Apple now Arm'd to the teeth: MacBook Air and Pro, Mac mini to be powered by custom M1 chips rather than Intel Apple on Tuesday unveiled three computers based on homegrown Apple Silicon processors, marking the start of a two-year transition that will replace Intel microprocessors with Cupertino-designed chips compatible with the Arm microarchitecture. "In June, we announced that the Mac is taking another huge leap forward by …
So we've gone from OSX to macOS and now MacOS 11....
But that raises the question... Is that eleven or the roman numeral two? (considering the previous roman numeral).
My biggest questions though... How much will we have to fork out for new software this time/how broken will existing stuff be when plugged into it?
That makes it sound worse than it was; the original Rosetta appeared in 2006 along with the Intel transition and remained part of the current version of the OS until 2011. Which is essentially the same lifetime as the Classic Environment. So if you believe history has any effect then expect Intel translation to be available for about five years.
Apple are not calling it macOS 11. That's The Register labelling it as such (though version 11 is shown in the Apple > About menu).
Apple have dropped the OS X moniker some time ago in favour of macOS and will no doubt continue to use California landmark names for the foreseeable future.
Imagine if Win10 was called Windows XP (Ver 5.10.1234 or something)
Well, it is Windows NT version 10.x.
More confusingly, there are a few extra half-steps in there. All of PowerPC, x86 and Arm have had architecture changes from 32-bit to 64-bit versions within those steps (32-bit Arm only on iPhone though). These sneak past because the hardware itself manages a lot of the backward-compatibility, so hardly anyone whinges, but they are extra instruction set changes that have to be handled by the OS and software ecosystem (compilers, developers, etc).
The upshot is that Apple developers are really quite good at it by now, after all that training.
newspupoy is counting how many architectures Macs have used. Fair enough. But there's a case for counting how many architectures OSX / NeXtStep has been on - it was designed to be as architecture agnostic as possible, and ran on a few things before Jobs rejoined Apple.
When contemplating putting OSX on Intel, an Apple engineer was sent out to buy a high end Sony VAIO laptop, and they got OSX running smoothly on it within two days.
All the heavy lifting for porting to x86 had been done between NextStep 3.1 and NextStep 3.2 back in 1992/93. A decade before the Apple PPC supply debacle which forced the x86 transition. Cant remember who NS suckered to pay for that particular port.
I can assure you the original NS3.1 port to x86 was a lot longer than 2 days. Given how unstable every NS x.0x release was ( every initial NS 3.x release broke the SCSI driver for example) I suspect it took quite a lot of engineering man months to get from the 68k/SPARC code base (at the time) to something that could run the core NS frameworks on x86 without crashing. Although back in those days NS always was just betaware. Sometimes barely even that.
Mac OS 9 was named as such for marketing reasons. It was originally Mac OS 8.7.
I'm fairly sure that Mac OS 9 came about as a bridge between Mac OS 8.x and the new Mac OS X.
I still believe that the X in Mac OS X is primarily UNIX. Until XML came along, X is a name meant UNIX.
It's macOS, with a lower case "m" and they promote it as "macOS Big Sur".
The "major version", which has been stuck at 10 all this time, moves to 11. There has been confusion over the names of some previous updates. For example, the current version is 10.15.7 and the patch last week was just "macOS Catalina 10.15.7 Supplemental Udate". Because they were misusing the semantic versioning conventions to mean 10.major.minor instead of major.minor.patch, there was no patch number available. Once there were even two identically named "Supplementals". You'd have to find a hex build number code, which even sometimes differed on different Mac models.
Old app builds see Big Sur's version number as 10.16.0, and new ones as 11.0.
All in all, it's a long-needed change.
Linus calls it a slow-motion dumpster fire (10 mins) and I'm inclined to agree.
Completely unexpandable, completely unfixable, performance figures as given by Apple are intentionally vague, the OS is even more locked down to the point where it becomes difficult to distribute software on Macs without Apple being the gatekeeper and the three models they've released (Air, Pro, Mini) have exactly the same specs.
The future of Apple hardware is here and it's crapper than ever.
Apple has made no changes at all to how locked-down the OS is today. You can still download and run unsigned binaries, you can still compile whatever you want. The creeping attack on that might continue but didn’t factor into today.
It’s also not clear that the three machines have the same specs, as it appears that one is passively cooled (Apple made a big deal about that) whereas the other two are actively cooled, and Apple hasn’t said a word about clock speeds.
So: worth waiting for reviews on that. It may be true but it may not.
Or it's passively cooled because it's a different version of the chip. Certainly the cheaper version of the Air has a lesser version of the chip because Apple actually gave us a technical detail: 7-core GPU instead of 8.
Since Apple refuses to tell us much else, I intend to wait and see. I certainly won't be following the opinion of a Youtube 'celebrity'.
Or it's passively cooled because it's a different version of the chip. Certainly the cheaper version of the Air has a lesser version of the chip because Apple actually gave us a technical detail: 7-core GPU instead of 8.
Yes, having 7-core GPUs somewhat suggests that the Air is getting the binned SOCs that have a faulty core. Drop the clock speed, disable the faulty core and they'll run stable with passive cooling. It's highly unlikely that they're crippling perfectly good chips in firmware - they're just the lower-binned version (like every foundry does - bin and brand them according to how many cores came out working & stable).
Sure.
Just seeing lots of people going "but it's the same chip across all of them" as if Intel's chips of a certain generation aren't all pretty much exactly the same.
Intel don't make 5 different i7s or whatever, they make one, bin it and then lock the frequency to create artificial models at different price points.
People fall for branding that Intel and AMD use as we're so used to CPUs being sold to us in that way.
The way I see it, Apple have no need to do that with their own chips. Instead they'll adjust the TDP and power envelope to suit the device at hand and maybe use chips with failed components in lower ranges. (Like the air with only 7 graphics cores)
"'Linus calls it a slow-motion dumpster fire (10 mins) and I'm inclined to agree."
