Hacker News new | past | comments | ask | show | jobs | submit login
Apple unveils M1, its first system-on-a-chip for portable Mac computers (9to5mac.com)
1365 points by runesoerensen on Nov 10, 2020 | hide | past | favorite | 1346 comments




Completely unconfirmed speculation incoming:

There's a solid chance that the logic board is exactly the same on all of the Macs announced today and the only difference is the cooling solution. If you play around with the Apple Store configurator, the specs are all suspiciously similar between every new Mac.

https://www.apple.com/shop/buy-mac/macbook-air

https://www.apple.com/shop/buy-mac/macbook-pro/13-inch

https://www.apple.com/shop/buy-mac/mac-mini


At Apple's volume and level of system integration, it doesn't make sense to do assembly sharing at that level between different models. Presumably the SoC package is the same between the different products, but binned differently for the Air, Pro, and Mini. The actual logic boards would be custom to the form factor.


Not just that. At 5nm there will also be yield problems. I.e they will put the best yield into high end and the worst yield into low end.


This is undoubtedly why they launched the Mac Mini today. They can ramp up a lot more power in that machine without a battery and with a larger, active cooler.

I'm much more interested in actual benchmarks. AMD has mostly capped their APU performance because DDR4 just can't keep the GPU fed (why the last 2 generations of consoles went with very wide GDDR5/6). Their solution is obviously Infinity Cache where they add a bunch of cache on-die to reduce the need to go off-chip. At just 16B transistors, Apple obviously didn't do this (at 6 transistors per SRAM cell, there's around 3.2B transistors in just 64MB of cache).


Okay, probably a stupid question, but solid state memory can be pretty dense: why don't we have huge caches, like a 1GB cache? As I understand it, cache memory doesn't put off heat like the computational part of the chip does, so heat dissipation probably wouldn't increase much with a larger chip package.


Nand flash is pretty dense, but way too slow. Sram is fast but not at all dense, needing 6 transistors per bit.

For reference: https://en.m.wikipedia.org/wiki/Transistor_count lists all the largest cpu as of 2019 amd's epyc Rome at 39.54 billion MOSFETs, so even if you replaced the entire chip with Sram you wouldn't even quite reach 1GB!

Dram would be enticing, but the details matter.


Nand is nonvolatile and the tradeoff with that is write cycles. We have an inbetween in the form of 3D Xpoint (Optane), Intel is still trying to figure out the best way to use it. It currently like an L6 cache after system DRAM.


Well not just Intel. Optane is a new point in the memory hierarchy. That has a lot of implications for how software is designed, it's not something Intel can do all by itself.


SRAM is 6 transistors per bit, so you're taking about 48 billion transistors there, and that's ignoring the overhead of all the circuits around the cells themselves.

DRAM is denser, but difficult to build on the same process as logic.

That said, with chiplets and package integration becoming more common, who knows... One die of DRAM as large cache combined with a logic die may start to make more sense. It's certainly something people have tried before, it just didn't really catch on.


> DRAM is denser, but difficult to build on the same process as logic.

What makes it so difficult to have on the same chip?


I don't know the details, but the manufacturing process is pretty different. Trying to have one process that's good at both DRAM and logic at the same time is hard, because they optimize for different things.


Or, well, how about putting all your ram on the package as Apple says they are doing with the M1?


Cost. Area is what you pay for (at a given transistor size).


the bigger the cache the slower it gets

so ram is your 1gb cache


Are you are referring to latency due to propagation delay where the worst case increases as you scale?

Would you mind elaborating a bit? I'm not following how this would significantly close the gap between SRAM and DRAM at 1GB. Since an SRAM cell itself is generally faster than a DRAM cell, and I understand that circuitry beyond an SRAM cell itself is far simpler than DRAM. Am I missing something?


Think of a circular library with a central atrium and bookshelves arranged in circles radiating out from the atrium. In the middle of the atrium you have your circular desk. You can put books on your desk to save yourself the trouble of having to go get them off the shelves. You can also move books to shelves that are closer to the atrium so they're quicker to get than the ones farther away.

So what's the problem? Well, your desk is the fastest place you can get books from but you clearly can't make your desk the size of the entire library, as that would defeat the purpose. You also can't move all of the books to the innermost ring of shelves, since they won't fit. The closer you are to the central atrium, the smaller the bookshelves. Conversely, the farther away, the larger the bookshelves.

Circuits don't follow this ideal model of concentric rings, but I think it's a nice rough approximation for what's happening here. It's a problem of geometry, not a problem of physics, and so the limitation is even more fundamental than the laws of physics. You could improve things by going to 3 dimensions, but then you would have to think about how to navigate a spherical library, and so the analogy gets stretched a bit.


Area is a big one. Why isn't L1 MB? Because you can't put that much data close enough to the core.

Look at a Zen-based EPYC core- 32KB of L1 with 4 cycle latency, 512KB of L2 with 12 cycle latency, 8MB of L3 with 37 cycle latency.

L1 to L2 is 3x slower for 8x more memory, L2 to L3 is 3x slower for 16x more memory.

You can reach 9x more area in 3x more cycles, so you can see how the cache scaling is basically quadratic (there's a lot more execution machinery competing for area with L1/L2, so it's not exact).

https://www.7-cpu.com/cpu/Zen.html


I am sure there are many factors, but the most basic one is that the more memory you have, the longer it takes to address that memory. I think it scales with the log of the ram size, which is linearly with the number of address bits.


Log-depth circuits are a useful abstraction but the constraints of laying out circuits in physical space imposes a delay scaling limit of O(n^(1/2)) for planar circuits (with a bounded number of layers) and O(n^(1/3)) for 3D circuits. The problem should be familiar to anyone who's drawn a binary tree on paper.


With densities so high, and circuit boards so small (when they want to be), that factor isn't very important here.

We regularly use chips with an L3 latency around 10 nanoseconds, going distances of about 1.5 centimeters. You can only blame a small fraction of a nanosecond on the propagation delays there. And let's say we wanted to expand sideways, with only a 1 or 2 nanosecond budget for propagation delays. With a relatively pessimistic assumption of signals going half the speed of light, that's a diameter of 15cm or 30cm to fit our SRAM into. That's enormous.


Well, the latest AMD Epyc has 256 MB L3 cache, so we're getting there.


but any given core only has access to 16 soon to be 32mb


in Zen 1 and Zen 2, cores have direct or indirect access to the shared L3 cache in the same CCX. In the cross-CCX case the neighboring CCX cache can be accessed over the in-package interconnect without going through system DRAM.


this!

when i started with computers, they had a few KB of L2 cache, L3 did not exist. Main Memory was a few MB.


With the DRAM this close, it can probably be very fast. Did they say anything about bus or bandwidth?


AnandTech speculates on a 128-bit DRAM bus[1], but AFAIK Apple hasn't revealed rich details about the memory architecture. It'll be interesting to see what the overall memory bandwidth story looks like as hardware trickles out.

[1] https://www.anandtech.com/show/16226/apple-silicon-m1-a14-de...


Apple being Apple, we won't know much before someone grinds down a couple chips to reveal what the interconnections are, but if you are feeding GPUs along with CPUs, a wider memory bus between DRAM and the SoC cache makes a lot of sense.


I am looking forward to compile time benchmarks for Chromium. I think this chip and SoC architecture may make the Air a truly fantastic dev box for larger projects.


That’s what “binned differently” means btw.


Interesting that commenter knew the process but not the terminology.


As someone who works with a lot of interdisciplinary teams, I often understand concepts or processes they have names for but don't know the names until after they label them for me.

Until you use some concept so frequently you need to label it to compress information for discussion purposes, you often don't have names for them. Chances are if you solve or attempt to solve a wide variety problems, you'll see patterns and processes that overlap.


Seconded.

It’s often valuable to use jargon from another discipline in discussions. It sort of kicks discussions out of ruts. Many different disciplines use different terminology for similar basic principles. How those other disciplines extend these principles may lead to entirely different approaches and major (orders of magnitude) improvements. I’ve done it myself a few times.

On another note, the issue of “jargon” as an impediment to communication has led the US Military culture to develop the idea of “terms of art”. The areas of responsibility of a senior officer are so broad that they enter into practically every professional discipline. The officer has to know when they hear an unfamiliar term that they are being thrown off by terminology rather than lack of understanding. Hence the phrase “terms of art”. It flags everyone that this is the way these other professionals describe this, so don’t get thrown or feel dumb.

No one expects the officer to use (although they could) a “term of art”, but rather to understand and address the underlying principle.

It’s also a way to grease the skids of discussion ahead of time. “No, General, we won’t think you’re dumb if you don’t use the jargon, but what do you think of the underlying idea...”

