Apple’s M1 silicon is all about platform control

  > However, that marginal improvement hardly seems worth upsetting the entire Mac OS developer and consumer software ecosystem.

Really? This entire article is really dismissive, but this line especially struck me. I'm not convinced the M1 chip is some strange alien tech bypassing the laws of physics, but it seems clearly superior at its specific niche (single processes with an emphasis on low power usage, using fewer cores but massive L1I and L1D caches with a shared L2 cache).

Architecturally it looks like they're sacrificing expandability and raw multiprocessing power (like massively parallel database server multiprocessing) for a single minded focus on laptop chips. Why would it be surprising it'd excel in that domain?

The author also makes the confusing suggestion that M1 is a ploy to get developers to publish to the App Store because ... Xcode produces fat mach-o's? Maybe I'm an idiot here, but all those third party apps are also made with xcode. Where is the correlation with the app store? Sure, apple has been pushing the app store, but how does "apple silicon" specifically push devs there?

If there's anything apple has done to discourage 3rd party distribution, it's through Gatekeeper and strictly requiring notorized executables. If they wanted to make non app store distribution even less appealing, they could only allow notorization through the app store, or change the default to "only app store" ... Oh wait, they did that. On Intel.

Apple has always had complete platform control. They've done a lot of crappy things -- but this article tries to dismiss the M1 for pretty absurd reasons IMO.

I consider myself a tech person and was immediately intrigued by the claims. Because I thought Apple could do it. And, after getting the Mac Mini in my hands as the only Apple computer in the house, I can mostly confirm it is quite astounding.

Maybe adding to the mystique helped some sales from curious people; it would help explain why benchmarks have been kept quiet and vague for the most part, even though they are very impressive.

I’ve witnessed single core performance on this device that I have not seen before and likely would need overclocking to hit if it can be done. For some workloads, it’s truly something.

Even thinking about it, I do expect some of this move was about performance, especially performance per watt. I doubt any other mobile processor is holding a candle at the moment. The control grabs, I think, might be secondary.

Apple is about platform control.

Apple Silicon is about performance and efficiency.

A friendly, open-hardware Apple would have moved to their own chips too, though they'd have documented their GPU, and a walled-garden Apple without a competitive laptop chip would have locked the system down anyway, as, in fact, they were doing on Intel.

As soon as you have kernel-level code execution ability you can "chainload" either u-boot or a Linux kernel directly. But yeah, the biggest problem will be the lack of documentation - for example the good old classic iPod had iPodLinux, but never got real traction.

The new M series? Probably Apple themselves will allow booting other OSes, but I would not count on any official support in getting stuff running.