That's not Linus™, there's only one rantyvocal Linus who gets to be referred to by only his first name! If, like me, you followed the link expecting commentary from the instigator of Linux, be aware that the video is actually from some other tech pundit.
Rather than calling them hardware and software Linuxes i prefer: the gurning clickbait clown from Youtube and the programming genius.
Anandtech has good coverage of the M1 up right now for those who are actually into hardware. Since they forgot to come up with a misleading provocative title and post a picture of somebody pulling a reaction face, they won't compete for Youtube Linus's audience.
Anandtech is getting very excited about the chip itself, mostly based on extrapolations from the A14, but as for the final products (or probably better to call them the final three form factors) it dutifully and uncritically rattled off Apple's meaningless benchmarks as they were live streamed without drawing very many conclusions at all.
Completely unexpandable, completely unfixable
How is that different from the previous 5 years of Macbooks (and even the most recent Mac Mini) with soldered RAM/storage?
Yes, it's undesirable - especially in a desktop like the Mini - but it's nothing to do with moving to ARM and everything to do with Apple's packaging decisions.
This post has been deleted by its author
This is great news for people buying a notebook but then use it as an iPad.
Seriously though, it's going to be years before you can do actual laptop tasks with these. Just this week I've flashed firmware in my Jeti ESC and receivers, used Fusion 360 to design a part for my plane, and used Cura to slice the file for printing.
Fusion 360 runs on Macs, but many of its competitors are on Windows only - so many Parametric CAD users would find an Intel Mac no more useful than you do an iPad (if it weren't for Bootcamp, of course). Computer Aided Design for Manufacture is in an area traditionally run on Windows (or once mainframes).
Whether Autodesk has an ARM version of Fusion 360 up their sleeves - I don't know (daft if they don't, since their x86 iOS version surely won them customers). However, most Mac users aren't CAD users, so some will find that Apple software and Adobe suite suit fulfil their needs. And there's no reason that the likes of Cura won't run under Rosetta II.
(Side note: CAD is a curious one - there are good reasons for running it on a cloud instance: every engineer can see the latest version, the data (possibly valuable, client confidentiality, national security, esp in aerospace) isn't stored on any stray laptop, RAM, CPU and storage can near unlimited, rented as needed. In this scenario, the power of the laptop is mostly irrelevant, as long as it runs a browser well. )
Upvoted for main text. The side note is only superficially true. Actually the cloud gives less security and for many would have performance issues. It would be going back to the 1960s; a wet dream for Cloud Service providers and Adobe.
Laptop power, RAM and storage has been pretty irrelevant for most CAD (simulation is different) for over 15 years. The issue is often been people needing to add external screens. Even my 2002 1600 x 1200 1.8GHz laptop wasn't retired till 2016 and was still fine for Spice and all 2D CAD/CAE. A bit slow for 3D rendering of a PCB layout!
Security of cloud is only as good as the Cloud vendor's security and the security of credentials. No physical access needed. Really critical security places the ability to copy to removable storage or unauthorised Internet is zero and you can't take the laptop/workstation away.
Depends upon the project and how many people are concurrently working on it. For sure, having engineers only access data on site, and that data held on the premises, is the most secure. So in this context, just replace 'cloud' with mainframe or server, and the upshot remains the same: computer on desk just needs to drive monitors. Changes an engineer commits to a model are immediately available to colleagues.
I didn't mean to suggest that the server / mainframe model is the definitively best way to use CAD in all situations, only that it does offer some plus points. The hybrid approach is popular - modelling done on local workstation, simulation / rendering done on local GPU farm or rented cloud resources.
"Whether Autodesk has an ARM version of Fusion 360 up their sleeves - I don't know"
That is the $64,000 question. Autodesk has already ended macOS support for their Alias and VRED software packages and more of the same could now follow.
I suspect that Apple management has already done the calculations - they'll ultimately gain more users overall by switching to Apple Silicon so allowing easy use of iOS apps on Macs but they'll lose out at the other end a bit when software suppliers just won't bother to release high end scientific and engineering software versions for ARM Macs.
Coming up on 25 years ago, I was doing sysadmin at a CAD shop using SDRC I-Deas Master Series (super-expensive mechanical CAD software) on HP-UX workstations (C-series was what we were moving to by the time I left, which I admit places this in time pretty well). Windows NT and AutoCAD was "for those who weren't doing serious work", but many of your points about the advantages of cloud CAD were solved way back then because every part had to be checked into a central server library. People would check out a part, work on it, and then using the software's built in version control, check it back in. I don't recall what the underlying VCS was, but it had the concept of formal checkout to lock a part so two engineers (as in mechanical) didn't accidentally work on the same part at the same time. The real reason we needed super-large horsepower was at that time, the CPU and GPU requirements for the serious work were prohibitive for an intel based desktop to get.
Part of my job there was to retrain the engineers to actually check in their components every day and make sure they understood that any changes they failed to check in each day, they should presume that their workstation hard drives would both crash and they would lose all their work.
Just this week I’ve done actual computing tasks (form factor irrelevant) including sending email, browsing the internet, writing code, compiling code, headbutting the keyboard over some obvious error, reading RSS feeds, doing the accounts in a spreadsheet, cursing loudly, and turning off wifi in the middle of a Zoom call to simulate failure*.
All of these are real tasks. Few of us are aeroplane designers. I doubt that there are even many people who pretend to be aeroplane designers.
* Not actually this one - I only just thought of it. But I think I might install the Network Link Conditioner test software to spinelessly cause my call to drop when I get bored of it. Good idea, no?
I found this from a month ago from the Autodesk forums:
"Fusion 360 will not be a native ARM application out of the gate; testing with Rosetta 2 has been positive so far, with my personal experience being that performance is largely unimpacted. No doubt a few tweaks and optimizations will be necessary over time, but if you were to install the x86-64 version of Fusion 360 on Apple Silicone today, the difference is practically imperceptible."
