“At today’s Intel Developer Forum, Intel CEO Brian Krzanich and other Intel execs showed off its 7th Generation Core, Kaby Lake,” Juli Clover reports for MacRumors. “For those unfamiliar with Kaby Lake, it is the third chip manufactured using Intel’s 14-nanometer process, following Broadwell and Skylake. It’s a semi-tock with optimized microarchitecture, offering support for Thunderbolt 3, native USB 3.1, and DisplayPort 1.2.”
“According to Krzanich, Kaby Lake processors are already shipping to Intel’s manufacturing partners and will launch in new devices coming this fall,” Clover reports. “Krzanich did not provide a further breakdown on when chips appropriate for some of Apple’s machines long overdue for updates will launch.”
“Intel often launches low-power 4.5W Y-series chips and 15W U-series chips first, neither of which are suitable for use in the machine that people are most curious about, the MacBook Pro,” Clover reports. “According to an old Intel roadmap, Kaby Lake chips appropriate for use in the MacBook Pro, the iMac, and the Mac mini won’t launch until the very end of 2016 or the beginning of 2017, meaning any Apple machines released in the fall of 2016 may be limited to Skylake chips.”
Read more in the full article here.
MacDailyNews Take: Same old, same old. Waiting for Intel.
I’ve always wanted to own and control the primary technology in everything we do. – Steve Jobs, October 12, 2004
[Apple’s] reason for being is the same as it’s always been — to make the world’s best products that really enrich people’s lives. That hasn’t changed. And we do that through owning the primary technologies. — Apple CEO Tim Cook, August 9, 2016
As we wrote back in January 2015:
There is no reason why Apple could not offer both A-series-powered Macs and Intel-based Macs. The two are not mutually exclusive…
Mac sales to grow in enterprise with new Apple A-series-powered Mac – October 14, 2015
Apple A-series-powered Macs are not only feasible, they may be inevitable – January 15, 2015
Why Apple dumping Intel processors would be disastrous – January 14, 2015
KGI: Apple is designing its own processors for Mac – January 14, 2015
Apple A9-powered MacBook Air? – December 16, 2014
Why Apple will switch to ARM-based Apple A-series-powered Macs – August 27, 2014
Intel-powered Macs: The end is nigh – August 4, 2014
Intel’s Broadwell chips further delayed; not shipping for most Macs until early-mid 2015 – July 9, 2014
Apple will inevitably drop Intel for their own A-series processors in the Mac – June 26, 2014
How long before Apple dumps Intel from MacBook Air? – June 26, 2013
I’m sure there will still be folks complaining “Why won’t Apple release Kaby Lake before next year?! UGH!!”
I said this over a year ago on this site… Anyone who thinks Apple will ship Macs based upon Kaby Lake processors in 2016 is nuts. Kaby Lake is NOT shipping in quantity. It’s barely past the prototype stage for any of its models. It will be sampling for *some* models in the next month or two. Full production won’t happen for *any* model until November or December (just so Intel can claim they high full production for Kaby Lake in 2016).
Yes, some vendors will announce Kaby Lake laptops (or laptop convertibles) this year with Kaby Lake, but they will not ship those in the tens of thousands per week. Not this calendar year.
The problem is that Apple *COULD* have refreshed the entire set of non Mac Pro lines with Skylake earlier this year. Apple could have even supported TB3 and USB 3.1 with stand alone chips. Then early next year updated the entire line with Kaby Lake with the CPU integrated capabilities. Many, many Mac users would have been happy with that.
Now Apple has the dilemma of either shipping Skylake systems before the end of this calendar year that they *could* have shipped four or more months ago (and be immediately declared as behind the times), or they can wait until 2017 to refresh the line base upon Kaby Lake and take the *very* real risk that many pro users are going to jump ship before then.
Yes. You did say this, and it bummed me out then, and it still does.