Might be a good phrase to use in other professional cultures. In particular in IT, because of the recursion of the phrase “term of art” being itself be a term of art until it’s generally accepted. GNU and all that...


Where can I learn more about the US Military culture's "terms of art" idea?


> How those other disciplines extend these principles may lead to entirely different approaches and major (orders of magnitude) improvements.

Fascinating. Would you happen to have any example off the top of your head?


But then how will developers exert their local domain dominance to imply and an even greater breadth of knowledge when patronizing other devs? /s


I have always assumed this term was in widespread and general use in the US business culture. Is that not the case?


This gets even more fun when several communities discover the same thing independently, and each comes up with a different name for it.

My favorite is the idea of "let's expand functions over a set of Gaussians". That is variously known as a Gabor wavelet frame, a coherent state basis [sic], an Gaussian wave packet expansion, and no doubt some others I haven't found. Worse still, the people who use each term don't know about any of the work down by people who use the other terms.


Reminds me of the Feynman story about knowing something vs knowing the name of something :-)


Reminds me of self-taught tech. I’ll often know the name/acronym, but pronounce it differently in my head than the majority of people. Decades ago GUI was “gee you eye” in my head but one day I heard it pronounced “gooey” and I figured it out but had a brief second of “hwat?” (I could also see “guy” or “gwee”.) It’s, of course, more embarrassing when I say it out loud first...


First time I went to a Python conference in SV, more than a decade ago, I kept hearing "Pie-thon" everywhere, and had no idea what the hell people were talking about.

I took me a solid half hour to at last understand this pie-thingy was Python... in my head I had always pronounced it the French way. Somewhat like "pee-ton", I don't know how to transcribe that "on" nasal sound... (googling "python prononciation en francais" should yield a soundtrack for the curious non-French speakers).


I thought numpy was pronounced like it rhymes with bumpy for a year or so.


Picture 18 year old me in 1995, I got a 486SX laptop as a graduation present out of the blue from my estranged father. I wanted to add an external CD-ROM to it so I could play games and load software for college, and it had a SCSI port. I went to the local computer store and asked the guy for a "ess see ess eye" CD-ROM drive, he busted out laughing and said "oh you mean a scuzzy drive?" Very embarrassing for me at the time but that's when I learned that computer acronyms have a preferred pronunciation so I should try to learn them myself to avoid future confusion.


> Very embarrassing for me at the time

it shouldn't be, it should be a badge of honor of some sorts - it points to somebody reading to expand their knowledge that is not available in oral form around them, so kudos to them !


It's even more visible in non-English speaking countries. In Poland: first everyone says Java as Yava and after a while they start to switch to a proper English pronunciation. Many times it divides amateurs from professionals, but I wouldn't really know, because I don't work with Java.



Not the one I was thinking of but same point :) https://fs.blog/2015/01/richard-feynman-knowing-something/


Great story, yes. But there's no such thing as a "halzenfugel" in German as far as I can tell as a native speaker. Even www.duden.de, the official German dictionary, doesn't know that word ;-0


That's OK, AFAICT there's no bird called the "brown throated thrush" either.


As a native English speaker and middling foreign-language speaker of German, "halzenfugel" sounds to me like a mock-German word that an English speaker would make up.


Hah, good to know. However, unless you are talking to people from the same domain, it's usually a better approach to spell out things instead of relying on terminology. Concepts and ideas translate much better across domains than terminology.


I think a bunch of people learnt a new thing from your comment, so it is a good one.

I hope my reply didn’t come out as gatekeeping, it was genuinely just to help put a name to a thing.


May just have skimmed GP and missed it.


Well then that was a good explanation because I didn’t know that!


That's now how yield works. Yield is the number of functioning chips that you pull out of a wafer.

I think what you are trying to refer to is frequency binning.


That's only partially true.

For example, AMD sells 12 and 16 core CPUs. The 12 core parts have 2 cores lasered out due to defects. If a particular node is low-yield, then it's not super uncommon to double-up on some parts of the chip and use either the non-defective or best performing one. You'll expect to see a combination of lasering and binning to adjust yields higher.

That said, TSMC N5 has a very good defect rate according to their slides on the subject[0]

[0] https://www.anandtech.com/show/16028/better-yield-on-5nm-tha...


Which is likely why there are some "7 core" GPU M1 chips.


Yep for the MBA. I think for devs that can live with 16GB, the cheaper 7GPU MacBook Air is very interesting instead of the MacBook Pro for $300 cheaper.


Plus, defects tend to be clustered, which is a pretty lucky effect. Multiple defects on a single core don't really matter if you are throwing the whole thing away.


Is that not what the parent comment said? I thought "binning" referred to this exact process.


Isn’t that what “binning” means?


If you compare the M1 Air and Pro, the only difference seems to be the addition of the Touchbar, 10% better battery life, and a "studio" speaker/mic on the Pro.

https://www.apple.com/mac/compare/

I assume the addition of a fan on the Pro gives it better performance under load, but there doesn't seem to be a hugely compelling reason to not just get the Air.


I think they got it wrong. I would pay money to NOT have the touchbar.


I got one of the MBP with the touchbar this year after holding out for many years (changed jobs so had to change laptop). Man it is so much hard to do the volume changes, and there has so far for me been zero benefit.


Not what you're looking for, but I'll mention it anyways just in case:

It's possible to set the default touch bar display to only ever show the expanded control strip (System Preferences > Keyboard > Keyboard > Touch Bar shows: Expanded Control Strip). In that mode you tap volume up and down instead of using a volume slider.

Again, I know you're looking for physical keys (aren't we all) but it's better than nothing.

I've been using the MacBook Pro 16 (with a physical esc key plus a touch bar) and I think it's a pretty good compromise between me who wants physical keys and apple who wants to push the touch bar.

The other thing that kept happening to me: I would accidentally tap the brightness button when reaching for ESC. For that, you can "Customize Control Strip..." and remove individual buttons, so that there's a big gap on the touch bar near the ESC key so that stray taps near ESC don't change the brightness.


I realise I'm an outlier here but I actually have grown to like the touchbar.

It's often unused, yes, but when I fire up Rider for my day job it automatically flicks to the row of function keys and back depending on which app has focus and it fits quite nicely for me between work and entertainment (I like having the video progress bar if I'm watching something on the laptop). Maybe I'm just strange but the non-tactile function keys didn't really bother me much either.

In any case, I could live without it, which is probably not a roaring endorsement in any case, but I'd rather have it than not.


I like it as well, especially in applications like CLion/IntelliJ which have tons of keybindings I keep forgetting because they are different between Linux and macOS. The context-sensitive touch bar is actually very useful in these applications for things like rebuilding, changing targets, stepping through the debugger etc. without having to use the mouse.

There's a lot of things to complain about with Apple products, but if you ask me there's been enough touch bar bashing by now and people should just get over it. It's pretty useful in some situations, and IMO no real downsides, especially now that the esc key is a real physical key again. Why all the hate?


Physical F keys are still useful so why not both?


If you hold Fn you get the traditional row of function keys, which seems like a pretty good tradeoff. And if you really hate that, you can simply configure the touch bar to always display them, in which case literally the only downside is that they are not physical keys anymore. Do people really touch-type the function keys so heavily that this becomes a an actual annoyance and not just an ideological one?

Adding an extra row of physical keys that do the same thing as the row of virtual function keys, at the expense of trackpad size and possibly ergonomic (harder to reach the touch bar) doesn't make a lot of sense IMO.


You can’t hit them without looking at the bar, because you have nothing to index your fingers on.

The touchbar is the second worst thing Apple has ever done in the history of the Mac, following closely on that abomination of a “keyboard” they used from 2016-2019.


I liked the keyboard too.


And if you ask me, It has not been enough touch bar bashing...

I’ve opted to buy a 2015 Mac Book Pro this year, it might be easier to get over apple than the touch bar even...


The only time that there would be enough touch bar bashing is when Apple listens and give users an option to have function keys instead.


Pretty sure they just did. The MacBook Air is virtually identical now save for the absence of a fan and the Touch Bar.


My guess - that you couldn't get a Macbook without the TouchBar. I'd like to be able to choose between a Macbook with or without a TouchBar, but with otherwise entirely identical specs.

I've been holding out on upgrading my 2013 MBP (mostly out of frugality) to a newer version, mostly due to the butterfly keys and the TouchBar.


Yup, me too. Especially when debugging in Xcode, the TouchBar shows a bunch of useful functions.


Those are f keys on a regular keyboard, and a few million of us have developed the muscle memory to use them over the say last thirty years that they’ve been assigned to those f keys in most IDEs.


