They've reinvented x86 at least twice: microcode and then x86_64. I wouldn't bury x86 yet.
Plus a good chunk of the M1 computing prowess is not even the CPU, it's the additional hardware such as the neural engine cores and such. This is something you could put next to a Sparc core, it's not ARM specific.
There is an fundamental complexity in the semantics of x86_64 instructions that makes them expensive to implement. I worked for Transmeta and am at another microprocessor company today so I do have 1st hand experience.
There are other things in M1 that helps make M1-based systems fast. I'm not talking about that. I'm exclusively referring to pure-core performance. It is true that the latency improvement and high-bandwidth of the nearby DRAM does help here, but it is far from the major factor.
FWIW, there's no evidence that TSMC's N5 node is the major factor, beyond letting Apple spend a lot of transistors.
Intel/AMD have so far been able to overcome the overhead with lots transistors and power, afforded by leading edge nodes. Apple is now at the same playing field and especially the power efficiency difference is obvious.
Nah, this is it for x86. The only relevant piece of x86 binary software is the Windows ecosystem ... and that runs on arm now. Everything else is either interpreted (Python, Ruby), JIT'ed (Java, C#) or can be compiled natively to arm anyway. All Apple have done is prove that it's possible.
However I'm doubtful that the PC market will move to a closed system similar to Apple's. So whoever comes in with ARM better be prepared for lots of competitions and low margins in a cutthroat environment. That's not a very attractive market proposition, especially since the x86 makers won't just take this lying down.
I think we'll be in a weird place soon.
Apple will be doing Apple things and increasing its market share but never taking over (maybe it will reach 20-25%? who knows).
Many server environments will switch to ARM, especially the big cloud providers that can make their own chips.
Qualcomm will push some more desktop class SOCs but I doubt they'll reach Apple's performance levels so they risk having their lunch eaten, especially by AMD.
So it's possible that the PC desktop/laptop market might be "stuck" with x86 for many more decades.
There is also a scenario where Intel or AMD come up with some brilliant designs and catch up completely or well enough to Apple in which case this will all have been a tempest in a teacup.
They've reinvented x86 at least twice: microcode and then x86_64. I wouldn't bury x86 yet.
Plus a good chunk of the M1 computing prowess is not even the CPU, it's the additional hardware such as the neural engine cores and such. This is something you could put next to a Sparc core, it's not ARM specific.