- Lance Carocci, Fusion 360 QA for UI Framework/Cloud Workflows
If Fusion 360 is happy running through Rosetta 2, I can't see a simple STL to G-code 'slicer' application being upset.
Seriously though, it's going to be years before you can do actual laptop tasks with these.
Sounds like the PHB: "Bring me a hard copy of the Internet so I can do some serious surfing."
This post has been deleted by its author
No more upgradable RAM in the Mac mini.
I was hoping for a new Mac mini running ape silicon that I could stuff with RAM & ssd 6 months or so after I bought it.
Its to replace my Lenovo xenon server that has 32gb dam and could do with at least 64gb.
I wonder how quick that ssd is and given the new changes I wonder if 16gb could actually be enough to run a shed load of docker (once it’s running on apple silicon, but I thought sur used docker containers?) containers.
All the low power gear is now actually pretty pokey and an interesting alternative to the intel alternatives. Apple silicon will be far quicker at image/ video processing than intel equivalents given all the extra image focussed processing cores.
Non replaceable RAM in the Mac mini is a real blow though.
Will be interesting to see the benchmarks.
If you want removable components, you're looking at the wrong product
Sad but true. It's crap that a new logic board is the only way to replace an SSD. Until now, there were two exceptions to the soldered RAM, though.
Mini: It's an enormous pain in the ass, but as long as you're careful (and don't mix up the 14 different kinds of almost-identical microscopic screws) it is possible.
27" iMac: Slotted RAM under a removable cover on the back panel, and it's considered a user-replaceable part.
Fusion Drive era and earlier iMacs had a plug-in HDD, but you had to deal with adhesives and risk breaking the screen to get into the @#$% case.
"Having the RAM in the same die as the CPU will have a HUGE impact on performance."
Where did you see/read the RAM is on the same die? Even in the article shot of the M1 processor you can clearly see it is not. It would be impossible for a number of reasons, including that TSMC do not fab SDRAM on 5nm. SRAM cache yes. But not the SDRAM.
have a read of the anandtech review. https://www.anandtech.com/show/16226/apple-silicon-m1-a14-deep-dive .
the RAM may not be on the same die as the cpu "package" but its part of the solution and physically a damn site physically closer than other solutions
certainly apples graphics make it look like a fully integrated solution
https://www.apple.com/mac/m1/
The reasoning for putting ram to the side is for heat dissipation.
regardless you need to buy the system with the ram you need now.
my suspicion is that coupled with fast disk 16GB would be more than plenty and 8GB would likely be more than enough for most use cases.
when review units arrive this thing will blow anything else out there for any meaningful tasks thanks to the other cpu's on the die that will chomp through image processing that most stress tests like to focus on.
if your a power user wait for the next lot of systems that may be multiple M1's or M1+'s
This is clearly not a server solution but a workstation solution.
Servers may still be best served by x86 as likely the M1 will not be suited to virtualisation.
Something something Commander Vimes economics *
I just had to trawl my email archives...
My wife's laptop was purchased 5 years ago (ok, it's about four weeks short of five years). I have, in the intervening time period, upgraded the storage (a not quite standard stick SSD), and for a couple of quid got a PCIe card to take the old one which is now sitting as nice fast app storage for an unRaid server.
It was ~£1k five years ago, and shows no signs of ageing badly at all - still runs the current version of macOS, and handles day to day needs flawlessly. I don't imagine it will need replacing for a good while yet, but the Air does have me interested.
My one real complaint is that they don't still have MagSafe. That is one of the best solutions to laptop power delivery ever, an I don't really want to have a USB-c magnet sticking out. OTOH it will likely only ever be charged overnight, so it possibly makes no difference at all.
Given that it's only a 30W charger - I'm slightly surprised it doesn't have a pair of wireless charging pads on the base... That could be rather interesting on a desk.
Even accounting for the SSD upgrade and a couple of cases that's ~£200/year (and falling). Not particularly expensive for a daily working machine.
My daughter as a Surface Pro (albeit a few generations old and second hand), another very well put together bit of hardware. But that's *more* expensive!
In fact an i5/8GB/128GB Surface Pro is about as much as the new Air (after discounts, including a keyboard, excluding Micros~1 265) - to get the same RAM/Storage it's £300 more.
In fact they charge £300 for the move from 128GB to 256GB storage, and £600 to go to 512 and £350 more to go to 1TB - their memory doubling is over £200 each time as well.
If the CPU power holds up to even half of their claims then it will be a worthy successor for the wife's machine when it needs it.
In fact if the CPU is sufficiently capable I could be tempted to replace the Mac mini (late 2012) that is starting to show it's age as a desktop - but would still make a good little server somewhere.
* Commander Sam Vines of the Ankh-Morpork watch:
"Take boots, for example. He earned thirty-eight dollars a month plus allowances. A really good pair of leather boots cost fifty dollars. But an affordable pair of boots, which were sort of OK for a season or two and then leaked like hell when the cardboard gave out, cost about ten dollars. Those were the kind of boots Vimes always bought, and wore until the soles were so thin that he could tell where he was in Ankh-Morpork on a foggy night by the feel of the cobbles.
But the thing was that good boots lasted for years and years. A man who could afford fifty dollars had a pair of boots that'd still be keeping his feet dry in ten years' time, while the poor man who could only afford cheap boots would have spent a hundred dollars on boots in the same time and would still have wet feet."
14 inch MacBook Pro, yes, please!
I have a 2015 13" MacBook Pro which has served me excellently, but there are times when it would be nicer to have a slightly larger (higher resolution) screen. I normally run it at 1440x900, but trying to use 1680x1050 is a bit too hard on the eyes.
16" is physically too large/heavy a laptop for me, but 14" would be only marginally larger (possibly very little so with smaller screen bezels?), and, hey, if I'm dreaming, space for Home/End/PgUp/PgDn/Del/etc keys would be very nice too…
First, you *can* plug an iPhone into an external monitor.
