Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Based on what?


based on it was the shift to x86 architecture that increased Mac OS popularity. Its became a common core to build on. Most cross platform frameworks work from this core. internally they go x86 is x86 (or amd procs)..the instruction set being that common thing to work with. Why a lot of applications have the mac os, windows, and Linux client for the low price of 49.95 universal licensing. The framework used makes this easier to do and its tweak issues as they arise for mac os or Linux (windows would be the primary development drive).


Apple goes this route it can't bank on 3rd party following them. Some 3rd party again. Computers need software to run to sell them, its an ecostyem. This is not a recompile and done process. Its a new architecture. Even with building from source on sparc this was fun in the past using source code. And the punch line...this was CLI application stuff. Should in theory be easier, no gui "bugs" when buttons don't work or graphics issues. It wasn't. Compiles in sparc could get complicated to nail it right and not have error reports galore from compiling.



Inb4 there are arm based systems out there. They are afaik amd based servers. They have coders on staff to code in this environment. When you hire in house devs, you can do unique things. this keeps many "dead" languages alive...they never converted to newer languages. Cobol good enough till now...good enough a while longer kind of thing.

In one real world implementation I read a about a bit back their use will be virtualized or containerized server solutions. If containerization used, end user doesn't even see the ARM inside. Nor tbh...does the container. I have docker containers that only know they are Linux servers, resources just there. They don't know they are containers in a Linux VM I created and run docker from in parallels running on MacOS. Virtualization the same concept. At least with ESX. I run dozens of windows server on ESX at work. ESX is actually a custom Linux based setup at its core. My windows servers don't see this custom Linux core.
 
I wonder how Intel has managed to keep a large portion of their manufacturing in a high cost location like the US. Given the way the rest of the US manufacturing sector has gone - I wonder how Wall Street has let Intel get away with this, almost a breach of fiduciary duty. Maybe because they just can't ship their tooling to China with all the export control restrictions.

Semiconductor manufacturing is not like any other manufacturing. The investment in the most advance level wafer fabrication facilities, like what Intel just built in Fab 42 is.... in a nutshell... insane by any measure. When I got into semiconductors in the 70's there were hundreds of manufacturing companies all with their own wafer fab. Today, while there are still quite a number of wafer fabs at what one might call the mid tier level, in the world there are only about 4 commercial companies with the capability of the most advance lithographic manufacturing. Intel, Samsung, TSMC, Global Foundries make the top. Intel has kept is manufacturing in the US, I would say primarily, so they can control the IP. China is working to catch up but they are still quite behind the rest of the world when it comes to wafer fabrication technology. A wafer fab has two main cost structures, fix and variable. The fixed is PPE (Plant, Property, Equipment). The variable is material and bodies. While the variable is important the PPE is what eats your lunch if you don't keep the fab full and PPE is going to cost close the world around.

Intel has been hammered by Wall Street and deservedly so but it's not for not offshoring manufacturing. It's for having such hubris and myopic vision that they did not see the portable hand held market in tablets and cell phones and by throwing away the opportunity to be an ARM licensee years ago so they could make products for that market. So now they are have a big crow dinner. They will be an ARM licensee and they will have a chance at keeping their fabs full at running products for the portable market.

A few years back I tried to explain the semiconductor market to my dad, a retired accountant. What I told him is you are a company that makes widgets. To manufacture your widgets you need a factory that will cost you about 1.5 to 2 billion dollars (this was a few years back as I mentioned. They cost more now). So you build your factory. Now to profitably pay for that manufacturing facility you MUST get at least 10% world wide market share for the widgets you make otherwise you will never have the volume to drive your product cost to an economical level. Oh.... and you get to gut your manufacturing facility in 3 years because it will be obsolete by that time so you get to start all over again. He told me my business was certifiably insane. I can't say I disagree.
 
  • Like
Reactions: mtneer
It will be interesting to see if this turns into anything in the long run for possible macOS/iOS hybrid devices, or simply arm-based Macs in general. Intel could certainly take things in an interesting direction.

My gut tells me this has a hint of ARM in Macs

Yup, agree with both of you. This would be a way for Intel to not only keep Apple's business, but facilitate their ARM migration by offering to throw some custom Intel-only hardware onto the chip to help it run x86 code faster. So they keep Apple's business and (some) of their own architecture in place going forward.
 
  • Like
Reactions: macduke
based on it was the shift to x86 architecture that increased Mac OS popularity. Its became a common core to build on. Most cross platform frameworks work from this core. internally they go x86 is x86 (or amd procs)..the instruction set being that common thing to work with. Why a lot of applications have the mac os, windows, and Linux client for the low price of 49.95 universal licensing. The framework used makes this easier to do and its tweak issues as they arise for mac os or Linux (windows would be the primary development drive).


Apple goes this route it can't bank on 3rd party following them. Some 3rd party again. Computers need software to run to sell them, its an ecostyem. This is not a recompile and done process. Its a new architecture. Even with building from source on sparc this was fun in the past using source code. And the punch line...this was CLI application stuff. Should in theory be easier, no gui "bugs" when buttons don't work or graphics issues. It wasn't. Compiles in sparc could get complicated to nail it right and not have error reports galore from compiling.



