Intel Kaby Lake CPUs suitable for MacBook Pro refresh said to be in manufacturers' hands

2»

Comments

  • Reply 21 of 38
    wizard69wizard69 Posts: 13,377member
    macxpress said:
    I'd say there will be a different Mac announcement sometime in October. I'm thinking Sept 7th won't have anything Mac related except for maybe macOS. I can already hear the pissing and moaning about the lack of a new MacBook Pro/iMac. Tim must be fired for this!
    No Macs september 7 has already been "leaked" so no surprised there.   Given that I don't think we are far off from NEW MBPs.   I just hope Kaby Lake is far better than Skylake which took basically for ever to get debugged.   While people don't want to listen to this, the reason there is no SkyLake MBP's right now is that Intel took literally forever to get the hardware right.
    netmagepalominefastasleep
  • Reply 22 of 38
    wizard69wizard69 Posts: 13,377member
    zoetmb said:
    If Intel has just started 'sampling' of these CPU's to kit makers then it could be several months before the yield rates get good enough for them to start shipping in volume to the likes of Apple.
    If it is the start of volume shipments then we may see finished devices available before the end of the year which can't come soon enough for many of us.

    True. It's possible Apple may announce new MBP's Wednesday with a ship date of November. Apple shouldn't have issues getting first dibs on the chips. 
    Apple isn't the #1 manufacturer of computers, so I don't see why they would have priority over anyone else.   And I don't see how if chips are just getting delivered to manufacturers now and not necessarily in large quantities, how Apple releases a new MBP with those chips in November, even though some Window machines are apparently already hitting the market.   It's not like there's no other design needed around those chips and they're plug-and-play.    Remember, although Apple usually flies some initial shipments to the States, most machines come by boat from China.  That in itself can take six weeks, which means shipments from China have to start by October 15th.   That's only 40 days away.       

    Is there any past history where Apple received samples of chips in late August or early September and still delivered machines in quantity before the end of the year?

    I hope I'm totally wrong.  I'm still using a late-2008 MBP and need a new computer.   I don't want to have to wait another year.   But since I keep my machines so long, I also don't want to buy old tech.   So I would love to be proven wrong and be able to buy a new MBP with Kaby Lake in November. 
    This article is pretty clear and confirms what I've read elsewhere, manufactures started receiving volume shipments in July. Note volume shipments do not mean engineering samples which they would have had very early this year.

     Given that Intel is still at 14nm there should be little in the way of process start up. So once the have solid masks, production volumes can happen quickly after the start of ramp up. In other words they have a tweaked 14nm process that they should be able to get going quickly, this assumes of course that they don't have any SkyLake type problems with the more advanced chips.
  • Reply 23 of 38
    wizard69wizard69 Posts: 13,377member
    Soli said:
    If Intel has just started 'sampling' of these CPU's to kit makers then it could be several months before the yield rates get good enough for them to start shipping in volume to the likes of Apple.
    If it is the start of volume shipments then we may see finished devices available before the end of the year which can't come soon enough for many of us.

    True. It's possible Apple may announce new MBP's Wednesday with a ship date of November. Apple shouldn't have issues getting first dibs on the chips. 
    In the past, they've gotten chips first or special SKUs from Intel, but I think those were only for Mac Pros and iMacs, not their relatively high-volume notebooks.

    If Apple can't get Kaby lake for a shipping MBP this year, then I'd be happy with Skylake, assuming all the other noted changes to the MBP are also coming down the pipe.
    If Apple was happy with SkyLake we would seem MBP's with those chips in them already. It is pretty clear that Intel long drawn out release of these chips has put Apple in a bad position and has caused them to ditch SkyLake. At least it seems that way, one has to understand though that there may be other hold ups for the 15" MBP that Apple might be waiting on. This includes the AMD GPU expected in the machine.

    Speaking of AMD GPU's I'm still thinking that AMD and ZEN has a shot in the Mac Mini.   That would give the Mini a very good GPU and a good enough CPU for the market it is in.   I'd dearly love to see ZEN in an Apple product just to put Intel on notice as a little completion goes a long ways.   I'd like to see ARM in a Mini too, but that is a very long shot.
    netmage
  • Reply 24 of 38
    I would love to see a 10 series GPU in the new Pros, even if its just the 1060...