Second, the point is the new Apple ARM processors are specifically designed for low power mobile use. They've been used for years in iPhones and iPads etc. so yes, they will run MUCH cooler than Intel MBPs. The MacBook Air doesn't even have a fan.
You got lucky.
Check out
https://forums.macrumors.com/threads/16-is-hot-noisy-with-an-external-monitor.2211747/
which has thousands of posts from disgruntled users. Speaking personally, mine is OK until I actually tickle the CPU into doing something a little demanding. But the point is, a previous model (the one with the flaky keyboard) didn't do this.
For the purposes of the people trading Apple stock, today's announcement doesn't tell anything more than the June 20th WWDC announcement - i,e "Macs, potentially very fast and long battery life, two year phase in, most software native, potential of compatibility niggles likely to cause some consumer wariness".
Apple's share price dipped slightly after June 20th, before rising in a month to this year's highest... though of course iPhones and services are big factors.
> Until Spectre/Meltdown type flaws are found in the M1. Developing your own chips and doing security is *hard*
Indeed, but Apple aren't starting from scratch. They've had billions of chips from this family out in the wild for several years, at the mercy of motivated bug bounty hunters and security agencies.
Well, making a machine that competes with your rivals' machines on performance but costs you less money to make is going to be profitable. The trick is working out how to save costs. If it were easy, your competitors would already be doing it.
One would normally point out that the R and D cost of the silicon needs to included, but of course a huge chunk of that was paid for by iPhones. Again, sharing the R and D costs across phones, tablets and laptops is a smart move. A smart move that isn't easy for rivals to follow, since they don't control all the hardware and software.
A lot of the discussion here is how these ARM Macs compare to Intel. However, down the line it might be worth looking at other ARM machines running other OSs - flavours if Unix, Windows and whatever the hell Google is playing around with this month (ChromeOS, Fuschia etc). Could Qualcomm design laptop-grade ARM SoCs to compete with Apple's? Heck, even Google has played in the smartphone chip design space as they wanted ML stuff for their Pixel phones that Qualcomm wasn't doing.
As a long term mac user (8 of em in the house), I'm more interested in this than I thought I was - if it can really make editing my 4k video and 360 video quicker out traveling. But I reckon I'll wait for independant benchmarks. Especially for how fast rosetta2 stuff runs - since other than apple apps that's what you'll be getting for a good while.
I'm concerned about memory for the pro stuff though - that 8gb now has to do GPU and CPU duty. Before you had your discrete 4 or 8gb GPU memory also. Seems to me that GPU intensive tasks are going to hog most of that 8gb now, and cause real problems ?
For example, on my 2017 macbook when I process/stitch 360 footage with Instal360 Studio it heavily uses the GPU. Since it only has intel graphics I found I needed to reduce the screen resolution down to 720p (not HiDPI) to free enough system memory to allow the GPU to do it's stuff. I'd have thought you'd have the same issues on the M1 ? Only here they are selling that as 'pro'.
Talking of which - what exactly IS the difference between the air and pro ? 1 extra GPU core (which you can have on the upgraded air) and a fan ? that seems to be it really ? unless you count the stupid touchbar. I'd hardly say that counts as 'pro'... the fact it has only 2 thunderbolts also suggests under the covers it's just the air with a fan....
> Talking of which - what exactly IS the difference between the air and pro ?
Easy. The Air is built using chips with an internal fuse blown to permanently disable the extra features.
(Which leads me to wonder if Apple have used the security features to lock-out rather than permanently disable the extra RAM so that there is the potential to sell over-the-air upgrades later?)
I have no idea how SoCs are manufactured. I was imagining one piece of silicon with everything etched onto it - including a full 16GB which is then disabled for the cheap model.
If that's not the case, are you saying that the SoC comprises separate processor and RAM chips and that their address lines are joined via a mini-circuit board hidden within the SoC's plastic/resin packaging, instead of external legs/pins/bga that would traditionally be connected via the motherboard?
Or something else?
I was thinking one of the speed improvements was because it was all on one bit of silicon and therefore had physically short paths. It also explained why the limit was 16GB. If it's not that then I don't see why larger RAM sizes couldn't be supported.
"2017 macbook"
if you had a mac with the t2 chip you'd likely see an improvement in transcoding speed.
https://appleinsider.com/articles/19/04/09/apples-t2-chip-makes-a-giant-difference-in-video-encoding-for-most-users
the t2 chip that runs the "stupid touchbar" and on the fly disk encryption etc accelerates image processing, leading the i3 mac mini with t2 to outperform the i3 imac without t2 by almost double.
t2 is a derivative of the arm chip used in iwatches.
the M1 with 8GB ram will make a huge improvement to your work.
In three years time, when Apple has ditched Intel processors completely, we'll see the effect of this switch on Mac sales. Many freelance professionals use x86 Macbooks BECAUSE they can also run Windows on it. With ARM-based Macs this capability will go away, resulting in many of them buying Windows laptops instead.
So Apple will lose a lot of sales. The question is whether this can be compensated by new sales from other audiences. My guess is not.
"In three years time, I expect ARM based Windows will be more of a thing and available to virtualise on an ARM based Mac."
I don't see this happening. Like at all.
MS is not even able to issue working patches outside of a grand notepad font update, so my expectations for them to issue Windows on ARM during the 21st century is strictly 0.
We'll probably see HP-UX on ARM long before !
He means that Windows on ARM will never be a big thing, even though it has been available for years.
In fact, its availability has highlighted many of the problems of switching to a completely new architecture. All the device drivers will have to be rewritten (many from scratch). Tons of applications will not be ported and vendors are loath to port their applications since it will cost them a bundle.
You have no idea what you're saying. Do you even have an INKLING how much software has been written for Windows on x86? And the effort required to port this to another architecture? Clearly you do not.
Maybe some other operating system will overtake Windows, but Windows will always be tied to x86.
