[HN Gopher] Radxa Orion O6 brings Arm to the midrange PC (with c...
___________________________________________________________________
Radxa Orion O6 brings Arm to the midrange PC (with caveats)
Author : goranmoomin
Score : 78 points
Date : 2025-05-10 12:16 UTC (10 hours ago)
(HTM) web link (www.jeffgeerling.com)
(TXT) w3m dump (www.jeffgeerling.com)
| wjnc wrote:
| Can anyone explain me the economics of why a novelty hardware
| supplier would Not invest in drivers / software in parallel?
| There must be a good reason for this to happen, it even happens
| at the other end of the scale, but I would think the very basis
| (just any drivers, not even good ones) are an affordable
| investment? What use is your product if every review will say
| "good product, alas Windows and Linux won't run"?
|
| A reason I can imagine that drivers are (I don't know!) somewhat
| interchangeable, so invest in drivers for your product and you
| are stimulating all current and future competitors as well.
| mschuster91 wrote:
| > Can anyone explain me the economics of why a novelty hardware
| supplier would Not invest in drivers / software in parallel?
|
| There is _no single company from Asia_ that deals in mass
| produced consumer goods that 's capable of doing decent
| hardware and decent software at the same time. As soon as you
| take a peek below the surface, no matter what, it begins to
| reek.
|
| Let's just go through the stuff I personally own or have
| experience in peeking... Samsung does decent hardware, but
| their modifications to Android, or their "hacks" for
| powersaving that keep messing up apps, or their "smart" TVs
| that are buggy and slow as fuck (not to mention riddled with
| ads!)... Sony makes excellent cameras hardware-wise but the
| software/firmware side sucks ass - the fact that they require a
| dedicated software to be used as a webcam instead of just
| exposing UVC is already braindead enough, but even more so
| given that they run on Linux and the Linux kernel already ships
| with UVC gadgets. Nintendo makes excellent games but even the
| new Switch 2 ships with a chipset that's years old. Mediatek's
| leaks for BSPs / Android are frightening in terms of code
| quality.
|
| Unfortunately, the competition just isn't there. Chinese
| companies are even worse penny-pinchers than Korean or
| Taiwanese, and Western companies outside of Apple and Raspberry
| Pi just don't give a shit because they can't compete with Asian
| price dumpers or because, like many things in the ham radio
| scene, get cloned in a matter of months.
| digisocialnet wrote:
| What about DJI?
| mschuster91 wrote:
| Let's just say I'm pissed about the Mini 3 Pro's
| controller. Has an USB-C host and client port but isn't
| capable of actually using them for anything relevant, and
| the built-in wifi is rotten. Zero way of extending the
| drone's functionality by writing one's own apps/scripts -
| but the cheaper N1 that just uses a phone, there's hacks
| for it...
| cgio wrote:
| For one their app does not work on a recent pixel but does
| e.g. on an oppo...Id say GP argument is valid.
| wjnc wrote:
| Any thoughts into the Why? Culture, finance, ...? I get your
| point. Also thinking about Sony or Nintendo. Billions more to
| be made if only those firms where more technically-commercial
| (make accounts and buying stuff for parents Easy).
| mschuster91 wrote:
| Cutthroat capitalism is definitely part of the reasons,
| especially when looking at China and the low end of the
| market. When there are dozens if not hundreds of factories
| that have the skilled people to put up a hardware design
| either from scratch or as a clone of something already
| existing, "time to market" trumps everything - the earlier
| you can get the product on Alibaba, Temu or the alphabet-
| soup sellers on Amazon, the better. As soon as the hardware
| is ready and the software somewhat stable, out the door it
| goes - there is no responsibility, no accountability along
| the chain, so why invest in it?
|
| For the "big ticket" brand items, honestly I don't know. If
| anything I wouldn't blame it on culture (partially because
| I lack enough knowledge of Asian cultures, partially
| because blaming systemic issue on culture can quickly
| devolve into outright racism), but on capitalist incentives
| once again - the common standard seems to be "as low in
| terms of quality as you can get away with", there is no
| market force pushing for better products, and no
| legal/regulatory pressure either.