  • Reply 25 of 38
    mcdavemcdave Posts: 1,927member
    Soli said:
    mcdave said:
    Lets hope they include their own S10X too. Only a small (though disproportionately vocal) number of users require x86 compatibility. The majority of runtime can be handled by ARM with x86 spun up if required.
    1) S10X? You mean Apple's ARM designs with CPU and GPU cores and clock rate, and RAM that's suitable for a notebook or desktop? If so, would they include that with an Intel chip? I don't think that LITTLE.big has been very effective, and that's all ARM, and I also don't see how macOS can be built to run under ARM and then allow for dual, cross-architecture processing if and when you need an x86-based program. If you mean that it would allow emulation of x86 apps the way that created Rosetta when they made the jump to Intel from IBM, I don't think that's feasible since PPC had been steadily falling behind Intel, which precipitated the migration. If this happens, I think it will be at the low-end of Apple's "PC" product line, use the App Store as the primary source for apps, and have no emulation of x86 apps since large app suites from MS and Adobe aren't necessary out of the gate when talking about an entry-level machine that could start a couple hundred dollars lower than the current Mac notebook without sacrificing perceived performance, or per unit or per machine category profits.

    2) Doesn't S-series refer to the ARM chip in the Apple Watch, and A-series refer to the chips in their iDevices? Wouldn't an ARM-based chip from Apple for their "PC" line then have its own name; perhaps an M-series starting with M1?
    Yes, in my mind the S-series SiP would allow better component integration/performance/reliability.  I was also thinking that x86 apps could be run in a shadow VM or some kind of app sandbox similar to Rosetta but passing x86 code to the x86 logic for full speed execution.
    The S10X could be used to prolong battery life in the full MBP or for a non-Intel option once a viable ARM software base had been cross-compiled.
  • Reply 26 of 38
    Just a reminder, MacBook Pros are not a relatively high volume product. If small volumes of the chips are all that's available, its all Apple needs for a new MacBook Pro release.

    While Amazon is not the best place to glean data, only 2 of the 10 MacBooks in their top 50 best selling traditional laptop list would use this processor. If I try to stretch that information a little (ok, big stretch), maybe only 20% of the MacBook line gets this processor.

    If you want to play the guessing game even more, what about an iMac release with this line of processors now and new MBPs later.

    https://www.amazon.com/Best-Sellers-Electronics-Traditional-Laptop-Computers/zgbs/electronics/13896615011/ref=zg_bs_nav_e_4_565108#1
    edited September 2016
  • Reply 27 of 38
    SoliSoli Posts: 10,035member
    wizard69 said:
    Soli said:
    If Intel has just started 'sampling' of these CPU's to kit makers then it could be several months before the yield rates get good enough for them to start shipping in volume to the likes of Apple.
    If it is the start of volume shipments then we may see finished devices available before the end of the year which can't come soon enough for many of us.

    True. It's possible Apple may announce new MBP's Wednesday with a ship date of November. Apple shouldn't have issues getting first dibs on the chips. 
    In the past, they've gotten chips first or special SKUs from Intel, but I think those were only for Mac Pros and iMacs, not their relatively high-volume notebooks.

    If Apple can't get Kaby lake for a shipping MBP this year, then I'd be happy with Skylake, assuming all the other noted changes to the MBP are also coming down the pipe.
    If Apple was happy with SkyLake we would seem MBP's with those chips in them already. […] At least it seems that way, one has to understand though that there may be other hold ups for the 15" MBP that Apple might be waiting on. This includes the AMD GPU expected in the machine.
    Are you certain? As you mention, there could be other holdups, and 3rd-party discreet GPUs have been a frequent bane of Apple's MBP line for many, many generations after they've aged a year or two. Also, when was Skylake chips availabe for the MBP available in quantity for Apple's needs? From what I recall, those only came available earlier this year and in not enough yield to service what may be the Apple's largest mass selling of MBPs in a single weekend once they are released.
  • Reply 28 of 38
    SoliSoli Posts: 10,035member
    mcdave said:
    Soli said:
    mcdave said:
    Lets hope they include their own S10X too. Only a small (though disproportionately vocal) number of users require x86 compatibility. The majority of runtime can be handled by ARM with x86 spun up if required.
    1) S10X? You mean Apple's ARM designs with CPU and GPU cores and clock rate, and RAM that's suitable for a notebook or desktop? If so, would they include that with an Intel chip? I don't think that LITTLE.big has been very effective, and that's all ARM, and I also don't see how macOS can be built to run under ARM and then allow for dual, cross-architecture processing if and when you need an x86-based program. If you mean that it would allow emulation of x86 apps the way that created Rosetta when they made the jump to Intel from IBM, I don't think that's feasible since PPC had been steadily falling behind Intel, which precipitated the migration. If this happens, I think it will be at the low-end of Apple's "PC" product line, use the App Store as the primary source for apps, and have no emulation of x86 apps since large app suites from MS and Adobe aren't necessary out of the gate when talking about an entry-level machine that could start a couple hundred dollars lower than the current Mac notebook without sacrificing perceived performance, or per unit or per machine category profits.