For most users there's no real benefit to switch to ARM, x86 works well enough even though it's less efficient compared to ARM. The drawback of having all your software run in emulation is however, unbearable, for a great many of them.
> The question is whether this [no x86 Windows Bootcamp users] can be compensated by new sales from other audiences.
Those Bootcamp users have a couple of years to see how things play out - more software may be coded for ARM Mac, emulation and virtualization might prove to be workable. In the meantime, pure MacOS users might just find the incredible battery life very enticing.
It's still *nix, so homebrew stuff just needs to be recompiled. Apple demoed Docker running on the dev kit when they announced it in June. There's also experimental support in Docker to run x86 containers on ARM. I tried it on a R-Pi 3 and got 32-bit Ubuntu up and running. That could be usable for dev work, just not for production environments.
No sensible person buys a mac for themselves
I'm no fanboi but I currently have two MacBook Pros (2016 and 2020 because I'm worried about potential restrictions on the ARM machines) and SWMBO has an 2020 Air. Depending on what you're doing the MacOS value proposition is not bad: a lot of power in a compact device with useful integration. I use the command line a lot and this is still way easier on MacOS than in Windows.
Comparisons with cheaper Intel notebooks without high resolution screens and similar weight/battery performance are misleading, even if my machines are generally attached to external screens and keyboards. When you need them for travel: weight and battery life suddenly become very important.
When it comes to phones, I couldn't agree with you more: you can get equivalent high-end phones but generally a budget one will be good enough for nearly everyone.
Where they are still in a league of their own is with the high end tablets. Google has fucked up here by not making Android more suitable and Microsoft is still favouring its own devices here. Again the ability to travel light is very appealing.
No sensible person buys a mac for themselves, they are a total rip off
My blind friend purchased a Mac because she uses the VoiceOver screen reader that's included with MacOS. She had too many problems with commercial screen readers on Windows. She also figures the Mac is cheaper because she saves $90/year for JAWS.
May I quote you directly when I tell her she was wrong?
Not in the sense of Intel chips are Intel.
So an ARM developed by additions to ARM's core by a bought in chip design house.
So it's not really Apple designed. They just own the group that enhanced an ARM design and Apple doesn't make any chips. All outsourced. Like the iPad and iPhone.
Originally the iPhone used the Samsung innovative Samsung SC6400, a Samsung enhancement of the ARM with the package containing three layered chips, the CPU-SoC, Flash and RAM. You certainly could not have done that with an Intel CPU then. It used a Samsung display, OS derived from MacOs, which was derived from Next Step, derived from BSD. Bought in Fingerworks to get the touch to work. The capacitive touch had existed for over 15 years, but because smart phones were corporate (since 1998) before iPhone data contract, they used high resolution resistive for annotation and content creation rather than low resolution capacitive.
Apple are really good at spotting tech and buying it in and spotting already succeeding trends and marketing them better. They also have good lawyers for the USPTO applications and court cases.
The chips are designed by individuals employed by Apple, based on an architecture developed by ARM, which itself was financially supported by Apple back in the Newton days. It suggests that someone at Apple bet heavily that spending huge amounts of cash on SoC development (rather than buying Qualcomm chips) would reap big returns.
Nobody here is too fussed about who invented or borrowed what first. It's often largely the integration of things that make an end product better fit for purpose.
The bottom line is, Apple have invested heavily in designing an ARM chip that is quite unlike any other (see Anandtech - the A series SoCs aren't mere enhancements of ARM's designs) to run software made by Apple to facilitate workflows it believes its customers want. If Apple believe that their customers want a laptop that do Zoom calls for nearly twenty hours on a single charge, so be it. If it believes that their chips can rush through AV work whilst sipping power, so be it. They can then select, fine tune and integrate the hardware and software required.
>>"big companies with money to invest" are really good at spotting tech and buying it in and spotting already succeeding trends and marketing them better. "big companies with money to invest" also have good lawyers for the USPTO applications and court cases.
That's almost always how companies become big, and those picking the good ones are the ones that stay big and become bigger.
Shareholders expect this of them - for their own pensions and savings.
And those acquired companies have seen their work to fruition and significant scale with this activity.
This is not an outsourced chip in the way the word "outsourced" is typically used. AMD x64 chips are not an outsourced chip. Anymore than the linux kernel is outsourced from Minix.
The samsung processor was a supplier relationship, there was no customization. Outsourcing relationships typically have customer specific customization done by the supplier.
This M1 is significantly different to anything on the market. The similarity for all purposes at this point is just the ARM ISA. ARMs purpose here is IMO to leverage toolchains, and that it works well enough with the phones/tablets.
For outsourcing to apply, someone out there *already* has an equivalent processor based on the ARM ISA to purchase, and Apple just brought in a customized variant of it. ARM does not - the pipeline and decode architecture is very different in the M and A series from Apple.
And at these specifications the answer is likely to remain no. Apple branding can command a price, and thus a margin, that allow for expensive architectural choices such as 192 KB L1 caches. Other companies cannot command the price for the end product (why not just buy Apple then) and so they cannot buy such a massive beast of a chip. This M1 is HUGE in size for a mobile consumer segment RISC processor.
Intel might make them now to compete, but it will swallow the margins that business line could historically report. CISC has inefficiencies. I don't know with their cost structure if they could even do it profitably. They have doing revisionist tweaks for years now since their 10nm debacles, so their architecture is not necessarily competitive to such threats.
"Apple’s early adoption of the 64-bit Armv8 ISA shocked everybody, as the company was the first in the industry to implement the new instruction set architecture, but they beat even Arm’s own CPU teams by more than a year, as the Cortex-A57 (Arm own 64-bit microarchitecture design) would not see light of day until late 2014."