I’m with you: I enjoy the TouchBar, and it’s genuinely useful for some apps I use


This is one option, but it still suffers from my main complaint about the touch bar -- it's way too easy to accidentally activate something. Therefore, I have my touch bar set to activate only while I'm holding down the FN key.

I will not remove that safety key until they make the touch bar pressure sensitive so that "buttons" on it only activate with a similar amount of force that was required to activate the tactile buttons they replaced. Until then, I consider it a failed design improvement.


I need my ESC, so I'm glad it's there. As for the rest of the keys on the top row, I was not in the habit of using them except in vim, where I hooked them up to some macros I had written. For them, I kind of like the touchbar now, because I have the fake keys labelled with sensible names. (No more trying to remember that I have to hit F3 to do such-and-such.)

I've also found the touchbar pretty useful in zoom calls, because my zoom client shows keys for common actions.

All in all, I think a physical escape key plus the touchbar is a slight win. I would not pay more for it, but I have reversed my previous opinion that I'd pay more not to have it.

I suspect these new machines are going to be quite nice, although I won't buy one for a while since I purchased a mbp a few months ago.


I don't understand why they don't put the physical keys AND the touchbar in. There is so much space on the 16" model they could easily fit it in and shrink the obscenely large trackpad just a touch.


I think that the trackpad needs to be bigger. Doing gestures is much easier with a big pad.


Especially on the 16 there is no excuse, they could really have a function key row and a touch bar :( more than enough space for them.


I ordered my 13" touchbar MBP with the UK keyboard layout. Adds an extra key to the right of left shift (mapped to tilde), letting me remap the key that is normally tilde on a US keyboard to ESC.


I've been really happy with the following mod for the last couple years of TouchBar usage:

https://community.folivora.ai/t/goldenchaos-btt-the-complete...

Fully customizable while being much better for muscle memory by giving you exactly what you want where you want it, gives you icon-shortcuts to script, and still allows you to have as much dynamic functionality / information as you like. So, for example, mine looks roughly like this:

- Fullscreen

- Bck/[play|pause]/Fwd

- CURRENTLY_PLAYING_SONG

- AirDrop

- ConfigMenu

- Emoticons

- (Un)Caffeinate

- (Dis)connectBluetoothHeadphones

- (Dis)connectMicrophone

- (Un)muteVol

- VolDown

- VolUp

- ScreenDim

- ScreenBright

- Date

- Time

- Weather

- Battery%

CURRENTLY_PLAYING_SONG playing shows the album cover, song name, and artist, but only shows up if there IS something playing. Same with AirDrop, which shows up only if there's something that I could AirDrop to, and then gives me a set of options of who to AirDrop to. The Emoticon menu opens an emoticon submenu on the TouchBar with most-recently-used first.

That all fits fine into the main touchbar, with other dynamic touchbars available modally (ie, holding CMD shows touchable icons of all the stuff in my Dock (my Dock is entirely turned off)), CTRL shows LOCK AIRPLAY DO_NOT_DISTURB FLUX KEYBOARD_DIM/BRIGHT, etc. ALT shows me various window snap locations.

Edit: BetterTouchTool also replaced a bunch of other tools for me. Gives you the same kind of tools for scripting eg Keyboard macros, Mouse macros, remote-control via iPhone/Watch etc with a lot of reasonable defaults.


I've heard a lot of complaints about the touchbar. The loss of tactile feedback is a fair one, and admittedly removing the escape key was a terrible idea. I recently upgraded to a machine with a touchbar, learned quickly why the default settings are pretty bad, and then quickly found BTT and set it up. The touchbar is not a revolutionary innovation, but it definitely improves functionality in some cases, and it's fun to mess with. Oh, and a button to "toggle mic in zoom" actually solves a real problem.

The people who complain about the touchbar functionality must not be putting any effort at all into it. I customize so many other things on my system, regardless of the OS. Why would a new hardware feature be any different?

I didn't know about this GoldenChaos thing though, thanks for that.


> The people who complain about the touchbar functionality must not be putting any effort at all into it.

I would say that people who complain about uselessness of F-keys must not have put any at all effort into using them.

Upon getting my MBP 2016, I spent numerous months trying to make the TouchBar useful; from customizing the contents where apps allowed it, to BTT.

What it came down to is that things worthy a keyboard shortcut are things I want to be able to do fast, reliably, and instinctively. I don't want to search for the button on the TouchBar – I'm using the keyboard, it needs to be as natural as typing, without the need to look down at it. I have a screen already, I don't need another one on my keyboard.

I firmly believe TouchBar can't even come close to the same realm of usefulness as F-keys, much less being worth the price hike it imposes. Function keys are twelve, tactile, solid, free, reliable(1) buttons for keyboard shortcuts; TouchBar is a touchscreen that sometimes(2) works.

> a button to "toggle mic in zoom" actually solves a real problem

I haven't used Zoom, but if it's a decent-ish Mac app, it either already has a keyboard shortcut to toggle microphone, or you can set one in Keyboard Shortcuts, in System Preferences.

(1) as far as anything is reliable on the butterfly keyboards.

(2) same story as butterfly keyboard – if it's even slightly sporadic, it is a shitty input device.


That's fair. Personally, I used the tactile function keys for exactly six things (brightness up/down, volume up/down, mute, play/pause). Those six functions are now available on my touchbar. They're no longer tactile, which, yes, is a minor inconvenience. I wouldn't use a touchscreen for touch-typing code, of course. But for a handful of buttons way off the home row, it doesn't affect me much.

In exchange, I get to add other buttons which can also display customizable state. Yes, zoom has a global shortcut to toggle the mic. The problem is, the mic state is not visible on your screen unless the zoom window is visible. This is a frustrating design flaw IMO, which really should be addressed in some consistent way across every phone/video app. But, it's not, and so I need to pay attention to my mic state. My touchbar button toggles the mic, and displays the current mic status. I would imagine that every phone/video chat app is similar.


I don't understand how Apple missed the boat so completely on haptic feedback for touchbar, considering their mastery of it on trackpads and touch screens.

I use this app alongside BTT; it attempts to supplement haptic feedback via the trackpad haptics. It's no where near as good as a real haptic solution would be but does provide some tactile feedback when pressing buttons https://www.haptictouchbar.com/


I have a haptic touchbar and I’m pretty sure it was enabled via OOTB with GoldenChaos on BTT, but maybe not.


Did you notice you can hold and slide to change volume? You don’t need to hit the volume slider where it appears. Same with brightness. Totally undiscoverable gesture.


Yup - I love this feature but I'd guess based on people I've shown it to that no more than 20% of users are aware of it.


Exactly this. I tried so hard to like it (since I paid for it), but I have found 0 good uses cases for it.

I would assume that macOS sends at least some basic usage data for the touch bar back to Apple HQ. I wonder how often it is actually used... and I would love the hear the responsible product manager defend it.


Same situation, my trusty 2012 rMBP finally gave up the ghost and I had to get a new one with this icky touch bar. It's useless to me and makes it harder to do everything. My main complaint is that I am constantly bumping it when I type, leading to unexpected changes in volume and brightness.


Oh yeah I forgot that. I keep hitting the chrome back button in the touch bar all the time. In the beginning I was not sure what was happening then I realized it was the touch bar.


My problem with the touchbar is that I tap it accidentally while typing all the time. It needs to be like another centimeter away from the keys.


Or use the same haptic feedback as the touchpad.


Haptics plus requiring a bit of pressure to register a press, just like the trackpad.



That's actually one of the things I like better on the touchbar. Just press down on the volume icon and slide.

I continue to be disappointed about the lack of haptics though. It's such a strange thing to not have when they've proven to know how to make very convincing haptic feedback. It works very well in both the Apple Watch and the MacBook's trackpad.


You CAN configure it to show the old-style mute/down/up with the touch bar, so you are not relegated to the ultra-shitty slider. No replacement for a tactile switch, but at least you are not stuck with the default arrangement.


Easiest way is to press and hold the Touch Bar on the volume control button and slide your finger left or right–that way you barely need to look at the Touch Bar.


You can use the touchbar to skip commercials on Youtube in Safari.

I love it.


Instead of press and hold, it's press, hold, and drag. Definitely annoying when it freezes, but when it's working it doesn't seem that much different.


The main difference is that I need to look down at the keyboard to operate the touchbar. With the keys I can rely on muscle memory.

Also I think every device which makes sound should have a physical mute control. The worst is when I want to mute, and the touchbar freezes, and I have to go turn the volume down with the mouse.


I intentionally took a performance hit by moving from a Pro to an Air almost entirely for this reason (although the low power and light weight are pleasant benefits). I'm glad that the new Air still has F-keys with Touch ID; but I'm flabbergasted that they're still not making the Touchbar optional for the Pro series, given how polarizing it's been, and the underwhelming adoption of new Touchbar functionality by third-party developers.


