Every single board computer I tested in 2025
bret.dk
bret.dk
Armbian do a great job of handling support for a whole host of boards (including most I included in this list), so you'll usually have Debian/Ubuntu-based flavours. Vendor kernels and vendor supplied images will be hit and miss. Mainline Linux support is a flag you filter by on the benchmark comparison site linked in the article, but it's a difficult one to keep up to date and define exactly. It could have some kind of support, but miss out on display functionality, or WiFi yada yada. What would we then class as having mainline support? All hardware etc functioning? If so, very, very few will meet that definition.
I get the desire for the information, and perhaps I should have envisioned these types of questions, but all I initially meant for the post to be was a recap for people following me to see which boards I'd tested that were released last year :D
If your standard is "supports suspend/resume", there's even plenty of laptops that won't meet it.
That gave me a laugh.
(From what I've learned so far, some magic incantation is required to convince Linux that a Lifebook E559 is a laptop not a tablet. I'm finding I have way less patience with these side-quests as I get older.)
Yup, i’m in the same boat. I’m so tempted to get one of the new macbooks and call it a day.
Disclaimer: I have never used any RISC-V board.
If you need software to be available in 2, 3, 5 years, get a raspberry pi.
Some might have some software available, some might have patches, some may need manual compiling, some only support debian with 2.4 kernel, some have binary blobs that only work on that 2.4 kernel, some have working usb ports on 2.4 and no gpio, but working gpio with 2.6 kernel but no usb ports, etc.
Just get a raspberry pi.
btw, i have inherited projects that used raspberry pis for the computing. every single one had to be reworked replacing the pi.
additionally, if the pi doesnt fit your RF footprint needs in an enclosure, it is not possible to get the chips standalone. plus the schematic is not open source. fuck broadcom and fuck raspberry pi foundation. acceptable for light hobby use only
Most of those SBCs have very poor software support. You will often need to go on GitHub or the manufactuer's support website to hunt down an OS image that hopefully works. If you want to stay up to date, tough luck. You will be lucky if your board is still receiving updates two years after release.
In the meanwhile in raspberry pi land, you can just go to download a reasonably new OS image from their website anytime you want and it will run on all their models. Even the Pi 1 model B+ which is over ten years old still receives updates, and will continue to do so until at least 2030.
Unless reviewing and playing with random boards is your hobby or job, in which case more power to you and thank you for providing valuable information to the community, you are likely better off buying a boring raspi so you can just get things done.
And then the construction quality/tolerance too. I've had Pis last for years and then cheap alternatives burn out after a few months of moderate use.
Interestingly, it’s the opposite for me, and I almost exclusively see comments about software support & linux mainlaine.
That said, I think 90% of the time it’s better to buy small x86 machine than a PI. Those have great software support, are more powerful, and can be cheaper (slightly larger & no GPIO, those two are the main reasons to go SBC)
I wonder if AI can help bridge the gap and provide the missing support that these vendors don't wish to provide.
Are you saying that even with the Raspberry Pi we are still at the mercy of the hardware manufacturer when it comes to OS images?
Raspberry Pi supports their images long term however, so you won't have to do that anytime soon.
Another benefit of raspberry pi is its popularity, there are just more projects out there compared to less known SBC manufacturers. Iirc the Archlinux arm project have images for the raspberry pi 4 (maybe 5).
Raspberry Pi usually requires customisation from the distro. This is mitigated by the fact that many distros have done that customisation but the platform itself is not well-designed for SW support.
Meanwhile many Allwinner and Rockchip platforms have great mainline support. While Qualcomm is apparently moving in the right direction but historically there have been lots of Qualcomm SBCs where the software support is just a BSP tarball on a fixed Linux kernel.
So yeah I do agree with your conclusion but it's not as simple as "RPi has the best software support and don't buy Chinese". You have to look into it on a case by case basis.
If you're lucky! Most of the time it's a questionable Google Drive link.
I can image having a very usable ARM linux laptop and tablet as a result of this — maybe even cellphone when the modems get mainlined or used via USB.