You are misinformed and biased as well. Have a downvote.
from https://www.anandtech.com/show/16226/apple-silicon-m1-a14-deep-dive/2 if you want to get informed
What I mean is there is this massive marketing hype of Magical, Revolutionary, Innovation, Patents etc, and it's mostly just marketing of similar, but reasonably well made, over priced hard to repair stuff.
We'd have smartphones, touch, wearable smart watches, ARM laptops, non-Windows OS, designer hyped products, mp3 players etc, even if Microsoft hadn't bailed out Apple and Steve Jobs hadn't come back, or even if he had and the iMac hype had failed.
I know Apple has a big USA penetration and is popular with a certain well healed demographic, but it's all massively overpriced, big profit margin and not that important outside the USA. A USA fashion brand for Tech.
The M1 is lovely piece of development by Apple, but achieved simply by developing the real innovation of others in the last 30 years.
Noboy is disagreeing with you Mage when you say Apple are very good at marketing, or that their pricing structures and practises can be frustrating or even nasty.
However, the way to express that isn't to keep stating that they they don't innovate or that everything of note they've ever done is just brought-in or stolen. When you take this line of argument it *appears* that you only value innovation of individual components, as opposed to the refinement of an integrated end product. It's as if you don't see the work of people in form engineering, ergonomics, human behaviour, testing, etc as being of any worth. You come across as valuing a product's history over its function.
I'm not saying that's what you actually think, I don't want to make you a straw man. But that's just how your posts can read from time to time.
This M1 can transform what we will be offered outside of Apple. Just like the iPhone did for the smartphone. Or the app store did for so many developers globally. Even on Android.
And marketing of engineering ideas is a critical part of the battle. Without Apple, the mass appeal to keep costs down for android would not happen.
As recently Airpods dramatically change the cost and offering outside of Apple of always existing wireless earbuds.
Your definition of innovation is narrow and myopic. Your dislike for apple is blinding you to their transformational role in many of their segments - outside of their product.
It's like saying the new COVID vaccine is irrelevant as it is using the "real innvoations" of people before, becaues they didn't invent mRNA. The current scientists are just making it right for "another virus".
Everything builds on the shoulders of others.
Ultimately a real innovation is what the most people benefit from, and the judgement is the money chasing it.
Although I don't use an iPhone, thank you apple for what I get on android.
>> reasonably well made, over priced
show a comparable TCO of an Apple mainstream product and an as per your definition appropriately priced device.
The iPhone TCO is very certainly definitely lower from my analysis if you take parameters like security updates and compatibilty.
So show your analysis of better value over the *lifetime* of the product, not for specmanship.
Nice but late: Apple were offered ARM silicon for the desktop in the Eighties according to Robin Saxby. Sir Robin, who was then fresh out of Motorola and the founding CEO of Arm, said he was pushing ARM beyond Newton with Apple. He and Larry Tesler were best buddies but Apple's desktop division wouldn't budge from PowerPC which was their choice to make but imagine what could have been if Apple had made this change decades earlier. DEC's StrongARM proved it could be done, ARM became a mobile architecture because Arm followed the Nokia money.
I realise I'm raining on a parade here but the truth is what it is and anyone who doesn't like it can always engage their Apple iRDF™ :)
The hardware (scroll wheel) was an integral part of the iPod UI. I had otherwise similar MP3 players, and the lack of a scroll wheel made navigating huge lists of albums less joyful than it could have been. Abba taptap tap tap tap tap..... Taptap tap Zappa!
(I'd previously had a Sharp 722 Minidisk player with a scroll wheel, but it was used for entering text, rather than for navigating tracks since albums typically didn't exceed twenty tracks. )
You cannot still install and run on the current version of Windows anything written for Windows 3.1.
Yet it'd be hugely unfair to accuse Microsoft of "forgetting about backwards compatibility".
In fact it would be better to point out that they fail by spending too much effort on bc....
If you look at Anandtech's article referenced, they compared Geekbench and SPEC2006 results between an iPhone 12 and the latest highest end Intel and AMD CPUs. The iPhone beat everything Intel has, and everything AMD has except the just released Zen 3.
The M1 certainly won't be SLOWER than the A14 in the iPhone (and it has a bigger cache and double the memory bus width) so early adopters will be just fine even if they are expecting a low end Mac to demonstrate better single thread performance than any PC you can buy at any price - which it will other than MAYBE the top bin Zen 3.
'Select industry benchmarks'
ie we ran all of them and on many of them the chip is only as fast or slower than Intel/AMD so we selected the ones that showed a nice speed bump :).
Can't say I blame them, everyone is at it, but i'd rather wait as see some proper independent benching before deciding whether to jump ship
One question I've never seen answered is why the Intel Macs are so slow under heavy loads. Filesystem throughput and swapping goes to hell on hardware that shouldn't be breaking a sweat. And let us not stop wondering why a minor system software update takes 45 minutes on a computer that should be able to perform cryptographic calculations and filesystem I/O in the GB/sec range. It makes me think that MacOS hasn't had a good cleanup since 10.6. If OS 11 was seriously cleaned up for ARM, it think it could easily outperform Intel Macs regardless of the "M" chip speed.
This is not correct - all Macbook pros do operate correctly under the TDP envelope defined for the chip to the intel specification.
Yes you can overclock or put bigger fans. You can make a fat fan cooled tablet for more performance or even use a desktop part on a tablet for cost by that logic. But that isn't the product offering.
Most people end up comparing just the first part of the Intel CPU model but for mobile CPU intel make variants (UL, etc) with different TDPs. You will need to compare the specific CPU variants with the same TDPs when you compare performance.
I expect Apple to use the same strategy with these M1s managing TDP for performance vs form factor vs fan noise etc.
So x86-64 is not slower vis a vis, but if you want the faster mobile parts with higher tdps and thicker profiles that other notebooks offer (with higher performance), it is correct than Apple does not address this product segment.
It is true that Apple does not address the non-overheating product segment that other manufacturers somehow manage to address:
Louis Rossman: Why Macbook runs hot - "They don't care if this thing cooks itself as long as it doesn't make noise... make no mistake, this is going to fry."