| drob518 wrote:
| IMO, it's company culture and focus. At Apple, they are
| focused foremost on designing the customer experience. If
| you've ever unboxed one of their products, you can tell
| from the first moment that they care about even this part
| of the product life. It's only 5 minutes and yet someone
| has clearly spent time on it. That carries through to the
| integration of hardware and software in the products. That
| doesn't mean Apple is the best at either hardware or
| software. IMO, there are companies that do just as well or
| better at either of those things. But Apple does whole
| product design better than anyone. That is something that
| is very difficult to add to an organization that doesn't
| understand it. IMO, it really started for Apple with Steve
| Jobs and being the founder, he was able to drive that into
| the culture. But it wasn't there even for them at the
| start. IMO, it really got going when Steve returned the
| second time and drove products like iMac and iPod.
| numpad0 wrote:
| Apple products aren't just well designed, but they feel
| like each has specific tiny set of target audiences they
| are built for. That would be a very politically unpopular
| move in East Asian cultures, one that could socially kill
| the targeted group.
| betterThanTexas wrote:
| > There is no single company from Asia that deals in mass
| produced consumer goods that's capable of doing decent
| hardware and decent software at the same time.
|
| Doesn't apple do most of their manufacturing in asia? I don't
| get your point. We certainly can't match this quality in the
| west.
| mschuster91 wrote:
| > Doesn't apple do most of their manufacturing in asia?
|
| Yes, but on the back of every MacBook there is the line
| "Designed by Apple in California" and the software is made
| in California as well.
|
| Asia is just chosen for manufacturing because of the close
| proximity of supply chain vendors and cheap but reliable
| labor cost.
| betterThanTexas wrote:
| Well then Asia is de-facto doing the hardware. Apple is
| just paying for it.
|
| Apple is famously sitting on a mountain of cash. How
| easily do you think they could replicate the supply chain
| of even one of their products without outsourcing it?
| drob518 wrote:
| Apple is designing the hardware in the USA, not Asia.
| Asia designs the manufacturing process, which they are
| quite good at.
| mschuster91 wrote:
| > Asia designs the manufacturing process, which they are
| quite good at.
|
| Not even that, for some parts like the aluminium cases
| ("unibody") or the laser-bored microholes in their old
| magsafe connector for the LEDs, Apple designed the whole
| process.
| mbreese wrote:
| I don't think there is a need to limit this geographically.
| As a rule hardware companies are bad at software. It doesn't
| really matter where they are from. If you're including things
| like TVs that spy on you, don't forget Visio. They are one of
| the poster children for making TVs with content tracking
| and/were they are based in California.
|
| The exceptions I see include Apple and Raspberry Pi. And even
| then, there are missteps.
|
| It's not intentional... it's just that companies rarely have
| integration as one of their core strengths. If you're a
| hardware company, you are good at making hardware. The skills
| necessary for that are very different than the skills needed
| for software. To get both, you need management that values
| both and can build the separate teams. Especially true when
| you can argue that you're working with the "community" to
| build out software and fix bugs. If you're still selling
| enough hardware, how can you say these companies are wrong?
|
| That's honestly a hard thing to do unless that is your
| competitive advantage. And for Apple and Raspberry Pi, I'd
| argue that is their competitive advantage in their markets.
| For a long time they were the small fish in big ponds. So
| they needed to have some trait that allowed them to command
| higher margins. Integration of hardware and software was it.
| bgnn wrote:
| This is slightly racist and totally wrong on many many
| points. Are American hardware companies better at software?
| Dell? HP? Vizio? Intel? AMD? Or European companies: Philips,
| Siemens? It's all the same. Even Apple has the abomination of
| iOS on their great HW..
| mschuster91 wrote:
| > Are American hardware companies better at software? Dell?
| HP? Vizio? Intel? AMD?
|
| Depends. If you're looking for consumer goods, they're just
| as dogshit as everyone else. Products intended for large
| commercial customers, particularly where support contracts
| are involved, tend to be actually decent.
|
| > Or European companies: Philips, Siemens?
|
| Never had an issue with Hue, never had an issue with Bosch-
| Siemens "white goods".
|
| > Even Apple has the abomination of iOS on their great HW..
|
| iOS certainly is not an abomination. I'm using both Android
| and iOS, and the latter is much more polished.
| markvdb wrote:
| s/from Asia//
| numpad0 wrote:
| _There is no single company from Asia that market electronic
| physical experience_. Other than Nintendo. IOW there 's
| Nintendo.