I’d say no point in shipping Skylake. All they have to do is break a tiny little bit of silence and say “We’re waiting for Kaby Lake” and most people will be satisfied. Lots of people will wait for Kaby Lake anyway. But they will be silent, and people will bitch and moan.
And that last sentence seems to be far too common of late, ie wait for a better solution, but end up waiting so long that their products ship for long periods with original more inferior solutions, and when they do launch its not far short of an opportunity for another update anyway had they gone with the less advanced chip option. Well that’s if they didn’t aim to try to save money by only updating well past the date when it is appropriate anyway, so exacerbating the timing problem further. The loss of sales by waiting as a result must surely put the logic of that policy into question.
Intel processors are two generations ahead of what ships in the MacBook Pro. You can’t blame Intel.
Apple didn’t ship with Broadwell.
Apple didn’t ship with Skylake.
When Apple doesn’t ship with Kaby Lake, you’ll be complaining that it is because Intel is late with Cannonlake.
Intel is progressing. Apple and Tim Cook are standing still.
The 2015 13″ MacBook Pro has a broad well chip, the 2015 iMac has a sky lake chip, the MacBook has a sky lake chip, the Mac mini and 15″ MacBook Pro are the only ones using a Haswell chip. The Mac Pro used the most current Xeons until this past May…. You’re very wrong.
Well, I’m referring to the 15″ MacBook Pro (probably because that’s what I have).
Regarding that, I am very correct.
You are correct about the 13″ MacBook Pro.
The Mac Mini is not only two generations behind; it uses a crippled processor (two cores only). Again, you can’t blame that on Intel. Blame Tim Cook.
This is why I’ve been so patient with the Mac refreshes. Intel is dropping the ball on processors, there is no reason they couldn’t give Apple early access to the Kaby lake chips for MacBook pros, like they did with the core 2 chips on the original MacBook Air. But with Apple now being one more company relying on intel’s roadmap and production, they are hamstrung what they can get from them. As MDN says A-series macs and Intel macs could easily ℅ exist, and I don’t think that’s the worst idea especially if the A10/A10x chips are markedly faster than their predecessors. Now if the Kaby lake chips don’t provide the generational improvement over sky lake? Then going with sky lake is fine for performance, and making some battery optimizations, storage improvements, and Polaris 11 graphics would make up for that in many ways. They may have to build a custom thunderbolt 3 board, but they’ll do it. Goddamned Intel.
I give them slack, too. I have the money to buy (been saving money now for years LOL!) and I have the desire for a new system (haven’t bought since 2013) but since I’ve already waited this long, I’ve been waiting for the “next” thing. Looks like the next thing won’t be coming until next year at the earliest, so I’m content to continue to wait until Intel can actually make something that Apple wants to use. OH, and if the chips next year are all dual core until late in the year, then we’re still not going to see an update. 🙁
Intel is NOT “dropping the ball on processors”!
Intel has announced its roadmaps for processors 12 – 18 months in advance. Yes, there was a significant slip a couple years back. But since that time Intel has been very up front with their plans and have matched those plans within a couple months for the last couple years.
Ax series Mac will not do what Intel Macs will do. A lot of the Mac users (even pro users) depend upon software written for Intel and/or Windows/Linux. If you want to run that on an A10/A10X architecture you’ll have to do software translation at the lowest level. That is a huge hit. The only reason why Apple could get away with it for the 68000->PowerPC jump was the PowerPC was *so* much faster than the 68000 series chips of the day. It barely worked (but it did work) for they PowerPC->Intel shift because the chips being offered by IBM had pretty much hit a wall. No one in their right mind is suggesting Intel has hit a wall with their CPUs. Are you seriously suggesting Apple move to Ax series chips which are slower than Intel’s current chips for the vast majority of what Macs do then take a translation hit on top of that?
Kaby Lake is the “Optimization” stage of what is now a three stage process (is: shrink, new architecture, optimize — was: shrink, new architecture). Kaby Lake’s optimization includes things like updated PCIe lanes, TB3 onboard, USB 3.1 onboard, better internal graphics engine, Optane storage, etc. There are reasons to want to move to Kaby Lake.
