Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Should Apple do Microsoft a solid and sell them m1 chips?
- Thread starter TouchOfClass
- Start date
- Sort by reaction score
They could, but would be better for everyone would be for Microsoft to dictate to Qualcomm to make something specific for their laptops that meets or exceeds it.
It will take Qualcomm a while to get there, and they’ll dictate that Microsoft has to make very strategic investments to get there.
Just to be clear, Apple has been doing this for however many years internally with their own teams. “We’re going to increase your budget to make something that puts Intel to shame”. At the same time Apple has cut throat deals with their partners that essentially says “we’ll invest this amount of money in you to get your manufacturing up to scale. When you get there we get to use it exclusively for this period of time at this cost.”
This is going to happen in some form or another. Look at Ceramic Shield or whatever it’s called. I can all but guarantee you part of that investment said Apple will invest money to do it with Corning, but Corning can only use this for a set period of time.
Tim Cook, I love him to death. He isn’t the visionary Steve Jobs was, but he understands how to get blood out of a rock.
Probably for the next year or two we’ll see the PC side try to figure out how to make their ultra thin and lights run like the new MacBook Air. I used one for my entire work day today with nothing plugged in. I’m use to a MacBook Pro 16, and during conference calls with “all of the services” the fans were loud, and at points there was varying points of it not performing well working. The MacBook Air just performed better, running apps with Rosetta. I figured at the end of the day it would be basically dead. I went to plug it in, and saw there was 38% battery life. The translated Intel code ran just as fast, in some cases faster (longer conversation).
TSMC, if the rumors are true, has given Apple all of the 2 year capacity for 5nm manufacturing. Theres no reason to assume this is true, outside of the fact no one else has 5nm.
It’s important to be honest, and say that Microsoft really took the first move with the Surface Pro X. Microsoft knows this is coming. They clearly didn’t take their foot off the gas when Windows RT died. But more shocking, is that I’ve been running all day today on a MacBook Air with no fan. It was cool to the touch. It’s got 8GB of memory compared to my MacBook Pro 16 that has 16 and the MacBook Air with Apple Silicon wasn’t actively cooled.
Performance for native apps was shocking. Performance on Rosetta apps was on par, if not better. With no active cooling.
My MacBook Pro 16 is sold. I have a MacBook Pro with Apple Silicon with 16GB and 1TB storage is coming. My wife doesn’t use her laptop all the time, but I know I can use her MacBook Air and it just destroys the MacBook Pro 16 with Intel with what I did today. Sustained better than my Intel MacBook Pro.
I’ve heard people complain that the same chipset in all 3 products is weird. If the active cooling in the MacBook Pro Apple Silicon is able to give me that sustained at all points, and with double the memory I’m all in. I’ve spent 24 hours with the Air, but if my assumptions are right, macOS is giving the bulk of the load to high efficiency cores. If my assumption is right, the high performance cores are steady at 30% load. If my assumption is wrong, I should have plugged in halfway through the day. I shouldn’t have had 38% after working all day.
Satya is a very pragmatic business person. He’ll probably make the ask of Apple to get access to the M1. I know whats happening on the early sides of him pressing the matter already. Microsoft is offering to help LOB (Line of Business) applications to work on Windows on ARM. Satya will look at what he controls first which is the software and funding (term usually used to mean Microsoft helps), get control of the supply chain (ask Apple to buy), and incentivize other suppliers to meet the same metrics (tell Qualcomm that they’ll give them the money to make it happen).
I’ve got a lot of love for Satya Nadella. Truly, he deals with things he understands proactively. I think he knew this was coming, which is why Surface Pro X exists. I don’t think he knew how fast this train was going.
The benefit for us is competition. Microsoft being an actively aware of what they control, and can somewhat influence other things means that Apple has to be on their toes. Microsoft doesn’t control the whole widget. Intel has been very proud of the parts they have controlled.