|
| East Asian companies hasn't shifted into maximizing perceived
| values over functions. I think it just feels fake and wrong
| to many. Another factor that I think might exist is, it might
| be simply hard to inflate values and sell experiences as East
| Asian company entities without good connections and/or
| cultural understanding to sell to developed Western markets.
|
| East Asian engineers don't share the pain points as Western
| audiences. People don't use webcams - less sympathies exist
| for the urge to see faces. People don't use Linux on laptop -
| Windows is normal and fine. Don't recognize annoyances as
| annoying as often - it's considered signs of weakness. (Not
| sure about Nintendo complaints - people everywhere happily
| pay $80 for another Mario remakes on top of $500 console.
| Isn't that what you're asking for?). Kids aren't as
| interested in ROM cooking or piracy or software freedom in
| general - way more interested in enforcing IP rights
| themselves than demanding something from IP holders.
|
| Or, widespread "just don't give a shit because they can't
| compete with Asian price dumpers" could be another reason.
| East Asian nations each has its own internal markets with
| massive surplus production capacities. Western companies had
| progressively moved into value engineering for survival, as
| Swiss watch industry famously did. The costs of East Asian
| physical products never represented its cost in the first
| place, only utility in context of economy at export
| destinations, and that might have affected how companies at
| the destinations have come to be.
| ZiiS wrote:
| To a certain extent if your reviewers can just turn it on they
| spend their time benchmarking.
| geerlingguy wrote:
| Some of us try out every possible interface... but it turns
| into quite a slog, as usually only about half the advertised
| features work by the time the board ships to the public.
| Things like GPIO, display interfaces, NPUs, etc usually
| require a lot of tweaks if they work at all.
|
| The O6 was better in that regard than many boards, but the
| bar is not very high.
| jrmg wrote:
| With the dearth of drivers as described, how do the
| manufacturers even know that all the hardware works? That there
| are no flaws in how it's all wired together? How do they test
| everything?
| Palomides wrote:
| the chip manufacturer usually craps out a hacked up android
| build
| dijit wrote:
| ok, but, naively... Android is Linux with a userland- so
| the drivers must exist in some form.
| kanwisher wrote:
| Android has some ability to have binary drivers that
| aren't easily reusable for normal linux branches
| notpushkin wrote:
| Yup. There's libhybris/Halium which wraps Android drivers
| for glibc-based Linux distros, but I think it's more of a
| hack / stopgap solution.
| arghwhat wrote:
| From the perspective of drivers Radxa is not the hardware
| manufacturer. They combine off the shelf hardware according to
| its documentation to create a product.
|
| Sometimes the original suppliers will have drivers, sometimes
| they just ship documentation and let it be up to the customers
| to write it, sometimes someone else contributed upstream
| support. When you get "drivers" from e.g. Lenovo, they didn't
| write them - they're just sending what they got along.
|
| Nothing would work if there weren't drivers in general, the
| issue is that hardware can be configured in multiple ways and
| it's not all going to have have proper support or be well
| tested. In Linux land, this stuff sorts itself out as people
| get their hands on the hardware, pretty similar to how e.g.
| laptop support comes to be.
| rjsw wrote:
| ARM themself have an open rec right now for someone to work on
| GPU drivers, the one in this SoC isn't supported by panthor.
| buyucu wrote:
| It looks like they are trying to mainline the code to Linux:
| https://patchwork.kernel.org/project/linux-arm-kernel/cover/...
|
| My guess is that they wanted the board to be available to devs
| early to get feedback. I might buy this board in a few months,
| when it will likely work out of the mainline kernel.
| jauntywundrkind wrote:
| ARM has such a mess. They don't actually make chips. And they
| don't make a sizable % of the design that goes into chips.
| There's all kinds of other IP, from power management to USB
| controllers, all manners of things they don't do.
|
| So then the task falls to the chip makers. Who each are trying
| to figure out what to do themselves.
|
| These folks don't usually want to do that or have the chops.
| There's usually one or many different folks taking off the
| shelf open source and porting it to this particular platform.
| These folks have a strong strong anti-incentive to do the right
| thing to upstream support: no one's gonna keep buying sdk's
| from these software vendors if they upstream support. And it
| _is_ a pain to upstream support, to spend possibly years
| figuring out the long term way to do something.
|
| (Notably some good players who focus on mainlining have
| emerged: Bootlin, Collabora.)