Louis Rossman: Louis opens new Macbook Air, immediately loses mind. - "Why is there a heatsink over here, that doesn't go anywhere, and then a fan over here? What is the fan blowing? ... the heatsink does not reach the fan."
The sound of a suffering Macbook Air CPU with no heatpipe - "Can any of you find the heatpipe on this MacBook that connects the CPU to the fan? Do you see it? Exactly. That's why the CPU clicks."
Louis Rossman - The horrible truth about Apple's repeated engineering failures. - "The airflow is going to be directly from the burning heatsink onto the bottom of the display assembly, the very part of the machine that's glued together. So every time you turn this computer on and do something processor or graphics intensive, you're actually heating up the glue that holds these two pieces of metal together and ungluing your machine."
LTT - Macs are SLOWER than PCs. Here’s why. - "I can fire up quite literally any stress test and within moments I will have temperatures rapidly approaching 100 degrees, the point at which most Intel CPUs will throttle back their clock speeds in order to protect themselves from damage."
Etc...
Louis Rossman is a very unknowlegeable youtube "celeb" - and that celebrity is built on anti mac ness for his audience. He even f***ing sells merch on his channel - like a fans site.
I had no idea who this guy was until his fankids claimed his "cracking" find about a FET design mistake in Macbook Pro 2018. He failed to apreciate anything about TDP and that the FET was exactly to Intel's spec for the revised part. (I actually checked the Intel design spec as it seemed like an incredulous mistake for a major manufacturer with extended warranties and branding to do - Apple or not). Between some Youtube celeb with no creds and the engineers as Apple and Intel - I was thinking let's bet on the latter knowing better.
After reading the builder spec, his entire commentary turns out to be nonsense.
The macos SW update (released the next day IIRC) corrrected the bug, not a HW redesign on the macbook Pro 2018 as he claimed would be needed.
He got his 15 secs of fame aka Youtube ads and clicks. And people refer to him as having found something "unique". Yet his complete nonsense on 2018 MBP about a HW design failure seems to have no effect on his fans when it was so clearly fixed by SW and benchmark performance restored.
This rossman character never does explain how if his finding were true it was fixed in SW? The CPU would never get the voltage required for the higher clocks as per him - no SW can fix that. But lo and behold a SW update did exactly that and the higher clocks were working.
Indeed you clearly put so many of his references so you haven't picked it. You probably think the mbp 2018 still has the issue perhaps.
I have not seen the other videos nor will I watch what you post as "sources". The video I did see was BS. If it's BS in an area I do know, then the rest is BS for what I don't know. And I don't want my new knowledge coming from a known bullshiters.
This is the common problem with these Youtube "techies" - they are not engineers nor do they want to learn facts, as their focus is on learning celebrity, followers, fans, merch and clickbaits - this is for entertainment and not for fact, and you shouldn't bring these entertainers to forums lie these, worse as references for anything except on a fan site discusiing rossman episodes perhaps. (ZOMG his chair is so cool like where do I get that!!)
Someone posted a Linus (Youtube celeb) video earlier, which again tripped me up - yet again another where clickbaiting is the name of the game. I don't know if he changed his name for the fame - I seriously wouldnt put it past these celebs making season 4 of their dramedy shows.
Same with fitness celebs - yo that exercise you be doing (and everyone has and still is doing in professional training) is so so the wrong one here be why... change now with my new rotating upside down hinge flip abs crunch... click (coz no one else figured this out but me..)
OMG just realised your last reference is the other Linus character I referenced - what a unsuprising coincidence.
From a video that advertises paid promotions for HW component upgrades - I can see how Apple products might shall we say "influence" those advertisers revenues.
Get a fat thicker laptop compare to another form factor and say oops there are faster machines out there. I can do a benchmark and say alienware laptops are faster than dell ultrabooks. Epic fail dell (hmm but dell owns alienware, so no ads dang..)
Or should I do a why tablets are slower than a macbook air video? (Nah that can't sell any ads can it)
No I did not watch the whole video -
entertainment with paid promotions - OK
reviews with paid promotions - not really
reviews with paid promotions with a conflict of interest with the reviewed products - hella "get the f*$@ outta here" NO.
You shuold have post anon, coz Dan55 your credibility is rather shot since you got to own your references.
You are here arguing that Apple copy-pasting an Intel reference design and practically hermetically sealing the CPU inside a laptop of their own design which has no air flow which Intel has no control over is a good thing and anyone who complains obviously doesn't get it.
You are also arguing that undervolting the CPU in software to try and retain some performance in temperature conditions which would normally make it throttle back to below base clock speed is also a good thing instead of it being obviously another fix for the same cause - terrible thermal design.
My 2012 MacBook Pro had to be taken back (happily a few days before one year after purchase) because it has the well known GPU-frying problem. And no, it wasn't covered by the extended guarantee because they non-retina discrete graphics version is not covered by the extended guarantee.
Heck, there's someone else here saying if they want to make their Mac run hot all they need to do is plug in an external display (and give the GPU more work), plugging in an external display was never contemplated in their thermal design.
The internet is replete with stories about Apple's problems with heat, so much so that the adage about the plural of anecdote is data holds true. By the way, Louis Rossmann's job is not a YouTube celebrity but a business where he does fix people's hardware problems and he does it well otherwise he wouldn't be in business. The YouTube videos are done on the side to bring in more business.
I have no doubt they will finally get on top of their thermal design problems after more than a decade or having them, but not because they've fixed the thermal design, simply because the M1's power consumption is lower than Intel so it chucks out less heat. Over a decade of workarounds because they couldn't come up with a laptops that cool as they should (extended guarantees due to heat problems are documented so you can't handwave that away)... until they finally fixed them by changing the CPU.