They brought back the escape key which is what really matters.


Honestly, I think it's only polarizing here and among some developers.


I mean if you think about other professions that use Macbook Pro they don't need it either. Are video professionals using it to scrub through video? Nope. For any audio professional it's useless. No one who is serious about their profession would use it.


I’m serious about my profession and I use it daily. Wonder what that says about me.


Please tell me what it can do that hot keys and just generally knowing a program can't do? I'm honestly interested.


I'd be curious to know what portion of their user base for the Pro series are developers. Anecdotally, quite a lot of devs seems to use Macs; but I have no idea what that fraction is, relative to the rest of their market.


The touchbar is probably why I'm probably getting the Air for my next upgrade, and not a Pro.


Honestly, the idea of a soft keyboard is a great one, particularly one as well integrated into the OS as the Touch Bar is. However, while I have a Mac with the Touch Bar, I never ever ever ever use it intentionally, as I spend 90% of my time on the computer using an external keyboard.


Just put that control panel at the bottom of the screen. It would still be close enough, and it would be out of the way of my accidental touches.


My daughter loves it. It's her emoji keyboard


As someone using a hackintosh considering a real Macbook, what's so wrong about it?


There are loads of rants out there that are easy to find, but personally it's mostly: you can't use it without looking at it to make sure the buttons are what you think they are (nearly always context-sensitive, often surprising when it decides it's a new context), and where you think they are (can't go by feel, so you need to visually re-calibrate constantly). Button size and positioning varies widely, and nearly all of them have a keyboard shortcut already that doesn't require hand movement or eyes (or at worst previously had an F-key that never moved).

The main exception being things like controlling a progress bar (mouse works fine for me, though it's a neat demo), or changing system brightness/volume with a flick or drag (which is the one thing I find truly better... but I'd happily trade it back for a fn toggle and F keys). But that's so rarely useful.


When I watch non-HN type people use it, they like it. They never used Fn keys in the first place.

I just hated the lack of ESC key (which they brought back, though my Mac is older). I have no muscle memory for any other key in that row.


I think the touchbar was my favorite part of my old MBP, specifically because of the contextual buttons that are always changing.

I'd probably pay a little extra to get one on future non-Mac laptops, but not too much extra.


Yeah, most people I know almost never use F keys (except perhaps F1 for help). They leave it on the media-keys mode... which is the same as the touchbar's default values, but without needing to know what mode it's in.

With the physical media keys, if they want to mute, it's always the same button. Pause music, always the same button. They're great in a way that the touchbar completely destroys.

(and yes, I know you can change this setting, but if we're assuming non-techy-users we also generally have to assume default settings.)


Honestly, I hardly ever used the function keys either. As a result the Touch Bar doesn't really bother me -- but neither does it seem the slightest bit useful for the most part.


Lot of non-hn people also type while looking at the keyboard and some with single finger from both hands.


It's just not useful. The context-aware stuff is too unpredictable, and I'm never looking at the keyboard anyway so I have never learned it. So the touchbar ends up being just a replacement for the volume and brightness keys, but a slow and non-tactile version of them


For me at least (and I'd imagine most of the other folks who hate it) - I had the key layout memorized. If I wanted to volume up/down/mute I could do it without taking my eyes off the screen. With the touchbar ANYTHING I wanted to do required me to change my focus to the touchbar - for the benefit of...?

I'm sure someone somewhere finds it amazing, but I have no time for it.

To me it's no different than volume controls in a car. I've been in a cadillac with a touchbar for volume, and a new ram truck with a volume knob - there's absolutely no room for debate in my opinion. One of these allows me to instantly change the volume 100% up or down without taking my eyes off the road. The other requires hoping I get my finger in just the right spot from muscle memory and swipe enough times since there's 0 tactile feedback.


For me it hides the things I use all the time (media and volume controls) to make room for application specific controls that I never use.

If it was more customisable I wouldn't mind it, but the apparant inability to force it to show me the things I actually want is annoying.

I can imagine there are some people for whom the application specific buttons are useful, but for me they are not worth it for what they displace.


Not sure if it's helpful for you but you can customize the behavior by going to your System Prefs > keyboard settings and toggling "Touch bar shows:".

I did this on like day 2 of having my MBP for what sounds like the same reason you want to. The setting I have turned on is "Expanded control strip" and I never see any application-specific controls, only volume, brightness, etc.


Omg, thank you. Somehow I'd missed that setting.


Check out BetterTouchTool if customization is holding you back.


FYI you can customize it, and force it to always display certain controls.

I had the exact same frustrations as you. Took me 10 mins digging into settings to figure it out. Now I have my touchbar constantly displaying all of the controls that are buttons on the Air (ie a completely over-engineered solution to get the same result)


The latency on the touch bat is terrible. Perhaps 1/2 second to update when you switch apps, for example!


It doesn't provide tactile feedback.


https://www.haptictouchbar.com/ is a great app I use, provides haptic feedback for the touch bar.



They do such a good job on the iPhone with this that it is quite mystifying why not.


In addition to what everybody else said, because the touch at is flat and flush with the body chassis, I find it’s very easy to accidentally press when you press a key in the row below. Eg, accidentally summoning Siri when pressing backspace, muting the audio when pressing “=“. And then you’re forced to look down and find the mute “button” to fix it.


No haptic feedback, mainly. A lot better with a real escape key, but still.


hahaha 2 times Air user, touchbar is a big NO.


Air doesn't have a fan, so if you want consistent performance, you have to buy the touchbar.


Maybe some kind of clip-on, thin, third party cooling solutions will become a thing?


so true! i hate the touchbar. if i would have to change my laptop today, i'd buy air just because i hate the touchbar.


I think it’s good actually


My Y-series Pixelbook with passive cooling performs as well as a U-series laptop from the same generation -- until I start a sustained load. At that point, the U series systems pull ahead. Actively cooled Y-series systems get pretty close in lots of applications only falling short due to half the cache.

If you are doing lightweight stuff where the cores don't really need to spin up, then they'll probably be about the same. Otherwise, you'll be stuck at a much lower base clock.


Surely they have different clock rates, but Apple isn't referencing clock rates in their marketing material for Macs with M1 chips.


Yeah I am surprised more people are ignoring this. If the two computer models have identical chips, then why does Apple even bother putting a fan in the pro?

To me the fact that the Air is fanless and the pro has a fan would indicate to me that they have different clock rates on the high end. I am sure the Air is capped lower than the Pro in order to make sure it doesn't overheat. It is probably a firmware lock, and the hardware is identical. But once we do benchmarks I would expect that the pro outperforms the air by a good margin. They added a fan in the pro so that it can reach higher speeds.

Surely the Air is capped so that users don't ruin their computers by running a process that overheats the computer.

But of course Apple doesn't want to reveal clock speeds. The best they can give us is "5x faster than the best selling PC in its class". What does that mean? The $250 computer at Walmart that sells like hotcakes for elementary age kids that need a zoom computer or the Lenovo Thinkpad Pro that business buy by the pallet? Who the hell knows.


They said in the presentation for sustained workloads. I get the impression they're the same peak clock speed but the air throttles faster.


The fan is only there for sustained loads. They all have identical chips (aside from the 7 core gpu option Air). They all hit the same pstates and then throttle accordingly due to temperature.

The MBP and Mini are there for people who want maximum sustained performance.


I recently got an Air after using a MBP13.

Aside from the loud fan and slow performance which should be fixed in this release, my biggest complaint is that they only have the usbc plugs on one side of the laptop.

Really obnoxious when the outlet is in a difficult spot.

Unclear whether the new MBP13 also has this problem...

Edit: the new M1 MBP13 has both usbc ports on the same side. No option for 4 (yet). Ugh.


The two-port and four-port 13" MacBook Pros have been separate product lines since their introduction. This new A1 MBP only replaces the two-port version. Presumably the higher-end one will share an upgraded processor with the 16".


I'm confused of their new pricing scheme / spec tiers for Macbook Pros.

There's no more core i7 for Macbook 13. You have to go to Macbook 16. I'd rather get a Dell XPS or other Core i7/Ryzen 7 ultrabooks.

So now, spec-wise, Macbook Air and Macbook Pro are too close.


I'm guessing the MBP13 is now a legacy model, being refreshed more to satisfy direct-product-line lease upgrades for corporate customers, than to satisfy consumer demand.

Along with the MBP16 refresh (which will use the "M1X", the higher-end chip), we'll probably see the higher-end MBP13s refreshed with said chip as well, but rebranded somehow, e.g. as the "MacBook Pro 14-inch" or something (as the rumors go: same size, but less screen bezel, and so more screen.)