The longer term question is how can Microsoft either get control of everything in the stack, or at least influence it. They have two routes, existing relationships with NVidia and Qualcomm. Microsoft has to make Windows on ARM be as smooth as the M1 on the software side. Point blank they can’t do it without it. It’s going to be the Visual Studio team, and the people that build everything that supports the platform, and the Windows team.
2 years. I’ll be happy to be wrong on that. I don’t think Microsoft would be willing to help companies move line of business applications to Windows on ARM if it was easy, or if there was a quick fix in a year.
There is always ARM X1 and it’s derivatives. They won’t be nearly as fast as M1 but they are faster than the mobile ARM chips.
Bigger problem is graphics. ARM Mali is bad and it has a lot of limitations.
That's so true.
Apple literally took all of TSMC's current 5nm capacity. QC had to go with Samsung's inferior 5nm for their 875.
From a business perspective: No.
Currently M1 Macs are unique and even if a lot of it is the tight knit interaction between hardware and apples software it is better for apple to keep their hardware to themselves
Currently M1 Macs are unique and even if a lot of it is the tight knit interaction between hardware and apples software it is better for apple to keep their hardware to themselves
Trust me I get that, thats why I think it will be 2 years for a full fledged effort will be clear.
The PC (not Windows Exclusive) is use to the model of “well we can take this collection of parts, and we can put them on a board, and let Mother Nature sort it out”.
It’s going to take time for the PC market to realize that. It’s going to take time for the people to try to figure out how to do that, all while Intel gets to 5nm, and Apple is running prototypes for whatever is sub 1nm.
If I was facing Apple in the ring, I’d give up. There is no clean path.
This is going to take time, compared to how everyone else is doing business.
Let me just give you an example. I know at one point, if you bought very specific models of Dell computers, if you were on the lower rungs of buying, you got what I called the “sweep the floor of what doesn’t fit enterprise needs, and if it powers on and installs Windows ship it for $500”. That worked for a long time. I wish that statement was an exaggeration.
I think what being “competitive” changes.
Reread part of what you said after the fact. It’s close to 2AM the night before Thanksgiving.
ARM Mali at least from what I’ve read isn’t great. I’m happy to be wrong. That means Apple needs to step it up. I don’t think Qualcomm is using Mali on their specialized Windows chips but I maybe wrong. If NVidia makes a specialized chip for Windows I can all but guarantee you that it won’t be using Mali. Nvidia is very proud of their flag ship line. They’d sell their GPUs with ARM chips for Windows. They’ll leave Mali for the leave behind chips long term for things like Raspberry Pi and the rest
Last edited:
The problem is also that it is not clear whether Windows world will benefit much from investing into ARM. New Tiger Lakes and upcoming Cezanne (Zen 3 mobile) are very good mobile CPUs. It is not even clear whether Cortex X1 will even be more energy efficient. Apple is clearly an outlier here and I am sure that a lot of chip design companies are trying to puzzle out how exactly they did what they did.
Yep, that's exactly how it is. Most of PC companies just don't make good things (by very rare exceptions). It's literally "put together some scraps and ship it out". Even if fits enterprise needs
At least Apple is trying (although sometimes they are trying to much and they are surely opinionated).
A better thing to do would be for them to work out a solution for running WoA on ASi Macs.
The problem is also that it is not clear whether Windows world will benefit much from investing into ARM. New Tiger Lakes and upcoming Cezanne (Zen 3 mobile) are very good mobile CPUs. It is not even clear whether Cortex X1 will even be more energy efficient. Apple is clearly an outlier here and I am sure that a lot of chip design companies are trying to puzzle out how exactly they did what they did.
Yep, that's exactly how it is. Most of PC companies just don't make good things (by very rare exceptions). It's literally "put together some scraps and ship it out". Even if fits enterprise needs
I think we’ve had this discussion before a few times.
When it comes to AMD I have nothing but love. Trust me, I’m invested there.