>> Apple copy-pasting an Intel reference design
Now I know you don't know what you are talking about. It is not the reference design it is the builder spec. Notes for a system builder to make their *own* design
>> thermal design
You are ignoring form factor, and comparing thermal performance *of another form factor*. Can you achieve a different balance of themal design and performance. Yes. At the same form factor?
>> running hot
thermal failure and running hot are different things. If you are saying that at heavy load it should not turn hot, then well I'd say what is the point of the fan. Now for form factor reasons the Air does not have a fan, this is the design, it is not meant for prolonged max loads. What would be a failure is if shut off.
I suppose at the end of the day, you are claiming that max load must run indefinitely with no thermal backoff in *any* form factor. This is a disservice to the very, very vast majority who do not need the extra weight and thickness to support such a peak load scenario, but would appreciate the burst in performance for shorter max loads. (For that 30 minute compile). It also makes a big difference to responsiveness.
A 10 hour compile workload ought to be moved to a different form factor device. Given the technical nature likely to be involved in such sustained peak workloads, the buyer ought to be aware of their requirements and that they will have to compromise on form factor benefits for their workload requirements. I agree that the Macbook Pros (indeed almost all ultrabooks) are not suitable, but I equally am not aware at the same weight/thickness class if there are other alternatives that would handle such a sustained peak workload at cpu max clock indefinitely.
Note that these design principles are what is done for all mobile phones, android included. What Apple are doing is mobile style system builder design for laptops. What you are looking at it is the PC system builder approach, where laptops are portable desktops. You are right that Apple is not addressing this "portable desktop" segment in their macbook line.
If you ran a compile on your phone, and the clock was throttled for thermal management, do you complain it was badly designed, as it does not stay at the max cpu clock?
if rossman is a repair man he should not comment on circuit design. He repairs designs, he does not design. He can comment on the repairability of a design.
He reads the repair manual written by the manufacturer with the design engineer defining how the repair is done.
He can comment that a *repair* done used the wrong FET part, because the manufacturer service manual said so.
Claiming that a FET part, used by the manfucturer, is incorrect, is a design comment. Not a repair comment.
Best laptop CPU performance (Geekbench 4.1)
1. Asus ROG Mothership: 34,879
2. Alienware Area-51m: 32,591
3. MSI GT76 Titan: 32,167
4. Apple MacBook Pro (15-inch): 31,012
https://www.laptopmag.com/benchmarks/cpu-performance
So best in class x64 CPU performance for the form factor.
Here are some of the apps I have on my MacBook Pro, and I’m holding off Apple Silicon till I know the all work. It might sound harsh, but this is a sample of real apps people want;
Spotify,
Dropbox,
Final Fantasy 14,
Steam (plus games),
Visual Studio code,
Unity3d,
Libreoffice,
Adobe D.C.
IntelliJ,
HashiCorp tools,
VirtualBox,
Bootcamp - Dev tooling for Nintendo switch,
Homebrew.
I don’t want to but I see myself have to switch to a dual boot Windows and Linux system.
I wasn't expecting this to go live so soon, I was thinking to upgrade my 2012 MacMini as I use Parallels a lot for Windows VMs on my Mac.
I don't really see how this will be workable on ARM Macs since they can surely only run ARM guest/virtualised OS and Windows ARM is pretty niche.
Anyone else in this boat and have better intel? (!)
They didn't stop selling x86 Macs, and they'll be available on the used market for years, if running Windows x86 stuff is important to you.
Running Windows/ARM in a VM should be fine for most people though, Microsoft has built in x86 to ARM translation so you can transparently run x86 Windows applications if there isn't an ARM version (which there isn't yet for most third party stuff, but Microsoft has their own stuff ported)
Unless you need to run Windows applications that are 1) only available for x86 and 2) demand a lot of performance like CAD tools or the like you'll be fine. If those things are the case, and things don't change WRT to a Windows/ARM port (or native Mac port) in the next five years or so, then you might have cause to worry.
They didn't stop selling them yet, but they already reduced the range. The lower spec Mac Mini is gone for instance. 2nd hand is a option of course but not ideal.
On the other hand, how do you get an ARM version of Windows in the first place? I don't know you can buy it and if you can, does that mean another license purchase?
Dunno where to get it, you'd have to ask Microsoft. When Parallels Desktop is released for ARM Macs they'll presumably have a way to download it or whatever.
Obviously it would mean another license purchase, just as Microsoft requires for running a Windows VM even when run under Windows. They will always get their pound of flesh!
I too use Parallels, but these days, only occasionally (Windows XP and 10) to run Windows software I wrote over a number of years, and the occasional specialised technical application. We may be OK: Parallels on Apple Silicon.
""In June, we announced that the Mac is taking another huge leap forward by transitioning to Apple Silicon," said Apple CEO Tim Cook during a internet-streamed presentation. "And we promised that the first Mac with Apple Silicon would arrive by the end of this year. Well, that day is here.""
Great news, Cookie ! Now enjoy your 2 grands laptops with 0 performance, barely 8GB of RAM, 256 GB of disk space, but having 20 hours of battery life !
""The transition to Apple Silicon is going to have a profound impact on the Mac," said John Ternus, Apple VP of hardware engineering."
Sure, mate ! Sure ! The first one will be no virtualization of Windows what so fucking ever ! Brilliant ! Carry on this strategy !
Gawd, I've gone from '040, to PowerPC, to Intel, now this. I'm really tired of the BS that goes along with these transitions. *rolls eyes*
The GOOD NEWS is, that this M1 chip sounds like it can really deliver efficiently, but I wish I knew this was on the horizon about a year and half ago. Looks like some really solid performing new-ish Apple Intel kit of mine could get sidelined prematurely. Bugger!
Intel and AMD are toast, period.
Even under Rosetta 2 emulation, apps are ALREADY faster than any current Mac in native mode. As for native apps, even a single ARM iPad is already faster than virtually any Winblows desktop out there - the M1 will extend that lead so much that Intel/AMD will probably file for Chapter 11 protection in a couple of years.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