Inb4 there are arm based systems out there. They are afaik amd based servers. They have coders on staff to code in this environment. When you hire in house devs, you can do unique things. this keeps many "dead" languages alive...they never converted to newer languages. Cobol good enough till now...good enough a while longer kind of thing.

In one real world implementation I read a about a bit back their use will be virtualized or containerized server solutions. If containerization used, end user doesn't even see the ARM inside. Nor tbh...does the container. I have docker containers that only know they are Linux servers, resources just there. They don't know they are containers in a Linux VM I created and run docker from in parallels running on MacOS. Virtualization the same concept. At least with ESX. I run dozens of windows server on ESX at work. ESX is actually a custom Linux based setup at its core. My windows servers don't see this custom Linux core.

I agree with you that it will take much work and migration to the new platform and understand what you're saying but x86 design and power consumption is becoming a hurdle to longer battery life and power efficiency. Their has to be a massive action taken to break that plateau on the end user side starting with the Macbook (perfect candidate for an ARM CPU) and ending with the MBP.
 
Yup, agree with both of you. This would be a way for Intel to not only keep Apple's business, but facilitate their ARM migration by offering to throw some custom Intel-only hardware onto the chip to help it run x86 code faster. So they keep Apple's business and (some) of their own architecture in place going forward.
I was somewhat surprised to hear that Intel was moving to ARM production, and most certainly Apple is behind it as you'd never expect Intel to be willing to do something like this of their own accord. I figured they would beat x86 to death until they went under.

As a hardware partner, Apple may have been telling the higher-ups at Intel that their days using x86 were numbered. Intel knows that as Apple goes, so goes the market. At least that's how it generally works these days. They could sweeten the deal by promising iOS A-series contracts to make the switch worth their while.

You're probably right about Intel putting custom instructions and other optimizations into their chip designs to help with porting legacy software through the transition. Considering how relatively smooth the transition from PowerPC to Intel was, I imagine this transition will be even smoother with another 10 years of experience along with experience being a chip designer.

It makes so much sense when you look at the buyer's guide and see "DON'T BUY" pretty much across the board for the Mac lineup. Why would they let every single Mac fall behind? Especially given all their chest puffing saying "Can't innovate my ass!" when they announced the Mac Pro just a few years ago? Why let that thing get so outdated? I never thought they would be crazy enough to move everything to ARM at once, but now I have to wonder. The dark horse announcement for an autumn event? Makes sense why they're adding important yet likely easy to port features from iOS to make things even more cohesive. Could they have the iPhone/iPad event in September and then a crazy ARM Mac event at Apple Campus 2 in late October or early November?

Tim: "Welcome to our new home, Apple Campus 2. This building gives us a solid foundation to design and develop innovative products for decades to come. And that innovation and solid foundation starts today with the Apple X1 processor. To tell you more about this processor, I would like to invite to the stage Intel CEO Brian Krzanich and Apple Senior Vice President of Hardware Engineering, Dan Riccio to the stage."

Dan: "Thanks, Tim. For many years now our team here at Apple has been collaborating with Intel on a new chip design that bridges the gap between mobile and desktop. A chip that possesses the compact size and efficiency of a mobile processor, with the sheer performance found on a desktop. This is a single chip that can run both iOS and Mac OS software. And we're going to show you that chip now. Brian?"

Brian: *Holds an incredibly tiny chip with a pair of tweezers. The camera zooms way in for dramatic effect.* "This is the X1, the first ARM chip fabricated by Intel, designed in partnership with Apple. Watt for watt these chips are about 50% faster than their current generation i-series counterparts when optimized by Apple's multi-core framework. The chip I'm holding is the X1-C, which is suitable for the MacBook. It has six cores and runs at 2GHz, but only consumes 6 watts of power. This is possible due to Intel's industry-leading 10nm process."

I got tired or writing dialog, but they would go on to talk about the performance advantages, maybe Craig would come up and do a demo, etc, showing cross-compatibility with iOS apps and talking about Xcode updates and how the scaling works to port iOS apps. Someone like Phil would come up on stage and talk about innovation, go through all the new specs across the Mac line, etc. They would show performance numbers and how they compare to older Intel chips. They'd talk about backwards compatibility with x86 applications and demo that. They would talk about the different performance tiers. I imagine they would have a few different tiers for the processor based on performance and how much power it uses. I envision the iPhone and iPad running something like the X1-C. The iPad Pro and MacBooks running something like an X1-B, and maybe the MacBook Pro/iMac/Mac Pro running some variation of the X1-A (maybe the MacBook Pro could have one X1-A with 12 cores or something, and the iMac would have two X1-As, and the Mac Pro would have four X1-As. And yes, I realize my naming convention isn't the best. Hopefully they come up with something better. But what I'm getting at is that it would be important to denote different levels of performance, much like they do with the i3, i5, and i7. Although the iPhone having a "lower tier" chip could hurt it's branding as a premium device. So IDK, maybe they'll just change the first letter, or keep the A-series for the iPhone and only use the X-series on the Macs and maybe the iPad Pro. It would be a really fun show!
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.