The Ibm chips didn’t “hit a wall”. Performance wise, they’re still well ahead of intel offerings. It was “performance per watt” since apple was focusing on mobile systems. Intel has dropped the ball on processors, since Rachel iterative generation hasn’t provided the performance increase that apple came to expect in the first 6 years of the Intel relationship. For example, my 2012 MBP (as I’ve stated previously) isn’t much slower than the current systems. And in raw compute, is maybe 6 seconds slower on large encodes. Now if the kaby lake chips are significantly faster? Then use them. (The Polaris 11 gpus will be much faster than what we currently have) but if they’re not? Then use the sky lake. No one else seems to give razer any shit when they use last gen tech in their “gaming notebooks”….
“there was a significant slip a couple years back”
You could have stopped right there as we’re in agreement. 🙂 Now, if we can agree that Intel has been REALLY late with a 4 processor mobile ready Skylake processor, we’ll still be in agreement.
Intel JUST announced that they will be producing ARM processors. I’m sure it’s because they know Apple’s serious about going the A9 way and they don’t want to be left out.
Sure, Intel really really needs to be in the 7% of PCs that happen to be Macs.
Let Apple put Ax CPUs in Macs and watch their market share drop by at least half. Perhaps Mac sales will even drop below Win phone sales? Now that would be spectacular.
Apple move on Intel is Motorola and IBM all over again.
ARM based macs are out of the question for now, that’s a fact we know.
How, you may ask? Well, Bitcode.
(Bitcode is the intermediate language of the LLVM compiler that enables Apps to automatically support any CPU architecture without changes. With that Apple can change the CPU architecture and all Bitcode-Apps would just magically work with native speed.)
If Apple wanted to change CPU architecture, it would first enable, then encourage, then mandate all new App Store submissions to use Bitcode, and then launch the new architecture after there’s enough Bitcode-Apps available. Without that they’d need emulation, which is impossible with current (low) performance difference between ARM and i64.
If Apple had any plans for ARM macs, Bitcode would be mandatory in the Mac App Store. But Mac App Store doesn’t have *any* support for Bitcode at all. So, there you have it: no ARM macs in the pipeline for the foreseeable future.
However, Apple supports and encourages the use of Bitcode in (iOS) App Store. They are closer to running iOS with i64 than running macOS with ARM (xBox killing game console version of AppleTV or iOS Apps on macOS?)
Rather than waiting for ARM based Mac, I’m hoping for an i64 CPU designed by Apple. I don’t have very high hopes, but that would explain waiting for so long without updates.
That’s a good argument, and would make sense if there wasn’t a history of Apple holding back compilers until the release of a “universal” binary support like during the Intel transition. Rosetta wasn’t even hinted at in code until it was announced with the 1st gen Intel iMacs and MacBook Pro. And I doubt that any developer who is trusted with that type of information would put any evidence of it in their systems. Apple recommends bitcode for arm devices because although they are different architectures (32 vs 64 bit) they are all still the same instruction set (risc) and it is required for older versions of software to run on older iOS devices that still use 32 bit arm devices. With the advent of iOS 10, and it dropping support for the A5, this won’t be necessary at the OS level anymore, which is another reasons that iOS 10 is smaller than previous releases, but is still required on apps that will support older hardware. On the Mac side, since i-64 chips are x86 cisc chips with the amd developed 64bit extensions, it makes little sense for Apple to tip it’s hand with any mention of a universal solution like Rosetta until they announce systems running on arm processors. And if they are able to get a hypothetical A10x “desktop” chip with better performance than any Intel offering, llvm support would be baked in to the version of macOS that ships with these machines. However, I doubt Apple would even hint at or announce such a transition before they actually do it so they can leapfrog the industry just like they did with 64bit mobile CPUs. Remember that PA-Semi had a PowerPC risc variant that was much faster than the core 2 series when Apple bought them, and there is no reason that team couldn’t have been developing both the arm processors for mobile and a true Intel replacement. Every time one of apple’s chip partners has gotten lazy, and not allowed them to move the way that they want, they drop them…. It’s happened 4 times. (Mos 6502- Motorola 68000, 68k-PowerPC, PowerPC-Intel, Samsung arm/strong arm-A series) and there weren’t any clues in any of the OS versions that hinted at these switches until Apple had publicly unveiled them.
