If the trend of this chart continues, 86x processors will start to fall behind more and more every year. At some point the gap will be so big people won't buy 86x PC's anymore. The industry has to keep up. I'm sure the big players are already testing out ARM processors for PC's.
If all ARM SoCs/CPUs were created similarly and equally, I'd agree that this would threaten x86 pretty directly. However, Apple seems to be the only one designing SoCs this way, and they're only doing it for Apple and needs that specifically benefit the execution of macOS and the Mac platform. I think ARM64 versions of Windows and Linux will run well on Apple Silicon, but making that so is not Apple's primary objective (just as it never was with Intel Macs and x86 and x86-64 versions of Windows and Linux). But I think we're not going to see another chipmaker that isn't solely making chips for Apple (let alone Apple itself) producing ARM64 CPUs that are optimized as well for Windows 10 for ARM64 or other ARM64 distros. ARM is still relatively new in the desktop computing space (having mainly been for smartphones and tablets); Apple has one hell of an advantage in that Mac OS X/OS X/macOS has always been architecture independent, and in iOS and iPadOS (being mostly the same OS below the UI layers); they've pretty much had a thirteen year head-start to this point.
Maybe some enterprising folks will build a massive compute clusters with the Mx Macs in the future, seeing that it isvery power efficient compared to what's on the market.
You're still going to be limited in that M-series SoCs are optimized for Macs first and foremost. Apple doesn't care about the datacenter the way that Intel, AMD, Qualcomm, and Microsoft all do.
With Docker coming on board and more appliances are released as native ARM binaries, it may turn into a data center monster as well, if Apple ever want to enter that piece of the market.
It's possible. Though, I'm not sure the M-series scales as well as some of the other ARM SoCs out there that have 128 cores and insanity like that.
I hope this move results in a "Industry game changer", just to get rid of crappy x86. But that is probably just wishful thinking from me.
But it is a game changer for Apple, it is good to them to make their products unique and not just "another PC in a fancy enclosure". Now I know they have always had macOS to diffrentiate and that for me was enough reason to go with Intel macs. But it is much more interesting when the entire computer actually is different.
It is also a good thing that macs ran on x86 prior this. Imagen a alternate reality where apple stuck with PPC (and it could keep up with intel performance wise) and the switch happened now in 2020 from PPC to ARM. My guess is that it wouldn't be as exciting because there would be much less software and apples market share would probably be much smaller than what it is in our current timeline. Now, much thanks to the x86 years, mac has a much wider software catalogue and bigger user base and this means that ARM will get much better software support.
The move to Intel was huge in that it gave more developers a reason to port to x86 Mac OS X. I think Mac gaming got to the point where it was bigger than ever circa 2010-12 as a result. But I fear that may shrink down. Companies like Aspyr seem to be more focused on Linux ports these days (and with all of the hell that Apple has put developers through over the last few years, I really can't say I blame them at all). Luckily, Feral Interactive still seems to care. But the loss of games to Catalina ditching 32-bit Intel binary support was huge. And I'm sure that games weren't the only pieces of Mac software affected in this way.
I think you're right that the switch to Intel brought more developers into the Mac fold. I think the switch to Apple Silicon is going to be true test of loyalty to the Mac app ecosystem.
I cannot understand this complaint. OpenGL/OpenCL work perfectly fine and will probably continue to do so for the foreseeable future. They are implemented as a wrapper around Metal on the Apple Silicon and Apple can continue to ship them at no cost. Deprecation simply means that one should not use that stuff anymore in new applications.
Deprecation, especially with Apple (and Microsoft nowadays too) usually means the actual axe is coming. Apple isn't usually one to leave deprecated technologies around in macOS all that long. So, I definitely understand the concern of that deprecation.
ARM ISA will finally begin to takeover from x86 ISA.
Cloud offerings
Server side
Linux
ML / scientific calculation
Video production
These will all flock to the speed and efficiency of these new chips. Energy efficiency (and heat dissipation) are key drivers for large server data centers.
Will they change overnight? Will all be Mac offerings? Will Intel disappear? The answer is NO in all counts.
But the wave is coming and a new era of computing is finally upon us!
You need more people attempting to do what Apple has done for this to actually take off. It's like Apple came up with the cure to COVID-19, but only inoculated themselves with it. Puts them in a really great spot, but it ONLY puts them in a really great spot. That's not to say that x86 isn't in a jam. I think the fact that TSMC is also manufacturing AMD's chips puts AMD in a good spot to keep innovating x86. But Intel is in a bad bind (due not in small part to their poor upper management) and they're a key player. But on the whole, for ARM to really take off, more people need to be building with, innovating, and employing it. Though, I would agree that this does seem inevitable right now.
AMD not being included doesn’t paint an accurate picture of the future of x86, as AMD clearly is executing better than Intel right now. Apple isn’t selling their CPU core designs either, so it may be awhile until we start seeing other ARM designs catch up and also challenge x86 in the same way.
Legacy tooling and software will likely pull folks to AMD for a while in the desktop/laptop market. At least until Microsoft gets serious about Windows on ARM, and some manufacturer other than Apple gets serious about ARM on the desktop enough to deliver good performance there.
I definitely agree with this. The fact that there isn't a desktop variant (and only a server-focused edition) of Ubuntu for ARM64 is telling. Because, otherwise, Apple would've demoed that at WWDC instead of Debian (which, while the basis of Ubuntu is not as popular by comparison). ARM for desktop is still bleeding edge. Apple has just had more time to get comfortable with the idea of moving their own desktop platform there.
This certainly differentiates the Mac, and Apple undoubtedly sees the barriers that other PC companies have in replicating what they have done as an advantage. That said, M1 certainly demonstrates that there is a viable alternative to x86. Microsoft is probably in the best position to vertically integrate, since they write the OS and now design hardware. They could introduce reference designs, and work with a company like Qualcomm to design an SOC around a future update to Windows.
They did that with the SQ1, the SoC that went into the first Surface Pro X, and the SQ2, the SoC that went into the second Surface Pro X. It performed native OS functions and apps well enough; the latter doing a bit of a better job than the former. But it's a far cry from Big Sur on M1.
/article-new/2020/11/perf-trajectory_intel-apple-axx-anandtech.jpg)