|
| It's all so terrible. Theres some attempts to mature the
| platform, to make some standards so at least folks can boot
| something maybe (SystemReady and an array of neighboring
| acronyms). But man, it's so bad, your question is so searing,
| so obvious. This whole world systematically seems unable to do
| the right obvious good thing for itself, has resolutely
| remained a shitty backwater for 3+ decades versus x86.
| ChocolateGod wrote:
| > to make some standards so at least folks can boot something
| maybe
|
| There's still no incentive to do this outside of server
| hardware (Red Hat doesn't support rebuilds of RHEL). Why
| bother making a firmware stack that allows anything to boot
| when you can just modify the Linux kernel to work on your
| non-standard hardware and work around any bugs, publish an
| Ubuntu image that you'll only update a few times and then
| call it a day. Then, maybe, just maybe, support for your SBC
| will be upstreamed a few years later, but long after it's
| actually useful.
|
| Look how messed up and fragmented Android is still, even
| Windows Phone had a UEFI stack that in theory could allow all
| Windows phones of a certain architecture to have a shared
| image (although I don't believe this was fully achieved).
|
| Imagine if Windows or Fedora for x64 had to be repacked for
| every single motherboard/CPU combination, it'd be insane, yet
| somehow it's fine for ARM, what a joke it's been allowed to
| become. I think electronic waste regulation is going to be
| needed to sort it out.
| numpad0 wrote:
| Someone told me once, yes we set breakpoint _on the CPU_ ,
| doesn't do anything to the AMBA bus on SoC it's a peripheral
| you know, the bus clock keeps going and DMA keeps going in
| and out if you stop the CPU, that's how we debug stuffs young
| man.
|
| I was like, that sounds fun...
| wmf wrote:
| Drivers are extremely expensive to develop (like 10x more than
| the hardware) and mostly interchangeable so every company is
| trying to free-ride on someone else to develop the drivers.
| zabzonk wrote:
| Is this so they can say they are "Orion Arm", which is where we
| are in the Milky Way galaxy?
| drob518 wrote:
| Clever
| zettabomb wrote:
| I bought one of these, with the full 64GB of RAM. So far it's
| been a fun machine to play with. In UEFI mode I can install
| Fedora 42 with essentially zero issues (it tells you that the
| bootloader didn't install but actually it did and works fine),
| which is quite smooth for ARM. It will be nice to see the CPU
| clusters, GPU/NPU drivers, and various PCIe snags worked out, but
| I really like what it has at this price point (assuming you're
| not in the US).
| irusensei wrote:
| Does the CPU includes something like an fTPM? The documentation
| also suggests there is an unsoldered TPM included.
| zettabomb wrote:
| As far as I can tell, no. It's a separate footprint on the
| PCB, not difficult to add if you've done a bit of soldering.
| I haven't tried that yet.
| cenamus wrote:
| I actually had the bootloader error message on my install too
| (regular intel amd64 CPU), so might be something else
| buyucu wrote:
| does Vulkan work? It's a must-have for llama.cpp.
| aseipp wrote:
| There is a fork of Mesa with some support for the onboard GPU
| (Immortals G720), but it's not upstream yet and might not be
| for a while. Some people on the forums have installed various
| discrete GPUs, which would obviously work (modulo bugs.)
|
| There is also an open source driver for the NPU somewhere
| (Zhouyi NPU) and some documentation, but nothing in an
| upstream kernel yet. https://zhouyi-npu-
| tutorial.readthedocs.io/en/latest/0_radxa...
| jasoneckert wrote:
| > Prices for those in the US (like me) just tripled due to import
| tariffs (ordering the 32 GB model went from $400 to $1500).
|
| This was the biggest takeaway for me in this post. In the past,
| to experiment with a particular piece of new hardware, we had to
| a) obtain the hardware, and b) obtain or create software for it.
| With a) fast becoming out-of-reach for most people, this puts a
| dampener on b).
| danieldk wrote:
| There are a lot of open source developers outside the US. While
| development will certainly take a hit, the world is larger than
| the US. For more niche boards, it may not even change things
| much - they are often hard to get due to limited supplies and
| now more boards will just end up in the hands of developers on
| other continents.
| _whiteCaps_ wrote:
| I'm not so sure about that. I think it's also possible that
| these niche boards won't get built if they aren't able to
| sell them in the US market.
| buyucu wrote:
| China has 1 billion more people than the US. These board
| would still get built, it's just that you won't be able to
| read the manuals because they will be in Chinese.