We can argue on points all day. I used a laptop literally all day with no active cooling. I used a MacBook Air that is literally the slowest one Apple will make, and it’s running with more consistent performance than the MacBook Pro 16 that I use every day. Trust me I can make up scenarios where to make the MacBook Air with Apple Silicon chokes. But it doesn’t change the fact that the fan is on all the time on my 6 “high performance cores” from Intel, and it’s begging for air.
The MacBook Air, with no active cooling in my daily usage is swinging out of it’s weight class. The fact I did that for a 8 hour day and still has 38% battery life, while performing like that is astounding.
Is the MacBook Air for my wife faster than my Xbox Series X? No, by any means. Is it more consistent, reliable, with better consistent all day performance, at $999 or whatever the 7 core GPU model costs, while leaving me with 38% battery life? Yes. With no active cooling.
If you have a prototype laptop somewhere laying around that has no fans, beats it in performance, and Intel is selling the chip for laptops that literally have no heat with no cooling I’d call that amazing.
You do realize I am discussing the low end of the Apple Silicon Macs that we’ve ever seen right?
I also updated my last post to clarify points better.
The cheapest MacBook Air doesn’t cure Covid. I actually asked the Apple sales rep that wanted my questions before I purchased the laptop. She confirmed, it doesn’t. But from someone selling a MacBook Pro 16, with 16GB of ram, to buy a MacBook Pro with Apple Silicon, considering the laptop that is the cheapest MacBook Air available just beat the snot out of it, for 8 hours with fewer performance dips, I’ll take it.
Apple Silicon doesn’t fix Covid, and it’s almost certainly slower than a chip that Intel claims does something at some point. But only after they put a giant fan on it, make it Desktop for a few years, while they claim that they are really close to 7nm, despite Microsoft and Sony have been shipping at higher volumes for years.
It’s impressive. I’m glad you are telling me that Intel has a roadmap. They’ve been very good at having a roadmap.
I think we are in complete agreement here. In terms of chip design, Apple is so far ahead everyone else that it's not even funny.
What I am trying to say though that I am skeptical whether any other chip designer will be able to replicate Apple's success in close future. It's not just about ARM vs. x86, it's about chip design, and these things take time.
Apple has truly a unique advantage here, because they have approached CPU and GPU design from a very different angle. Their aim from the start has been making desktop-class chips while being constrained by a thermals of a mobile phone. It's a tough engineering problem to crack. Senior members of their engineering teem have stated in interviews that it has been a ten year journey. Intel and AMD approach things from a different perspective — they were never thermally constrained and their architecture doesn't scale down very well. ARM (for it's own Cortex) in comparison only focused on low power consumption, performance was always a secondary thing. They might be able to scale it up somewhat, but it is unclear whether they have the know-how to implement really wide designs like Apple.
I remember hanging with a bunch of AMD execs during their very successful Athlon64 run. They turned to me and said something like "Conroe sounds like the real deal, so we're in for some hurt". That would have been early 2006 or late 2005 (I don't remember) but it took until Ryzen to get their footing back, so that's what? 8 years?
CPU engineering roadmaps are long, and I don't believe anyone can catch up quickly when behind (look at Intel). It takes YEARS. Intel lucked out with Conroe because they had a very sharp B team working on mobile. I'd be shocked if Qualcomm or any other ARM partner is even close right now, and competitive products will likely be years from now.
Exactly. I mean, Intel is still essentially milking their Pentium M/Pentium Pro architecture — over 25 years after! AMD has been doing clean slate designs more frequently, but Zen still took at least 5-6 years to develop.
I don’t see the point, just as I don’t see the point of android smartphones using the A14 (or whatever) chip.
Apple’s processors are optimised for their own OS and software. Having it run windows will probably negative some of the performance benefits, to the point where you may not see significant gains compared to Intel.
Apple’s processors are optimised for their own OS and software. Having it run windows will probably negative some of the performance benefits, to the point where you may not see significant gains compared to Intel.