The market is already full of ARM development boards that are pretty powerful. Just need to scale these up and put some real power on them.

Put something with the power of an M series or a Graviton on these and you have the start of a great ARM PC market.

There's nothing inherently not-open about ARM, or at least it's no less open by nature than x86. The fact that most ARM devices are locked down is a secondary effect from most of them being phones.

RISC-V would be more open than either of these but it still lags on performance. I have a RISC-V board but it's kind of slow. Not terrible but wouldn't make a good PC for anything but basic uses.

▲

therein 2 days ago | parent | next [-]

> There's nothing inherently not-open about ARM, or at least it's no less open by nature than x86. The fact that most ARM devices are locked down is a secondary effect from most of them being phones.

I'd argue lack of something like ACPI to discover the device tree and memory map is why this impression exists. Besides the ARM CPUs not being socketed.

	▲	thw_9a83c 2 days ago \| parent [-]
		Exactly. There is no agreement on how the universal operating system should expect the generic ARM computer to boot and expose its hardware.

▲

NetMageSCW 2 days ago | parent | prev | next [-]

You’d have the start of a niche hobbyist market that no one would care about. Software is needed before a market exists.

▲

hulitu 2 days ago | parent | prev [-]

> There's nothing inherently not-open about ARM,

UEFI ?

	▲	okanat 2 days ago \| parent [-]
		UEFI doesn't help with hardware discovery. ACPI does. Commonly with non-PC systems the hardware addresses are hard-coded and they need to be known by the OS somehow. Device trees are that and there are nonofficial ways of exposing them as a UEFI driver but it is nowhere as official as ACPI on PC systems.