And then, next year, you'll see MBP14 and MBP16 refreshes, while the MBP13 fades out.


These are transitional products so it makes sense. I'm looking forward to see the replacement of the iMac Pro and Mac Pro. Will be interesting to see what those include.


Addendum: just got an email from the Apple Store Business Team about the MBP13. Here's the copy they're using (emphasis mine):

> "Need a powerful tool for your business that’s compatible with your existing systems? Let’s talk. We’ll help you compare models and find the right Mac for you and your team."

That's a corporate-focused legacy-model refresh if I've ever seen one.


Specs don’t tell you the thermal story. You can buy an i9 in a thin laptop and feel good about it until it throttles down to 1GHz after 30s.

The MBP should be built with better thermals to avoid throttling since you might be running simulations or movie encoding all day. The air should throttle after a certain amount of time.


And prioritize function over form? I think you just want to buy a PC.


This has to be the funniest take on the release of a whole new CPU architecture.


I agree, I use a small form factor desktop box which fits in my messenger bag.


    There's no more core i7 for Macbook 13
Sure there is – you just have to select one of the Intel-models and customize the processor.


You are correct. They hid the option.


it's there, you just have to configure it


I think MacBook Air is very compelling, and that's why got more screen time. Unless you want to run your system on >60% for extended periods of time - MacBook Air should be great.


Pro has one extra GPU core as well.


The base model Air has 7 GPU cores instead of 8, but the higher models have all 8 cores. Seems to be +$50 for the extra GPU core.


Note that it's a 7 core GPU only for the 256GB SSD


I am curious - my 2017 12" MB is an absolutely awesome machine (fast enough for casual use and light development while absolutely quiet and lightweight), but 30+ degree summer day is enough for it to get so close to it's thermal envelope that it throttles down to cca 1 GHz during normal (browser) use soon.

So, the sustained performance might make quite a difference for pro use.


It shouldn't really throttle to 1ghz - it's because Apple puts low quality paste in it, and sometimes because of dust.

My Macbook from 2014, is still excellent but it started throttling to 1ghz with video encoding.

After going to a repair shop and telling them about the problem they put high quality thermal paste in it for about 100 usd and the problem disappeared. Now i get 100% CPU no matter what i throw at it, pretty incredible with a computer from 2014!

Just fyi..


Thanks for your experience, but the passively cooled 12" Macbook is really a different thing. Basically, it isn't a 2.5 GHz CPU that throttles, but a 1.1 GHz CPU which can boost itself for a very short time, and the it runs into thermal limit and returns to a 1.1 GHz sustained performance.

And on hot day, that boost might last 30 seconds and then that's it.


My guess is different clock speeds, also base air has one gpu core less...


And this might be the old production trick where one part of the core fails QA and so they shut it out and make it a cheaper part.

The GPU parts might be the tightest silicon and highest rate of failure so this approach reduces waste.


By "trick" you mean the only approach every chip-maker has been following for decades? Literally every single one. It's called binning.


The gp is using “trick” not with the nefarious connotation, but more along the lines of “hack” or “clever idea”.


yeah I guess I should have said 'hack'


I think it's proper way than hack


I think there’s one more notable difference: the M1 MBP also has a brighter screen at 500 vs 400 nits. Both have P3 gamut and are otherwise the same resolution.


The Pro screen is 500 nits vs 400 nits on the air.


Battery is not so important post-covid.

Air is clearly the better value option, if you really want to get one of these.


I guess the cooling let’s them tweak the CPU clocks accordingly? Wonder if we can hack the Mac mini with water blocks and squeeze higher clocks. The memory limitation makes it a dud though.


Wouldn't be surprised if the cooling solution was serialised and it had a detection whether the cooling is originally programmed for the particular unit like they do now with cameras and other peripherals (check iPhone 12 teardown videos). I bet that the logic would check the expected temperature for given binning and then shut down the system if it is too cool or too hot. Apple knows better than the users what hardware should work with the unit.


For a while, the fan was broken in my 2017 MacBook Pro 13". Didn't spin at all. The MacBook never complained (except when running the Apple hardware diagnostics). It didn't overheat or shut down unexpectedly. It just got a bit slower due to more thermal throttling.

I expect it would work the other way, too. Improve the cooling and performance under load would improve.


https://www.youtube.com/watch?v=MlOPPuNv4Ec

This is a video from Linus Tech Tips that demonstrates that no matter how much you cool it, they've physically prevented the chip from taking advantage of it.

And if it could be fixed with software, they would have worked out how, they're into that kinda tweaking.


Intel chips, on the other hand, are designed to work with a varying degree of thermal situations because they don't control the laptop it is put in. In this situation, Apple could potentially get more creative with their approach to thermals because they control the entire hardware stack.


Sure, if by "more creative" you mean handicap the CPUs by not delivering them any power because they know they can't cool them.

https://www.youtube.com/watch?v=MlOPPuNv4Ec


Intel processors use Intel designed throttling solutions... which exist to keep their own processors from overheating because they have no control over the final implementation.

These new M1 laptops are the first laptops that have complete thermal solutions designed by a single company.

As an example, there is the potential to design a computer with no throttling at all if you are able to control the entire thermal design.


> As an example, there is the potential to design a computer with no throttling at all if you are able to control the entire thermal design.

This is not true. A laptop needs to work in a cold room, in a hot room, when its radiator is dusty, etc. If your CPU is not willing to throttle itself then a company with Apple's scale will have machines overheating and dying left and right.

For a computer to never _need_ to throttle, either (1)the cooling system has to be good enough to keep up with the max TDP of the CPU, or (2) you "pre-throttle" your CPU by never delivering it more power than the cooling system could handle. Apple refuses to accept solution 1, so they went with solution 2. If you watch the video I posted, it shows that even when there is adequate cooling, the new macbooks will not deliver more power to the CPU. In effect, the CPU is always throttled below its limit.


If Apple actually did that Louis Rossman would be out of a job.

No, not in the sense that the cooling lockout would make him unable to fix MacBooks - he clearly has the industry connections to get whatever tools he needs to break that lockout. Yes, in the sense that many Apple laptops have inadequate cooling. Apple has been dangerously redlining Intel chips for a while now - they even install firmware profiles designed to peg the laptop at 90C+ under load. The last Intel MBA had a fan pointed nowhere near the heatsink, probably because they crammed it into the hypothetical fanless Mac they wanted to make.

Apple actually trying to lock the heatsink to the board would indicate that Apple is actually taking cooling seriously for once and probably is engineering less-fragile hardware, at least in one aspect.


So, essentially their new Macbook line is a glorified iPhone/iPad but with a foldable display (on a hinge)?

Not too far-fetched when you see the direction MacOS is headed, UI-wise. And it sounds nice, but if it means that repairability suffers then we'll just end up with a whole wave of disposable laptops.


To be fair to apple, people keep their macbooks for years and years, keeping them out of landfill longer. They are well made and the design doesn't really age. Written on my 2015 Macbook pro.


To be fair to the rest of the world, this comment is written on a 20 year old PC. It has had some component upgrades, but works like a champ after 20 years.


If you keep replacing failed/failing components or give needed upgrades to the system every few years, is it fair to call it 'working like a champ for 20 years'?


I'll take it a step further. Is it fair to even call it the same system after 20 years of changes?

Like the Ship of Theseus thought experiment, at what point does a thing no longer have sufficient continuity to its past to be called the same thing? [1]

[1] https://en.m.wikipedia.org/wiki/Ship_of_Theseus


Some parts are kind of like tyres on a bike, just need to be replaced from time to time, it doesn't mean the bike is bad or not working like a champ.


Yeah, but it does mean it is no longer the same bike. If you replace every part of a bike, even one at a time over years, it is no longer the same bike. So it all depends on what GP means by "replacing some parts". Is it entirely new computer in a 20 year old case? Or is it a 20 year old computer with a couple sticks of RAM thrown in?

Regardless, I have a hard time believing a 20 year old computer is "working like a champ". I've found the most people who say their <insert really old phone or computer> works perfectly have just gotten used to the slowness. Once they upgrade and try to go back for a day, they realize how wrong they were. Like how a 4k monitor looks "pretty good" to someone that uses a 1080p monitor everyday, but a 1080p monitor looks like "absolute unusable garbage" to someone who uses a 4k monitor everyday.


Definitely not if the metric we care about is keeping components out of landfills.


I don't understand the landfill argument here.

A typical "Upgradable" PC is in a box 10 times the size of the mini. If you upgrade the GPU on a PC, you toss out an older GPU because it has pretty much zero resale value. Typical Apple hardware is used for 10-15 years, often passing between multiple owners.