That’s a good argument, and would make sense if Apple was not supporting Bitcode for iOS. Why they can give that hint for iOS but could not do the same with macOS too? There would be less of a hint if both had support.
In fact Apple told about Intel transition in WWDC, June 2005. First Intel Mac came out in January 2006, half a year later. It was not a surprise at that point.
Half a year even though as emulator Rosetta worked with existing PowerPC binaries “as is” — developers didn’t need to do very much for the transition so Apple got full app library support for free (well, for the cost of emulation.)
Back then Apple could afford emulation because the new architecture was so much more powerful. Not so this time — Intel to ARM is on par at best, so emulation is not an option. You need at least twice the power, and the power per watt is even bigger problem — emulating i64 on ARM would consume MORE, not less, than native i64. You don’t want an ARM based MacBook that’s both slower and hotter than an Intel machine.
So the only viable way to have comprehensive library of supported apps on day one (and Apple simply must have it) is Bitcode. Just adding Bitcode support in macOS would help only if the apps were compiled to Bitcode beforehand (as it is not an emulator.) Zero 3rd party apps are.
There are two ways to get 3rd party apps compiled to Bitcode: promote Bitcode years in advance or reveal the transition at least half a year early like with the Intel transition. But not on launch day. That’s simply too late when emulator is out of question (and as seen with Intel transition, even emulation was announced half a year early.)
Early announcement could kill sales, so that’s not desirable. That leaves promoting Bitcode as the best option. But that does not mean revealing a transition too early, or at all: Apple can require Bitcode for all its platforms and it still would not reveal a transition.
There are multiple other reasons to require it, like a possibility to reoptimize the apps when there are new extensions for instruction sets or security fixes in compilers etc. Apple can have good reasons for requiring Bitcode for all platforms forever without ever making a transition.
So, by requiring Bitcode for other reasons Apple could really make the transition as a real surprise, unlike the Intel one that they were forced to announce half a year early.
Bitcode support is not a hint of transition but as long as there is no Bitcode support, that’s a hint of no transition.
That’s a little bit circular but I get what you’re saying. I guess my response to that would be that I predicated my argument on the hypothetical A10x desktop chip being more powerful/faster than any Intel offering, providing similar benefits to how Rosetta worked. I don’t think Apple would transition to a less powerful chip design. And I don’t see a reason why the pa semi team couldn’t out design Intel since they did it 8 years ago already, I guess we’ll see but I think putting bitcode into the developer release would tip their hand since Apple doesn’t support 32 bit macs anymore of the same instruction set. That’s why it makes sense on iOS platforms and the watch. The watch effectively runs a shrunken and optimized A5 called the S1, it’s 32 bit, so having the llvm built in makes sense, also because the watch works with 32bit hardware in the A6 (iPhone 5/5c). And because there are still a ton of 32bit iOS devices that can run iOS 9, which is when at wwdc 2015 they started requiring it. When >90% of the install base is on 64bit hardware, they’ll probably drop support for it. I think we just are coming at this from different angles, and I do see your point, I think we’re just not agreeing.
They do need to have these ahead of everyone else. There’s no reason they shouldn’t unless they are developing their own processors.
Apple can upgrade the processor in the 13″ MBP but not the 15″ MBP. Sounds Apple just doesn’t give a crap to me.
One can look at the Mac Pro and tell that Apple doesn’t give a crap about Mac Pro users, or Mac Mini users, either.
In fact, Apple must hate Mac Pro users… Look at the POS they served up. Obsolete when it went on the market.