| doawoo wrote:
| Reading this line made me so mad. I've been a huge fan of lower
| power ARM CPUs... this literally just halt/hampers progress in
| this country. I can't believe we have to put up with 4 full
| years of this crap.
|
| I would have purchased this board in a heartbeat otherwise.
| Ugh.
| cyanydeez wrote:
| ....atleast 4 years
| irusensei wrote:
| The whole tariff situation will probably dry out a significant
| amount from the incentives for such boards and it seems the only
| good options for a powerful ARM personal computer or home server
| are basically Macs or Ampere.
| walterbell wrote:
| Asahi Linux is making slow but steady progress on Macs.
|
| Ampere was acquired by Softbank, owner of Arm.
| buyucu wrote:
| This sounds like a cool board. Does anyone know if the APU works
| with llama.cpp?
| ksec wrote:
| ARM China, and CIX, Cix CD8180 SoC, Armv9.2 Architecture.
|
| I was under the impression that ARM China doesn't have the latest
| license to Armv9 and stops at Armv8. While ARM HQ opened a
| separate ARM Unit in Shanghai under a different name ARM
| Something ( Some Chinese Phonetics ). But CIX has had this SOC
| with Armv9 announced a while ago. So I assume ARM China is now
| officially back under ARM HQ / Softbank control?
|
| By Control I dont mean just swapping a new CEO but the actual
| power structure of the company.
| qwertox wrote:
| Radxa has been notoriously bad at supporting their hardware from
| the software side.
|
| Their hardware is great, but they launch a product and then won't
| offer a proper distribution for it, hoping that some developers
| will take care of this for them, for free.
|
| This is why I love Raspberry Pi so much: they care about the
| software just as much as about the hardware, if not even more.
| And that is great. Because the hardware, once you have it, that's
| it, it won't change.
|
| I don't know about Raspberry OS, but when I installed Raspbian 12
| Bookworm on my first-gen Raspi with 500MB RAM, it worked. It's
| now working as a VPN server.
|
| See this, for example: The Zero 3e is a really great board, but
| this is the software they offer for it https://github.com/radxa-
| build/radxa-zero3/releases
|
| 3 weeks ago: internal test build
|
| Apr 8: internal test build
|
| Jan 10, 2024: beta 6: Currently there is an issue preventing the
| Debian CLI image from booting, and we suggest users to use the
| Desktop variant instead for now. Ubuntu CLI: This flavor is
| provided as-is except for critical issues. Users should look at
| Debian CLI as an alternative.
|
| There are now community maintained Armbian variants, but it took
| a long time for them to appear. There was also a distibution by
| some other volunteer, but Radxa did nothing.
| zozbot234 wrote:
| AIUI, even the Raspbian is only supported by a somewhat hacky
| downstream kernel to begin with. It takes time to achieve
| proper upstream support, and even then the community can only
| succeed due to how popular the Raspberry Pi hardware is.
| als0 wrote:
| It at least supports SystemReady. Presumably any OS that is
| SystemReady compatible will just work?
|
| https://github.com/ARM-software/arm-systemready
| geerlingguy wrote:
| The SystemReady firmware is currently a bit limited. Not only
| does it still have high idle power consumption, it disables
| 4/12 cores.
| solarkraft wrote:
| > once I installed the Nvidia proprietary driver with sudo
| ubuntu-drivers install nvidia:570, it was quite stable
|
| Didn't expect this to just be available for ARM. It really is
| making its way out of ,,weird niche platform" territory to ,,it's
| just a PC"! Especially together with the SystemReady firmware.
| wmf wrote:
| The ARM Nvidia driver is needed for Tegra/Jetson and Grace
| Hopper so it has been in the works for a long time.
| walterbell wrote:
| Nvidia spent time trying to acquire Arm, before regulators
| blocked. Project Digits client on the way.
| sunshine-o wrote:
| > If you're just doing AI stuff or GPU compute, this board might
| actually be a decent option, all things considered.
|
| Looking at the benchmark [0]
|
| When I first saw the board I thought this might be the RPi of the
| "AI age" since it is IMHO the most affordable option with 64Gb of
| RAM.
|
| But I am always cautious if we can really make the most of the so
| called "30 TOPS NPU"
|
| - [0] https://github.com/geerlingguy/ollama-benchmark/issues/13
___________________________________________________________________
(page generated 2025-05-10 23:01 UTC)