It's a shame we don't have charities that would take such parts and then distributed them to less fortunate countries. Ten years ago a ten year old graphics card would no longer be quite usable, but now 10 years old card should work just fine for most of the tasks, except more advanced gaming.


I don't see the point. There is nothing to put it into. It's far cheaper to just ship modern CPUs with integrated graphics which will be faster and more efficient than that 10 year old GPU. The era where computer components were big enough for it to make sense for them to be discrete parts is coming to a close.

This is particularly true on the lower end where a 10 year old part is even interesting.


I thought you could donate any part of a computer and then people could sort and match, but I think you're right.


if only two parts got replaced, then landfill mass was reduced.


Why do I think of Trigger's Broom when I read this?


Apples and oranges. I've never kept a laptop for five years.


That's only applicable to Macbooks made upto 2015.


I guess I'll throw my 2016 MBP out then.


You probably will before I throw out my 2010 MBP thanks to easily replaced parts.


To me it looks more like they swapped the motherboard out with their own, keeping the rest of the hardware the same.

With RAM and SSD already soldered to the motherboard, repairability can't really get much worse than it already is.


It's not difficult to replace RAM or SSD with the right tools (which may be within reach of an enthusiast), problem is that you often cannot buy spare chips as manufacturers can only sell them to Apple or that they are serialised - programmed to work only with that particular chip and then the unit has to be reprogrammed after the replacement by the manufacturer. I think they started doing it after rework tools became affordable for broader audience. You can get a trinocular microscope, rework station and an oven for under a $1000 these days.


You can get a screwdriver (allowing you to replace RAM and SSDs in most laptops, including older macs) for $5. There's really no excuse for them to do this all the while claiming to be environmentally friendly.


Depends on the model. My 2012 mbp15r uses glue and solder, not screws. Maxed out the specs when I got it, which is why it's still usable. Would've been delighted for it to have been thicker and heavier to support DIY upgrades and further improve its longevity while reducing its environmental impact, but that wasn't an option. Needed the retina screen for my work, bit the bullet. Someday maybe there will be a bulletproof user-serviceable laptop form factor w a great screen, battery life and decent keyboard, that can legally run macOs... glad to say my client-issued 2019 mbp16r checks most of those boxes. /ramble


Something like ATX standard but for laptop shells would be awesome - imagine being able to replace a motherboard etc, just like you can with a desktop PC.


Intel tried this more than a decade ago. The designs were as horrible as you might imagine, and a few OEMs did come out with a handful of models and parts.

As I recall, consumers didn’t care or wouldn’t live with the awful designs that they initially brought out. I don’t remember. I remember thinking I wouldn’t touch one after seeing a bunch of engineering samples.


Maybe it was too early for this kind of thing. I could imagine today such shell would be much slicker.


Except the RAM is in the M1 now. Pretty good excuse Id'say.


Is it? I thought only memory controller is in the chip, not the memory itself.


The M1's RAM is integrated into the SoC package. But it's still separate RAM chips, not actually on the same die as the M1.


Mmm... it's certainly better than they had before. But really they ought to be designing repairable machines. If that makes them a little slower then so be it.


My 2007 MBP, yes. I don't think that's true of my 2017 MBP, nor my 2012 MBA.

It's been years since Apple did away with this stuff, and nobody expected them to suddenly allow after-market upgrades.


Serialized components should be illegal, frankly.


There are good privacy and security reasons that someone might want serialized components.


Sure, but you add the option to ignore the serialization, or options to reset the IDs as part of the firmware or OS. That way the machine owner can fix it after jumping through some security hoops, rather than requiring an authorized repair store.

Mostly because, its doubtful if state level actors (or even organized crime) aren't going to pay off an employee somewhere to lose the reprogramming device/etc. Meaning its only really secure against your average user.


I don't believe those reasons are more important than open access and reducing the environmental impact of planned obsolescence, outside of the kind of government agencies that are exempt from consumer electronics regulations anyway.


Surely there is a better (and I'd bet, more effective) way to handle environmental regulations than mandating specific engineering design patterns within the legal code.

Perhaps instead, it might be a better idea to directly regulate the actions which cause the environmental impact? i.e. the disposal of those items themselves?

Engineers tend to get frustrated with laws that micromanage specific design choices, because engineering practices change over time. Many of the laws that attempt to do so, backfire with unintended consequences.

It is quite possible that your solution might be just that -- many industries with high security needs are already very concerned with hardware tampering. A common current solution for this is "burner" hardware. It is not uncommon for the Fortune 500 to give employees laptops that are used for a single trip to China, and then thrown away. Tech that can give the user assurance that the device hasn't been compromised decreases the chance that these devices will be disposed of.

As a side note, I don't think serialized components is even one of the top 25 factors that does(/would) contribute to unnecessary electronics disposal.


I think resetting instead of bricking doesn't compromise security, but saves a burner laptop from ending up in landfill. I get your point, but I think company would have to demonstrate that e.g. serialising meets particular business need that is different from planned obsolescence. Could be a part of certification processes that products before getting marketed have to go through.


Based on what legal principle should they be illegal?


In practice, such a law could resemble right-to-repair bills like the one recently passed in Massachusetts, which requires auto manufacturers to give independent repair stores access to all the tools they themselves use. A bill like this for consumer electronics could practically ban serialized components, even without mentioning them explicitly.


Illegal, no. Taxed extra.


Why beating around the bush? If the function of extra tax is to stop producers from implementing planned obsolescence, then why not just stop them directly and require that components are not serialised etc. as a part of certifications products need to go through? If you add tax, then all you do is restricting access to such products for people with lower income.


the point is to push the market into the correct^Wdesired direction without outright banning anything. non-serialized would be cheaper, hence more accessible. there are use cases where serialized parts are desired (e.g. if i don't want somebody swapping my cpu with a compromised part).


Normally I prefer nudges to bans, but I'm not sure they work on giant monopolies. Unless the tax were high enough to have no chance of passing, Apple would dodge it or write it off as cheaper than being consumer-friendly.


With Apple Silicon, the RAM is not even on the motherboard. It's integrated into the SoC package!


I don’t think that’s true for M1.


Yes, it is true of the M1. RAM is integrated into the M1 SoC package (but not on the same die).


> So, essentially their new Macbook line is a glorified iPhone/iPad but with a foldable display (on a hinge)?

This isn't some new. Since day 1, the iPhone has always been a tiny computer with a forked version of OS X.

> but if it means that repairability suffers then we'll just end up with a whole wave of disposable laptops.

Laptops have been largely "Disposable" for some time. In the case of the Mac, that generally means the laptop lasts for 10-15 years unless there is some catastrophic issue. Generally after that long, when a failure happens even a moderate repair bill is likely to trigger a new purchase.


You won't be able to go beyond the built in P-states which in the end is a power limit not a thermal one.


8GB and 16GB configurations seem more than enough..


I had a quad-core Mini with 16GB in 2011. Almost 10 years later we should be much further, especially as the Intel Mini allows up to 64GB. (Which you probably would use only if you upgraded the memory yourself).


We're not any further in terms of capacity per dollar, but we are advancing in terms of speed.

The M1's memory is LPDDR4X-4266 or LPDDR5-5500 (depending on the model, I guess?) which is about double the frequency of the memory in the Intel Macs.

Apparently, this alone seems to account for a lot of the M1's perf wins — see e.g. the explanation under "Geekbench, Single-Core" here: https://www.cpu-monkey.com/en/cpu-apple_m1-1804

Bleeding-edge-clocked DRAM is a lot more costly per GB to produce than middle-of-the-pack-fast DRAM. (Which is weird, given that process shrinks should make things cheaper; but there's a DRAM cartel, so maybe they've been lazy about process shrinks.)


Not all types of processes shrink equally well.

Apparently DRAM and NAND do not shrink as well because in addition to transistors in both cases you need to store some kind of charge in a way that is measurable later on - and the less material present, the less charge you are able to store, and the harder it is to measure.


> The M1's memory is LPDDR4X-4266 or LPDDR5-5500 (depending on the model, I guess?) which is about double the frequency of the memory in the Intel Macs.

That's a high frequency, but having two LPDDR chips means at most you have 64 bits being transmitted at a time, right? Intel macs (at least the one I checked), along with most x86 laptops and desktops, transfer 128 bits at a time.

> Apparently, this alone seems to account for a lot of the M1's perf wins — see e.g. the explanation under "Geekbench, Single-Core" here

That's a vague and general statement that site always says, so I wouldn't put much stock into it.


No virtualisation -> I’m guessing no Docker.

Am I missing something?

Mind you with 16Gb, Docker won’t be that useful.


Why do you think there is no virtualisation? Apple showed Linux running in a VM during WWDC already.