    2) Doesn't S-series refer to the ARM chip in the Apple Watch, and A-series refer to the chips in their iDevices? Wouldn't an ARM-based chip from Apple for their "PC" line then have its own name; perhaps an M-series starting with M1?
    Yes, in my mind the S-series SiP would allow better component integration/performance/reliability.  I was also thinking that x86 apps could be run in a shadow VM or some kind of app sandbox similar to Rosetta but passing x86 code to the x86 logic for full speed execution.
    The S10X could be used to prolong battery life in the full MBP or for a non-Intel option once a viable ARM software base had been cross-compiled.
    It's 2016 and we still can't an iGPU and dGPU working that well together using OpenCL, so I'm doubting that we'll be able to have an full integrated and seamless handoff between Apple's ARM chips and Intel chips running on the same machine. There are just too many issues and not enough benefits for for me to see that as an advantageous solution.
    netmage
  • Reply 29 of 38
    bigpicsbigpics Posts: 1,397member
    entropys said:
    mcdave said:
    Lets hope they include their own S10X too. Only a small (though disproportionately vocal) number of users require x86 compatibility. The majority of runtime can be handled by ARM with x86 spun up if required.
    An idle thought: seeing as Apple can do excellent ARM processors, why not design out their own x86 chip for Intel or someone else to build? Got to be some reason for that massive R&D spending.
    Not so idle.  There's podcast chatter that Apple will be having Intel make all their ARM chips to remove their helping out Ssamsung, and with Intel making more than ever with Apple, Apple might then have AMD build them an X86 variant specially optimized to MacOS......  ...and they have a good working relationship with AMD as well.  
    netmagepalomine
  • Reply 30 of 38
    wood1208wood1208 Posts: 2,905member
    Apple's MAC hardware update schedule is dependent on the Intel's processor schedule. Initially Apple targeted redesigned Macbook Pro with Intel 10nm ICE lake in mind but Intel delayed it till next year. Moreover, redesigned Macbook pro update could not wait any longer (waiting for many years) so Apple-Intel worked out to use Kaby lake in-between until next refresh of Macbook pro in 2018 which perfectly lines up with Intel's ICE lake processor schedule..
    edited September 2016
  • Reply 31 of 38
    @"Apple isn't the #1 manufacturer of computers, so I don't see why they would have priority over anyone else."

    Actually, that might be a plus. Get the chips in the hands of a high-profile brand but not have to deliver so many during the production ramp-up stages that the manufacturing system will be totally overwhelmed. 
    bigpicspalomine
  • Reply 32 of 38
    mcdavemcdave Posts: 1,927member
    Soli said:
    mcdave said:
    Soli said:
    mcdave said:
    Lets hope they include their own S10X too. Only a small (though disproportionately vocal) number of users require x86 compatibility. The majority of runtime can be handled by ARM with x86 spun up if required.
    1) S10X? You mean Apple's ARM designs with CPU and GPU cores and clock rate, and RAM that's suitable for a notebook or desktop? If so, would they include that with an Intel chip? I don't think that LITTLE.big has been very effective, and that's all ARM, and I also don't see how macOS can be built to run under ARM and then allow for dual, cross-architecture processing if and when you need an x86-based program. If you mean that it would allow emulation of x86 apps the way that created Rosetta when they made the jump to Intel from IBM, I don't think that's feasible since PPC had been steadily falling behind Intel, which precipitated the migration. If this happens, I think it will be at the low-end of Apple's "PC" product line, use the App Store as the primary source for apps, and have no emulation of x86 apps since large app suites from MS and Adobe aren't necessary out of the gate when talking about an entry-level machine that could start a couple hundred dollars lower than the current Mac notebook without sacrificing perceived performance, or per unit or per machine category profits.

