The next generation Ryzen series (7000 series) looks like it's going to be absolutely insane. That window may end up being shorter then expected.
The next generation Ryzen series (7000 series) looks like it's going to be absolutely insane. That window may end up being shorter then expected.
Note that their portable / consumer devices have always been ARM. Like, back to the Newton even.
Apple now makes their own chips. They've dealt with that vulnerability. And they're fully onboard the ARM bus in a big way and they've managed to bring their customer base with them without a lot of hassle.
Over the last 40 years Apple has been burned multiple times with having their key ISA be dropped by its manufacturer. 6502 became a deadend (quickly.) Motorola dropped the ball on the (relatively successful) 68k, and tried to get people onto the ill-fated 88k and then pulled the eject cord and everyone went PowerPC. That lasted maybe a decade and Motorola/Freescale&IBM started to flee the ship. I think Apple has learned their lesson. They don't want to be vulnerable on this front anymore.
Apple's systems may very well end up being slower. From a business standpoint, I am not sure that will matter.
Neither AMD nor Intel have shown much indication of catching up to Apple's (well, ARM's) mobile device battery life advantage. Apple doesn't sell systems that go in data centers, they don't need to smoke anything.
When Apple was using PowerPC the number of PowerPC units shipped was tiny compared to the PC market.
The huge volumes of mobile devices that Apple now ship has got to at least help share R&D costs across product lines.
Consumers already know this. They'll buy the new Mac even if it's worse (see: 2016), the new iPhone even if it's slower (see: 2022), and the new Magic Mouse even if it's still broken. It doesn't matter how fast or slow the Mac is, Apple has long proven that they exist in a market segment of their own, for better or worse.
Essentially this was because PC volumes supported Intels R&D and investment being greater than anyone else could afford. We’re no longer in 2000 though. Apple sold 240 million iPhones last year which is not far off global PC sales. Meanwhile TSMC has overtaken Intel with a broader and growing product base whilst Intel is shrinking.
So Apple bought themselves a win, which makes perfect sense. And they had already chosen ARM for iOS (for historical reasons, really -- the original iPhone picked an off-the-shelf Samsung SOC), so that's what they dropped in the Mac.
But indeed, this is probably transitory. Eventually TSMC will either lose its lead or ramp its best processes to the point where other manufacturers can buy it. Already, as we see here, they're only mildly ahead of the rest of the pack.
But I do think there's a certain amount of kool aid being drunk by the Mac community about these chips. They're great. But maybe not insanely so.
[1] It's absolutely true that Apple's whole-system power integration story remains at the top of the industry. But that was true even when they weren't on the best process.
Sorry you are totally right and my phrasing was poor, what I was trying to say is that Apple's TSMC 7nm chips (A13) are still very competitive with Qualcomm and AMDs offerings on the same node. They beat out AMD (5000 series) on perf/w while tying perf and beat Qualcomm (865+) on perf handily with (iirc) similar wattage. M1 and M2 are certainly benefitting from TSMC 5nm compared to everyone else currently and AMD in particular is a lot closer than I think many Apple fans give credit.
A low power ARM MCU will kill anything in perf/watt. Performance gains aren't linear.
That meant that the performance requirements of a laptop were simply never going to be IBM or Motorola’s priority. The last PowerPC Mac was the G5 tower, and that was an amazing case with incredible cooling setup because it needed it. The G5 used a ton of power, and generated insane amounts of heat (seriously those G5 heat sinks were larger than most pc gamer/modder heat sinks if there era, before considering the enormous airflow, and even a flow shroud. Simply put there was no way a G5 was ever going to be in a laptop that was actually portable.
That’s what led to Apple migrating to intel - intel at least was interested in consumer hardware, and portable computing as well.
So the important thing is that Apple’s own cpus are presumably always going to be aligned with what Apple wants, in a way that they can’t rely on others to do.
The only game in town that Intel/AMD can really play now is to make their chips cheaper and more palatable (reduce power consumption, offer better integrated GPUs, etc) for x86 based consumer products or slowly see Apple's share in the laptop market increase more and more.