| > just pointing out the CPU isn't some standard design with nothing to open up simply because it licenses the ARM ISA The implementation is not an ARM reference design; but it's Apple's own implementation. Why should they be forced to open it up? How are they limiting their competitors? > they are nearly 2x ahead of Qualcomm using the same ISA at the same power level in 3rd party userspace computation only benchmarks I'm not sure whether you're speaking about M1 or A14 here. A14 is just a bit faster than other mobile SOCs, certainly not 2x. If you're speaking about M1... well, what are you comparing it against from Qualcomm? Do they have a similar product, with a similar target and power envelope? The recently announced Snapdragon 8cx Gen 2 from Qualcomm is around 6-7W TDP, which is 50% that of M1. > There isn't any vertical integration in those, the raw performance is just better. I'm unconvinced. We need to consider the 5nm production process, the everything-on-soc approach, and the fact that Intel is selling more-or-less the same CPUs as 5 years ago. The Ryzen 7 4800U is quite close to the M1 both performance wise and TDP wise. Now, take a 4800U, make it 5nm, build a SoC with it which includes the memory and build a finely tuned OS - aka "vertical integration". Would you really think you could appreciate a performance or thermal difference? |
I was referring to the M1 at hand but actually the A14 is arguably in the 2x range as well, certainly not "barely ahead" as much as I love my Android with a Snapdragon. Qualcomm has been in this "ARM laptop" space for lack of a better term with things like the 8cx and now 8cx gen 2. While these Qualcomm laptop devices aren't completely blown out in multi-core (though it is a significant loss) Qualcomm's single thread performance isn't really that much better than the mobile chips it's based of off. Which to be honest is also true of the M1, it's not ridiculously better than the A14 by any means... the A14 just already trounced the 8cx even though it's for a lower segment. Hell I'd rather have an A11 from 2017 in my brand new Android phone.
I'm not really going to speculate with you on "what ifs" when we have the actual quantitative data from 3rd party benchmarks saying general purpose code ported by 3rd parties has nearly the same performance uplift as Apple got with its own software. Actual data tells more story than I could ever spin from personal theory.
As for the future Zen 4 going to the same TSMC 5nm process isn't going to close the wattage gap by any means but we'll have to see if AMD can make larger overall gains in single thread performance. Apples challenge with the M1X/M2 is going to be "does is scale to more cores" to compete with the likes of the *950X or Threadrippers as they move towards workstation hardware but that's a more tried and true challenge than trying to get x86 cores to run faster on less power. I don't want to count AMD out and I'm actually typing this from a 3900X with a 5950X on order but I can tell you right now it's noticeably faster to e.g. browse the web on my Fiancee's A14 Ipad than my liquid cooled 3900X with a 3090 and as Chrome has shown that's not because of magical software integrations because it's all got the same company name attached it's because the CPUs run the workload better.