Unifi released the UCG-Fiber around a year ago that can also apparently finally handle it, but plenty of threads about slow performance with their UDMs since it's entirely done on the CPU [0].
I'm not the biggest fan of OpenWRT and would prefer something like OPNSense, but it's x86 only and good PPPoE performance isn't guaranteed either - need a CPU with good single core performance that costs more than the BPI-R4, or apparently virtualizing OPNSense allows it to process PPPoE with multiple threads.
0: https://community.ui.com/questions/What-is-the-max-performan...
I've found it to be the better choice for x86 hardware, because it performs so much better on older CPUs. FreeBSD has gotten better with driver support, but the Linux kernel in OWRT is just a better base to build off of.
I still use it though, can't complain in terms of actual routing/switching
Edit: never mind. Not official images.
Then the results aren't comparable across different boards across RAM sizes. It'd be better to test a set of different model sizes on all and report -- if it didn't fit. But could you report the full ollama model name and version size slug for each?
> I pull Jeff's fork of the ollama-benchmark software
A link would be nice.
For the 2nd, the file I grabbed initially was https://github.com/geerlingguy/ai-benchmarks/blob/main/obenc... - which I now notice wasn't modified in his repository, so I can check that out, but either way, the same version has been tested across everything thus far.
Also the zero 2 is a little under-powered. It's fine as far as it goes, it just feels like there's probably performance on the table with a more modern chipset.
There’s also a fairly usable UEFI implementation.
Had that issue with some Odroid boards, where the vendor kernel supported MFC hardware acceleration but the vanilla kernel didn't/doesn't. I'd like to avoid that
I believe I needed the vendor kernel to use video through the USB-C port, and to use the HW acceleration for transcoding in Jellyfin. This situation may have changed since my last attempts.
"Most ARM single board computers depend on proprietary binary blobs to boot or ship with outdated vendor kernels that are never upstreamed and quickly abandoned. Libre Computer takes a different approach: we fund and contribute to the mainline Linux kernel and U-Boot directly, ensuring our platforms run on upstream open-source software with minimal proprietary firmware."
I just scanned for WiFi networks and that worked fine. I also see that GPIO is not enabled for CIXP1 devices in Debian's kernel; I'll ask the kernel team to enable it.
Uh, what were they doing before? No, seriously, given that every laptop ever and most legacy free desktops, not to speak of the 1U servers you find in data centers use a single PCB, what makes a SBC a SBC?
I've seen Raspberry Pi and the like being referred to as 'open frame computers', which I thought describes them more fittingly.
We could debate about strict definitions all day, but I think the vast majority of people differentiate between a Raspberry Pi 5 (not a great example given it needs storage, pick anything with eMMC if you truly want everything on a single board) running a full Debian-based OS, and a Pico running MicroPython (or whatever your poison of choice is) and one task at a time, at least I do when it comes to this kind of thing!
I mean really though unless you have a lot of time I don't think there's anything close to worth leaving the software support of Raspberry Pi for.
Rpi and also nvidia jetson I would add.
https://www.waveshare.com/product/cm4-disp-base-2.8a.htm
I think being able to switch between a phone and desktop is a great use-case for SBC. Important to be able to attach a power bank and the real bottleneck to adoption: a telco addon.
Some sort of vibration absorbing case would also be great with a few different varieties around too: https://www.gttwireless.com/raspberry-pi-outdoors/ https://github.com/NebraLtd/IP67-Enclosure
You could take that SBC with you on a bike, car, around the home, to the worksite. Even a mobile web server for temporary gatherings placed on a high vantage point. Signage is also an important area for SBCs.
I'd wager on Polymarket that Apple will end up doing something in this space.
That would be helpful before you dive into the details, for example, I build drones, and seeing nvidia Xavier specs I would be thrilled about it until I see its size and power consumption. Great article btw!
It wasn't until this thread that I actually clicked through to the sbc.compare. I saw you complement the CIX P1 and Qualcomm in the post but it took looking at the Geekbench numbers on sbc.compare to understand the magnitude of the situation.
(I guess I'm saying the table might've helped me. But I still appreciate the post as is.)