By: Ungo (a.delete@this.b.c.d.e), August 6, 2014 5:34 pm
Room: Moderated Discussions
anon (a.delete@this.b.c) on August 6, 2014 2:48 pm wrote:
> A8 is rumored to run at 2GHz. A12, if not A10 will probably have a CPU
> core running @2.5/3GHz or close by. iOS volume will help the CPU core.
Who cares about rumors, and on what basis do you claim they're "probably" going to be at 2.5 to 3 GHz by A10?
There are real examples of 2Xnm CPU cores with IPC similar to Apple's Cyclone and clock rates of 3+ GHz, Intel's Ivy Bridge and Haswell. Even when Ivy/Haswell clocks are restricted to about the same frequency as Cyclone, they consume a lot more power. Desktop/laptop performance potential requires tradeoffs unacceptable in low power mobile applications. iOS volume does not change this.
> Given all the on-package mem initiatives, (HBM/HMC/WIO) ongoing, Apple will have a
> high-perf uncore and on package memory soon any way. iOS volume will help here.
HBM/HMC are unsuitable for mobile, and you're acting as if this is a new thing for mobile SoCs. TSV interconnect is a new development, but package-on-package memory has been shipping for ages. WIO is actually designed for mobile, but it isn't likely to boost Apple's (or anybody else's) mobile chips up to laptop/desktop performance class for the forseeable future.
> It doesn't make sense in 2014. But in 2016? 2018? There, it looks to be a toss
> up to me. If it's a toss up, I think Apple will bring mac chips inhouse.
The reasoning you've given in support is nonsense. I don't claim to know that Apple won't do it, but if they do, I don't think it'll happen the way you're predicting.
As a footnote, like others I'm skeptical even of the proposed scenarios where they clock a derivative of a future Ax chip a little bit higher, call it good enough for the MacBook Air, and keep the rest of the Mac product line on Intel. Despite Apple being well prepared to support multiple CPU architectures, this would still create a lot of headaches for them, their customers, and third party software vendors.
Historically the Mac has changed CPU architectures twice. In both cases it worked because Apple committed and moved the whole line over quickly (leaving no doubt about the future of the platform), and the new systems were so much faster that they could emulate the old CPUs at a reasonable speed. That was very important: Apple was able to sell good products on day 1 even without a big native software base. The "put iPad guts into the MBA" concept is the opposite of that proven formula. So are most of the other ARM Mac scenarios that have been popular to speculate about the last few years.
> A8 is rumored to run at 2GHz. A12, if not A10 will probably have a CPU
> core running @2.5/3GHz or close by. iOS volume will help the CPU core.
Who cares about rumors, and on what basis do you claim they're "probably" going to be at 2.5 to 3 GHz by A10?
There are real examples of 2Xnm CPU cores with IPC similar to Apple's Cyclone and clock rates of 3+ GHz, Intel's Ivy Bridge and Haswell. Even when Ivy/Haswell clocks are restricted to about the same frequency as Cyclone, they consume a lot more power. Desktop/laptop performance potential requires tradeoffs unacceptable in low power mobile applications. iOS volume does not change this.
> Given all the on-package mem initiatives, (HBM/HMC/WIO) ongoing, Apple will have a
> high-perf uncore and on package memory soon any way. iOS volume will help here.
HBM/HMC are unsuitable for mobile, and you're acting as if this is a new thing for mobile SoCs. TSV interconnect is a new development, but package-on-package memory has been shipping for ages. WIO is actually designed for mobile, but it isn't likely to boost Apple's (or anybody else's) mobile chips up to laptop/desktop performance class for the forseeable future.
> It doesn't make sense in 2014. But in 2016? 2018? There, it looks to be a toss
> up to me. If it's a toss up, I think Apple will bring mac chips inhouse.
The reasoning you've given in support is nonsense. I don't claim to know that Apple won't do it, but if they do, I don't think it'll happen the way you're predicting.
As a footnote, like others I'm skeptical even of the proposed scenarios where they clock a derivative of a future Ax chip a little bit higher, call it good enough for the MacBook Air, and keep the rest of the Mac product line on Intel. Despite Apple being well prepared to support multiple CPU architectures, this would still create a lot of headaches for them, their customers, and third party software vendors.
Historically the Mac has changed CPU architectures twice. In both cases it worked because Apple committed and moved the whole line over quickly (leaving no doubt about the future of the platform), and the new systems were so much faster that they could emulate the old CPUs at a reasonable speed. That was very important: Apple was able to sell good products on day 1 even without a big native software base. The "put iPad guts into the MBA" concept is the opposite of that proven formula. So are most of the other ARM Mac scenarios that have been popular to speculate about the last few years.
