A couple of weeks ago I wrote about the ROCK 5 ITX coming soon and since then, samples of the Rockchip RK3588-based Radxa ROCK 5 ITX have been landing on doorsteps (or service points, screw you, UPS) of a lucky group of people and somehow I was one of those, so here’s a first look at Radxa’s latest Single Board Computer in a Mini ITX form-factor!
It’s going to be a photo-heavy post and I make no apologies for that, it’s a very nice-looking PCB, with the black and gold colour scheme looking very stylish. I imagine that was a very conscious decision seeing as, as expected, they’re marketing this as a low-power desktop option and you probably don’t want a plain Jane motherboard taking pride of place in your new system, right?
↫ Bret Weber
Now this – this, my friends, is exactly what the doctor ordered. I can’t wait for standard, ATX motherboard sporting ARM processors to become more common and readily available, hopefully standardised better than what we’re used to from the ARM world. I want my next (non-gaming) machines to be ARM-powered, and that means we’re going to need more of these ATX ARM boards, spanning wider performance levels.
No PCIe slots? I can’t believe an open PCIe slot is still a unicorn of a feature on ARM boards.
Drumhellar,
I’d love to have an ARM system that has uncompromising feature parity with x86. Including PCI & expandable ram/GPU/disks/BYO-OS/etc. The whole shebang.
Wereas this stuff is commonplace for x86 at commodity prices, unfortunately this sort of flexibility in the ARM world is niche and expensive. To an extent I’d be willing to give up certain “x86 privileges”, but I really don’t want to give up the ability to boot up my OS of choice without the obstacles that have made ARM such a pain to work on. This is the #1 problem with ARM at the moment IMHO and frankly it’s bitten me enough times that I dread the ARM experience. Honestly I rather hardware manufactures to have no official connection to the OS. “We don’t provide an operating system, all the drivers & patches you need for this hardware are already upstream, go install an unmodified instance of debian, ubuntu or your own distro” – this would be so refreshing!
It’s getting closer. ASRock Rack has an MB for the Ampere Altra. Proc plus MB is listed at $1500 which is within range of a high-end x86 board and proc. This is reasonable enough for me to consider it for a home server.
The Ampere Altra procs are the procs powering some cloud Arm services. GCP has Ampere, and Hetzner definitely has Ampere procs. The little Arm VPS I’ve rented from Hetzner has worked well. There are multiple Linux distro images available on the order menu, and the latest FreeBSD and OpenBSD versions are listed in the ISO menu.
https://www.asrockrack.com/minisite/AmpereAltraFamily/
They would have to pick a different proc for that. That little proc probably has it’s IO maxed out.
ASRack has an Ampere Altra board with lots of PCIe. That’s $1500 for the kit, which is quite a bit more then this, probably.
https://community.amperecomputing.com/t/asrock-rack-ampere-bundle/603/3
Especially when the box says PCIe 3.0, I would expect a slot, right?
The PCIe lanes are used for the storage, and probably other things. This chip isn’t designed to be a desktop chip, but Radxa is making it work.
I would love to skip over ARM and go straight to RISC-V. So, I am waiting for the Milk-V Oasis:
https://www.phoronix.com/news/Milk-V-Oasis
You can buy a RISC-V desktop today if you have the coin:
https://milkv.io/pioneer
Both the above boards have PCIe slots by the way.
The Milk-V stuff seems to be getting mainline kernel support. This has been the Achilles heal of ARM and an endless problem with SBC options. I think RISC-V may do better in that dept.
tanishaj,
I worry that realistically RISC-V might remain niche long term and competing with more expensive CPUs with worse specs. But I do hope you are right! If RISC-V manages to solve ARM’s notorious compatibility and support issues, I think many SMC/embedded linux tinkerers would benefit greatly.
You may be correct. That said, of the three segments of the market that I see ( low-end microcontrollers ), RISC-V is already making a dent. In that market, the cost of the ISA matters. The hassle and lack of agility required by licensing is also an issue. The inability to easily customize matters. In that space, the ability to customize on top of RISC-V has an advantage. I expect an Open Source hardware movement and ecosystem much like we have seen in software.
At the other end, you have customers that design their own chips. Apple is not switching ISA any time soon but, if you are Microsoft or Amazon, there is nothing really holding you back from using RISC-V to offer the software services that you sell to others. I would not hang my hat on it but perhaps the extensions system offers something interesting.
That leaves the middle of the market where people buy processor designs off of ARM. RISC-V by contrast will likley evolve into an entire market of CPU design providers. Historically, this kind of open ecosystem tends to dominate over a single-source provider. ARM is not exactly a classic single-source of course and there are no guarantees that a market of RISC-V providers emerges.
Another factor is that RISC-V is a safe option for centralized investors like the CCP. If you are in a legal fight with The West, the inability to prevent access ot the RISC-V ISA is a real feature. Goverments will invest in RISC-V designs. Universities may collaborate on RISC-V designs. You may get “alliances” like we see with AV1. My own view is that we will see enough of this to create a base of ARM competitors that will drive the RISC-V ecosystem. If this happens, I think we will see the big “custom chip” players adopt RISC-V too. If not, RISC-V will stay niche as you suggest.
Supposedly RISC-V standardized the boot process. It was mentioned somewhere, so that part might not be a problem.
I think they are also mandating feature sets, so this might solve the problem of different procs being wildly different.
The support issues will probably be there. Companies are jerks after all.
Flatland_Spider,
Yeah. I’m definitely intrigued. And I wouldn’t mind paying as much for a RISC-V system as I would an x86 system. However I don’t want to be an early adapter. It’s too much money for something that may not be well supported.
https://www.reddit.com/r/RISCV/comments/1bgkelm/milkv_pioneer_owners_how_is_your_experience/
Yeah. $100 dev board is different then $2500 for a complete system. Good for them for having a complete system so early though. It would be a good project for someone who wants to be a kernel hero.
You know, it’s not that bad considering what you get. If you’re already a developer wanting to target RISC-V it seems to be a worthwhile investment. US$2500 for a 64-core system with 128GB of RAM and 1TB of SSD storage, and a modest GPU, in a nice case and able to be upgraded easily, is not too shabby. Granted it won’t have the performance of a $2500 AMD64 or Apple Silicon machine, but for those seriously digging into RISC-V it’s pretty much a no-brainer.
Pico-ITX is back! This looks really cool. 🙂
It checks all of the boxes, and it looks like a fun little low-end server.
The one concern is the usual concern. Can a stock Linux distro or BSD be installed on it?
I just realized no ECC support on the RAM. 🙁 Rather disappointing for a “server” board.