I missed that, I assumed virtualisation was dependent on Intel VT.

Then again I would have expected them to have discussed it as much as the video editing.

I am guessing that they’d need a M2 type chipset for accessing more RAM for that. Or maybe they’ve got a new way to do virtualisation since that is such a key thing these days.

Edit: thanks for pointing that out though, that’s why I mentioned it

https://developer.apple.com/documentation/virtualization

https://news.ycombinator.com/item?id=23922846

And the mentioned Virtio here:

http://www.linux-kvm.org/page/Virtio

How well this fits in with current virtualisation would be interesting to find out; I guess this will be for a later version of Big Sur, with a new beefier M2 chip.


Are they virtualizing x86 though? Having Docker running arm64 on laptops and Docker x86 on servers completely nullifies the main usecase of Docker imo.


But you can run Docker on arm64 servers!


The intel Mac Mini is still available with the same 8GB in its base model, but configurable up to 16/32/64. RAM is definitely the biggest weakness of these new Macs.

On iOS they can get away with less RAM than the rest of the market by killing apps, relaunching them fast, and having severely restricted background processes. On Mac they won't have that luxury. At least they have fast SSDs to help with big pagefiles.

With the heterogeneous memory, your 8GB computer doesn't even have its whole 8GB of main system memory.

When the touchbar MBP launched in 2016 people were already complaining that it couldn't spec up to 32GB like the competition. Four years later, and it's still capped at 16GB.

Hopefully they can grow this for next year's models.


And the Intel Mac Mini had user-replaceable RAM. Tired of fan noise and slow response, I went from a 4 Thunderbolt 2018 MacBook Pro with only 8GB of RAM to a 2018 Mac Mini with 32GB of RAM (originally 8GB, bought the RAM from Amazon and upgraded it).

The difference was incredible


8GB ram is just soul crushing - even for basic office workloads. I need 16GB minimum.


What in a basic office needs 8GB RAM?! I used Word 6.0 under Windows 95 with 64MB of RAM!

Have you looked at Activity Monitor to see what is butchering your memory?!


Well, for starters, the most obvious answer, which is Office 365. Have you glanced at the RAM use of those apps, ever?

Second: web browsers, which can easily grab 5-10GB by themselves or even more if RAM is available.

So in other words: everything.


Probably Chrome.


I'm idling at 18gb right now and doing what I consider to be next to nothing.


It doesn't make sense for the system not to 'grab' a big chunk of your RAM. That is what it is there for. You want stuff to be preloaded into RAM so you can access it quickly if needed. You only want to leave some of it free so that if you launch a new application it has breathing room.

For example Chrome will scale the amount of RAM it reserves based on how much you have available.


> It doesn't make sense for the system not to 'grab' a big chunk of your RAM. That is what it is there for. You want stuff to be preloaded into RAM so you can access it quickly if needed. You only want to leave some of it free so that if you launch a new application it has breathing room.

Cache is excluded from just about any tool that shows RAM use, at least on desktops. If the ram shows as in use, the default assumption should be that it's in active use and/or wasted, not cache/preloading.

> For example Chrome will scale the amount of RAM it reserves based on how much you have available.

Which features are you thinking about that reserve ram, specifically? The only thing I can think of offhand that looks at your system memory is tab killing, and that feature is very bad at letting go of memory until it's already causing problems.


That seems like a hell of a lot of RAM for next to nothing.

I'm not a mac user but that seems ridiculous. I'd be investigating what's hogging it all.


I build my desktops with a lot of ram.

I have chrome, Firefox, photoshop, vs code, docker and a few other things running. As a kid I had to manage RAM. As an adult, I buy enough RAM to not need to think about it.

I was committed to buying an M1 on day one. I won’t buy a machine with only 16gb of RAM.


I'm the same, my current desktop has 32Gb, but still I'd be pretty concerned about 18Gb in use with nothing running.


The 2018 Intel Mac Mini has user-replaceable RAM. The 2014 mini has fixed RAM.


Another note on the Mini and MacBook Pro (in higher end SKUs) - these both used to have four USB-C ports, and now only have two. The Mini at least keeps its a pair of USB-A ports, but on the MBP you're back on the road to dongle-hub-land.

I'm betting this is due to Thunderbolt controller and PCIe lane capacity. They couldn't do four Thunderbolt ports with the M1 SoC, so they dropped the ports. Having four USB-C ports but only two supporting Thunderbolt would be a more obvious step back from the previous MacBook Pro. This way people can just blame it on Apple doing Apple things, instead of seeing a technical limitation.


Yes, based on my experience on a mac, I would not buy any mac with less than 32gb ram (I personally use 64gb and it's so much nicer)...

Yes, it seems crazy, yes it's a lot of ram, but I like to be able to run VMs locally and not have to boot up instances on AWS (insert provider of choice), I like to keep tabs open in my browsers, I like not to have to close apps when I'm using them and I like my computer to be snappy. 64 GB allows that 16 doesn't, 32 barely does.


Having read a bit more about the new M1 I really think it is designed and speced for the new Air. The RAM is on the package, which makes it very performant and 16G is a reasonable limit for an Air-like computer. The Mini got cheaper and more powerful, so it is not a bad trade off. I strongly assume, that there will be variations/successors to the M1 which are going to support more memory and also more IO (more USB-4 ports, more screens).


From their schematic, the two DRAM modules were directly on the SoC - possibly to improve bandwidth etc. So it looks like this cannot be upgraded / replaced. That said, it might be worth it to look past the specs and just use your applications on these machines to see how they perform. SSD storage is much faster these days and if the new OS has decently optimized paging, performance will be decent as well.


Also, lack of 10gbe is a big let down...


16GB limit with the latest MBP M1 13inch seems a big downer, I will wait for 16 inch MBP refresh now.


You have to factor in possible memory management improvements with the M1 chip, and ability to run iOS apps instead: https://twitter.com/jckarter/status/1326240072522293248


That is fine with the Air. But for a small desktop computer not to support more than 16GB in 2021? Its predecessor allowed up to 64GB (and possibly more with suitable modules).


They've also halved the 2 base SSDs to 256/512

I thought with the last update they'd finally seen the light and moved to 512/1tb, now we're back with the silly 256gb.

If you factor in having to upgrade ram to 16gb and ssd to 512 it's only £100 shy of the old price. Good, but not as good as it looked to begin with.


You can get an external M2 USB 3.1 Gen 2 (10Gbps) enclosure plus 1TB M2 SSD for $130 and a 2TB for $320. That makes the 16GB Mac Mini 256GB a decent buy at $970 imo.

https://www.amazon.com/dp/B07MNFH1PX


For the mini sure, but it's a massive pain having an external drive for a laptop. I use one all the time and as well as being waaaaay slower even with a good drive, I lose it all the time.


Yeah it's not feasible at all to use external storage on a two port laptop. Dongles that allow you to plug in power and monitor are still just not reliable enough for storage, the only reliable connection I can get on my 2 port MBP is with a dedicated Apple USB-C to A adapter.

Shocked they're still selling the two port machine, it's been nothing but hassle for me as someone who has to use one.


That's why I'm waiting to upgrade my laptop.


I've still got my 2013 MacBook Pro. Still chugging along.

I'm hoping it can wait for v2 of the MacBook Pro 16"

Apple's second version of everything is always worth the wait.


For pro users, the fact that 32GB isn’t even an option is pretty surprising


My guess is that the next wave will be Pro.

And they will have significantly upgraded CPU/GPUs to match the memory.


But it's right there in the name: 13" MacBook Pro


The 13" 'pro' has never really been a 'real' pro. They were/are always spec'd with less cores than the 15"/16" and never had dedicated graphics.


There are two lines of 13" MacBook Pro, the two-port and four-port versions. The two-port always lagged behind the four-port, with older CPUs, less RAM, etc. The four-port (which has not yet been replaced) is configurable to 32GB of RAM.


Entry level 13" MacBook Pro is for prosumers.

Think web developers, photographers, bloggers etc.


Web developers and photographers are the opposite of 'prosumers', kind of by definition. Plus, think of the size of a full res photo coming out of a high-end phone, never mind a DSLR.


Most of the professional photographers that I work with have PC workstations with 64gb to 256gb of RAM. Retouching a 48MP HDR file in Photoshop needs roughly 800MB of RAM per layer and per undo step.


Old undo steps could be dumped to SSD pretty easily.

And while I understand that many people are stuck on photoshop, I bet it would be easy to beat 800MB by a whole lot. But so I can grasp the situation better, how many non-adjustment layers do those professional photographer use? And of those layers, how many have pixel data that covers more than 10% of the image?


