Ask HN: Why hasn't x86 caught up with Apple M series?

There’s a number of reasons, all of which in concert create the appearance of a performance gap between the two:

* Apple has had decades optimizing its software and hardware stacks to the demands of its majority users, whereas Intel and AMD have to optimize for a much broader scope of use cases.

* Apple was willing to throw out legacy support on a regular basis. Intel and AMD, by comparison, are still expected to run code written for DOS or specific extensions in major Enterprises, which adds to complexity and cost

* The “standard” of x86 (and demand for newly-bolted-on extensions) means effort into optimizations for efficiency or performance meet diminishing returns fairly quickly. The maturity of the platform also means the “easy” gains are long gone/already done, and so it’s a matter of edge cases and smaller tweaks rather than comprehensive redesigns.

* Software in x86 world is not optimized, broadly, because it doesn’t have to be. The demoscene shows what can be achieved in tight performance envelopes, but software companies have never had reason to optimize code or performance when next year has always promised more cores or more GHz.

It boils down to comparing two different products and asking why they can’t be the same. Apple’s hardware is purpose-built for its userbase, operating systems, and software; x86 is not, and never has been. Those of us who remember the 80s and 90s of SPARC/POWER/Itanium/etc recall that specialty designs often performed better than generalist ones in their specialties, but lacked compatibility as a result.

The Apple ARM vs Intel/AMD x86 is the same thing.

Fair enough, but Apple Silicon is not a specialist chip in the way a SPARC chip was. It's a general purpose SoC & SiP stack. There is nothing stopping Intel being able to invest in SoC & SiP and being able to maintain backward compatibility while providing much better power/performance for a mobile (including laptop and tablet), product strategy.

They could also just sit down with Microsoft and say "Right, we're going to go in an entirely different direction, and provide you with something absolutely mind-blowing, but we're going to have to do software emulation for backward compatibility and that will suck for a while until things get recompiled, or it'll suck forever if they never do".

Apple did this twice in the last 20 years - once on the move from PowerPC chips to Intel, and again from Intel to Apple Silicon.

If Microsoft and enough large OEMs (Dell, etc.), thought there was enough juice in the new proposed architecture to cause a major redevelopment of everything from mobile to data centre level compute, they'd line right up, because they know that if you can significantly reduce the amount of power consumption while smashing benchmarks, there are going to long, long wait times for that hardware and software, and its pay day for everyone.

We now know so much more about processor design, instruction set and compiler design than we did when the x86 was shaping up, it seems obvious to me that:

1. RISC is a proven entity worth investing in

2. SoC & SiP is a proven entity worth investing in

3. Customers love better power/performance curves at every level from the device in their pocket to the racks in data centres

4. Intel is in real trouble if they are seriously considering the US government owning actual equity, albeit proposed as non-voting, non-controlling

Intel can keep the x86 line around if they want, but their R&D needs to be chasing where the market is heading - and fast - while bringing the rest of the chain along with them.