Interesting post - Passing of Wintel, thanks to ARM

[donawalt]

This is an interesting article written by Jean-Louis Gassee, a name you may remember as one of the execs in the Apple low years of the 80s, he worked under Sculley and Jobs. He writes an interesting article wondering if Apple Silicon won't put pressure on Microsoft to beef up their Windows ARM and compatibility with their own products. This will further put pressure on Intel and break up the Wintel monopoly since Windows/ARM will be a legit other choice; this will put pressure on the PC manufacturers to bring out ARM computers since they will prove to be more than competitive in the marketplace. It's one guy's theory but it's an interesting read.

Apple Silicon: The Passing of Wintel

by Jean-Louis Gassée

mondaynote.com

 

[KPOM]

Very interesting article. If he’s correct, then Microsoft is rooting for Apple’s success with their own silicon, and should be right there with native MacOS software from the start.

 

[Joelist]

There is a problem with his thesis - Apple Silicon is FAR more powerful than other "ARM". In fact all it has in common with other ARM is using the basic instruction set and even there they have added a lot of their own instructions. Apple can do this because of how powerful their SOCs are.

 

[burgerrecords]

Gap is closing - check out Cortex X1. I'm going to get an arm mac too but there are a lot of folks that will be gunning for Apple Silicon in the likely event that arm mac is awesome.

B

 

[hawkeye_a]

The article describes a likely chain of events to be expected, with the industry transitioning to ARM.

What if Intel/AMD stick with the x86 architecture and somehow make it competitive(TDP-wise) with comparable offerings on other(ARM) architectures? Unlikely i suppose.

What does this transition away from a de-facto industry "standard" to everyone designing their own processors and APIs mean for ..... software "standards" across manufacturers with theoretically different silicon and possibly different instruction-sets or "extensions" of those instruction sets? Developers primarily catering to the "least common denominator"(most widely supported instruction-set/features)?

It's going to get messy.

 

[alesales]

Before Apple’s keynote on June 22nd there was already some “movement” in ARM computing (not mobile), mostly on Raspberry PI (which in the past years created a demand of multiple Linux distros supporting arm64 architecture and, with Raspberry PI 4 created a first low end desktop replacement), on AWS servers (their 2nd generation ARM servers have terrific performance compared to Intel, and yet they’re beneficing from Linux ecosystem mostly created by Raspberry PI) and some shy attempts made by Microsoft with their Surfaces (which, in my opinion was mostly like throwing a stone and check what will happen).

After Apple’s announcement, considering the market recognition of Apple as one of the leader in desktop computing, I started to get interested in Windows 10 for ARM, mostly for checking what to expect from that version of Windows and have my own opinion on all the speculation about Windows on Apple Silicon. Based on what I found out (no hands-on experience yet, unfortunately) seems that 80/90% of work is done from OS side, with some notable exceptions:

• There is an implemented Intel x86 emulator (like Rosetta 2 on MacOS) but it supports only 32bit apps. Benchmarks said that the performance penalty is around 20/30% (in line with what Rosetta 1 was doing on Intel Macs) but it support caching, so once a piece of code has been translated once, next time the emulator won’t need to reconvert the software but will run the converted version – as far as I got Rosetta 2 will do something similar. At the moment there is not an x86_64 emulator (for 64 bits application) although is in roadmap.
• At the moment Microsoft did failed in creating a real ecosystem of applications natively running on ARM Architecture. Just an example, only this month Visual Studio Code has been ported to native ARM Architecture. Office for Windows is still only on x86 and x86_64 and I have no information about any notable software running natively on ARM. That’s I think the most blocking issue as Microsoft has no interest in making investment in a platform that has a very low audience.
• Not strictly technical and quickly fixable is the Windows 10 for ARM license, which at the moment is not generally available to public but it’s sold only to OEM partners.

Considering that most of the technical work is already in place, and having Apple going on the ARM direction, my speculation is that Microsoft will be for sure interested to “ride the wave” using the gained popularity of ARM to encourage the ecosystem in compiling Windows apps on ARM/

Nevertheless, I am not sure how much it will be relevant to people that at the moment are using Windows on Mac, except for the ones who are buying a Mac just because they believe that the hardware is much better than other PC brands and later spending most of their time in Bootcamp for using Windows-only applications.

I must say that in 2005 I was very excited when Apple announced the switch from PPC to Intel because it was opening a full range of new possibility both for myself (Mac User since I was big enough to handle a mouse) and for all the switchers which saw in Bootcamp and “insurance” in case after they bought their first Mac they wont be able to use the OS or find their favorite apps on Mac (at least that was happening in Europe, in US I think that Mac ecosystem was much more solid already). In my personal experience I although see that in the past years I reduced to almost zero the utilization of my Windows 10 VM, as I don’t have any software that I am using (or I am asked to use) which is Windows only.

 

[Boil]

...Jean-Louis Gassee, a name you may remember as one of the execs in the Apple low years of the 80s...

Ah, the BeBox (das blinken lights) & Jean-Louis' dream of BeOS replacing the Classic Mac OS...!

 

[jdb8167]

There is a problem with his thesis - Apple Silicon is FAR more powerful than other "ARM". In fact all it has in common with other ARM is using the basic instruction set and even there they have added a lot of their own instructions. Apple can do this because of how powerful their SOCs are.

ARM Holdings is getting religion when it comes to higher end CPU designs. They have the Neoverse line for server/machine room design and the Cortex-X1 (Cortex-X1 Micro-architecture). Qualcomm is now licensing ARM's CPU designs so these architectures are probably coming to a laptop/desktop sometime soon if Microsoft has real hardware ambitions.