their bad investment ever since. The risk is low that ARM falls into a Unity-level event, but the risk is much higher than people seem to think.

Raspberry Pi did join the Risc-V foundation a while back.

Reply on news.ycombinator.com