    2) Doesn't S-series refer to the ARM chip in the Apple Watch, and A-series refer to the chips in their iDevices? Wouldn't an ARM-based chip from Apple for their "PC" line then have its own name; perhaps an M-series starting with M1?
    Yes, in my mind the S-series SiP would allow better component integration/performance/reliability.  I was also thinking that x86 apps could be run in a shadow VM or some kind of app sandbox similar to Rosetta but passing x86 code to the x86 logic for full speed execution.
    The S10X could be used to prolong battery life in the full MBP or for a non-Intel option once a viable ARM software base had been cross-compiled.
    It's 2016 and we still can't an iGPU and dGPU working that well together using OpenCL, so I'm doubting that we'll be able to have an full integrated and seamless handoff between Apple's ARM chips and Intel chips running on the same machine. There are just too many issues and not enough benefits for for me to see that as an advantageous solution.
    I thought having control of their own supply chain and the ability to adapt hardware to suit the software would be highly compelling for Apple.
  • Reply 33 of 38
    zoetmb said:

    And I don't see how if chips are just getting delivered to manufacturers now and not necessarily in large quantities, how Apple releases a new MBP with those chips in November, even though some Window machines are apparently already hitting the market.  
    The ones hitting the market are low power Kaby Lake U/Y variants. Not the same thing.

    On MR I had said if Intel shipped quad core Kaby Lakes this year, I'd eat my 2011 MBP. WHOOPS. I didn't think anyone expected them to be ahead of schedule.  Are these with Iris Pro? Because the Skylake quads without IP have been out for months,  with the IP chips still barely available. 
  • Reply 34 of 38
    This is great news and makes a lot of sense for the big delay.  Why would Apple take so long to upgrade the MBP to only use Skylake processors since this could have been rolled out in June for the big back to school period.  My college age son has been waiting all year for this upgrade but went back to college with his 3 year old Asus which he hates.  
    The Skylake chips they'd likely use with Iris Pro 580 have barely seen the light of day in only the Skull Canyon NUC. Still not aware of any shipping laptops with those chips. 
  • Reply 35 of 38
    Soli said:

    If Apple can't get Kaby lake for a shipping MBP this year, then I'd be happy with Skylake, assuming all the other noted changes to the MBP are also coming down the pipe.

    I think I agree, with the same caveat you mentioned.

    Do I understand correctly that the significant difference between Skylake and Kaby Lake is the latter's integration of TB3 and USB3.1/G2 controllers with the CPU on a single chip?

    if so, could Apple get the same net result using Skylake CPUs with separate controller chips for TB and USB? Is there some significant downside to that approach?

    I very much want the new machine to come with USB-C connectors supporting the most up-to-date protocols possible, but is there a reason I or Apple should care which chips are used to accomplish that?

    (Just to be clear, I'm not armchair second-guessing Apple. I'm genuinely curious.)
  • Reply 36 of 38
    Do I understand correctly that the significant difference between Skylake and Kaby Lake is the latter's integration of TB3 and USB3.1/G2 controllers with the CPU on a single chip?

    if so, could Apple get the same net result using Skylake CPUs with separate controller chips for TB and USB? Is there some significant downside to that approach?

    I very much want the new machine to come with USB-C connectors supporting the most up-to-date protocols possible, but is there a reason I or Apple should care which chips are used to accomplish that?

    (Just to be clear, I'm not armchair second-guessing Apple. I'm genuinely curious.)
  • Reply 37 of 38
    Do I understand correctly that the significant difference between Skylake and Kaby Lake is the latter's integration of TB3 and USB3.1/G2 controllers with the CPU on a single chip?

    if so, could Apple get the same net result using Skylake CPUs with separate controller chips for TB and USB? Is there some significant downside to that approach?

    I very much want the new machine to come with USB-C connectors supporting the most up-to-date protocols possible, but is there a reason I or Apple should care which chips are used to accomplish that?

    (Just to be clear, I'm not armchair second-guessing Apple. I'm genuinely curious.)
    Yes there is a reason you should care. Kaby Lake integrates all the controllers you need for TB3 3.1. Apple will want to keep away from using a co-processor for TB3 or 3.1. There are some power issues to think about. Then drivers would be of concern. I for one would welcome apple to wait for Kaby lake so that it gets fully intergrated into the South Bridge on the CPU die.  
  • Reply 38 of 38
    sphericspheric Posts: 2,545member
    While I'm sure that Apple would love to integrate everything into one die, I have a hard time believing that an extra TB controller would not be compensated for by the power savings of Skylake over previously used processors. 

    Also, citing "driver" issues as a potential problem does not sound like an argument made from Deep Familiarity With The Subject Matter. Surely, this would be not much different from how Thunderbolt 2 is supported - by firmware. 

    What info do you base your claims upon?

    Note that I, too, am genuinely curious as to possible drawbacks, because whatever comes out next is what I am buying: I desperately need to replace my work MacBook Pro. 
    edited September 2016 Soli
Sign In or Register to comment.