From what I've seen, quite a lot of layers are effectively copies of the original image with global processing applied, e.g. different color temperature, blur, bloom, flare, hdr tone mapping, high-pass filter, local contrast equalization. And then those layers are being blended together using opacity masks.

For a model photo shoot retouch, you'd usually have copy layers with fine skin details (to be overlaid on top) and below that you have layers with more rough skin texture which you blur.

Also, quite a lot of them have rim lighting pointed on by using a copy of the image with remapped colors.

Then there's fake bokeh, local glow for warmth, liquify, etc.

So I would assume that the final file has 10 layers, all of which are roughly 8000x6000px, stored in RGB as float (cause you need negative values) and blended together with alpha masks. And I'd estimate that the average layer affects 80%+ of all pixels. So you effectively need to keep all of that in memory, because once you modify one of the lower layers (e.g. blur a wrinkle out of the skin) you'll need all the higher layers for compositing the final visible pixel value.


Huh, so a lot of data that could be stored in a compact way but probably won't be for various reasons.

Still, an 8k by 6k layer with 16 bit floats (which are plenty), stored in full, is less than 400MB. You can fit at least eleven into 4GB of memory.

I'll easily believe that those huge amounts of RAM make things go more smoothly, but it's probably more of a "photoshop doesn't try very hard to optimize memory use" problem than something inherent to photo editing.


So why are you blaming the end user for needing more hardware specs than you'd prefer because some 3rd party software vendor they are beholden to makes inefficient software?

Also, your "could be stored in a compact way" is meaningless. Unless your name is Richard and you've designed middle out compression, we are where we are as end users. I'd be happy if someone with your genius insights into editing of photo/video data would go to work for Adobe and revolutionize the way computers handle all of that data. Clearly, they have been at this too long and cannot learn a new trick. Better yet, form your own startup and compete directly with the behemoth that Adobe is and unburden all of us that are suffering life with monthly rental software with underspec'd hardware. Please, we're begging.


Where did I blame the end user?

> Also, your "could be stored in a compact way" is meaningless. [...]

That's getting way too personal. What the heck?

I'm not suggesting anything complex, either. If someone copies a layer 5 times and applies a low-cpu-cost filter to each copy, you don't have to store the result, just the original data and the filter parameters. You might be able to get something like this already, but it doesn't happen automatically. There are valid tradeoffs in simplicity vs. speed vs. memory.

"Could be done differently" is not me insulting everyone that doesn't do it that way!


You must not rate photographers very highly if you are mentioning them in the same sentence as "bloggers, etc".


I should wait for a 64 GB option. I've already got 16 GB on all my older laptops, so when buying a new gadget RAM and SSD should have better specs (you feel more RAM more than more cores in many usage scenarios).

It was surprising to see essentially the same form factor, the same operating system and not much to distinguish the three machines presented (lots of repetition like "faster compiles with XCode").

BTW, what's the size and weight of the new Air compared to the MacBook (which I liked, but which was killed before I could get one)?

Seeing two machines that are nearly identical reminds me of countries with two mainstream political parties - neither discriminates clearly what their USP is...


I don't think today's computers were aimed at those kinds of usecases.


Apple has a "missing middle" problem.

They have a ton of fantastic consumer-level computing devices, and one ridiculously-priced mega-computer.

But there are many of us that want something in the upper-middle: a fast computer that is upgradable, but maybe $2k and not $6k (and up).

(The iMac Pro is a dud. People that want a powerful desktop generally don't want a non-upgradable all-in-one.)


Apple's solution for upgradability for their corporate customers, is their leasing program. Rather than swapping parts in the Mac, you swap the Mac itself for a more-powerful model when needed — without having to buy/sell anything.


Apple has missing middle _strategy_


Apple doesn't care about your upgradability concerns on the notebook lineup. Once you get past that, it has traditionally done fairly well at covering a wide spectrum of users from the fanless MacBook to the high-powered MacBook Pros.


I have a late-2013 13" MBP with 16GB of memory. Seven years later I would expect a 13" MBP to support at least 32GB. I can get 13" Windows laptops that support 32GB of memory. The Mini is a regression, from 64GB to 16GB of memory. The only computer worth a damn is the new MBA.


Pretty sure my 2014 ish 13inch MBP with 16gb and 512 storage cost me around £1200, today speccing an M1 13inch MBP to the same 6 year old specs would cost almost £2000.

Seems absurd.


Wait just a bit and I'm sure your concerns in this area will entirely disappear.


They already disappeared, I switched to Windows in 2019.

I use MacStadium for compiling and testing iOS apps. I was wondering if the ARM machines would be worth a look, but they are disappointing. If I was still using Macs as my daily driver, I would buy the new MBA for a personal machine.


But 16gb is what I had in a computer 10 years ago.

I was just window shopping a new gaming rig, and 32gb is affordable (100 bucks), 64gb (200 bucks). Cheap as shit, what’s the hold up?


The memory is on package, not way out somewhere on the logic board. This will increase speed quite a bit, but limit physical size of memory modules, and thus amount. I think they worked themselves into a corner here until the 16” which has a discreet GPU and reconfiguration of the package.


A little bit up is was shown the memory in M1 is 5.5GHz DDR5. https://news.ycombinator.com/item?id=25050625

Can you please provide the link to 64GB DDR5-5500 for $200? I'd love to buy some too!


It's fair but if they choose fast but expensive and unexpandable technology, possibly the choice is failed in some perspective. I think most people who buy mini prefer RAM capacity than faster iGPU.


https://www.newegg.com/p/pl?d=64gb+ram

I guess DDR 500 runs you 350. It looks like Apple charges you 600 for ddr 400 32gb. I don’t know, what am I missing here?


Can you actually link to a product, not a search ? Because none of the items coming up there are DDR5-5500, they're all DDR4-3600 or worse, as far as I can see.


I guess I was wrong, everything is ddr4.


I’m confused. The link is for DDR4, it’s all too slow, and Apple doesn’t offer a 32MB M1 option at this time.


A new processor architecture. Wait a couple months and you'll probably have the computer you wanted released too.


The DRAM seems to be integrated on the same package as the SoC.


I went to Apple's website right after I finished watching the keynote with the intention of buying a new Mac mini ... the lack of memory options above 16GB stopped that idea dead in its tracks though.


Also no 10G networking option. The combination of those feature exclusions makes it a dud for me; I don't want to have a $150 TB3 adapter hanging off the back, not when previous gen had it built in.


I bet “pros” never bought it and it’s only been viable as a basic desktop. Probably nobody ordered the 10 gigabit upgrade.

I bet they’re only upgrading it because it was essentially free. They already developed it for the developer transition kit.

I commend the idea of a small enthusiast mini desktop like a NUC but I don’t think the market is really there, or if they are, they’re not interested in a Mac.


I think it is notable the new mini’s colour is light silver, rather than the previous dark ‘pro’ silver. Presumably there will be another model a year from now.


Over the years the mini has had a variety of different shades and designs. I wouldn't read too much into it.


I bet “pros” never bought it.


16gb can be limiting for some work flows today, and doesn't give you much future proofing (this RAM is permanent, right?)


Yes, it's in the SoC (or SiP now).


How do they get the ram into the SoC? Is it like a massive die?



Don’t touch Xcode then, it welcomes you to paging hell.


With that fast SSD, do you notice paging in Xcode? Would it be worth the extra $300 or however much Apple asks for extra 8GB of ram in the US store?


yes, you notice paging, even with 'fast' SSD.

or perhaps it's not 'paging', and just dumb luck I hit and see beachballs on multiple new higher-end macbook pros regularly.


It's not normally paging, but thermal throttling which involves the machine appearing to 'spin' but it's actually just the kernel keeping the cycles to itself, which typically give you beachballs as a side-effect.

And one tip is to use the right hand side USBC ports for charging, not the left hand ones as for some reason or other they tend to cause the machine to heat up more...


the right hand ones are the only ones that can drive external monitors (on mine anyway). I feel like I'm the only one that has this - I had a MBP 2019 - first batch - and I thought I'd read that one side was different than the other re: power. Power works on both sides, but monitors won't run from the left usb-c ports. but it's not documented anywhere. :/

thx for tip.


I just tried plugging my monitor into a right hand socket on my 2019 MBP, and it worked fine for me.

On my machine it is true that charging on the right is better. Charging on the left spins up the fan.


Just a thought, but maybe everyone should be appalled at that extra $300. And the lack of upgradability on a Pro machine, especially.


You're talking to Apple customers. Being gouged is a way of life for them.


The wording in the event supports this. Particularly when speaking about the Mini's fan "unlocking" the potential of the M1 chip.


This makes sense.

Most likely this is why the CPUs are all limited to 16GB. It's likely when they unwrap the 16 inch MacBook Pro, it will open up more configurations (more RAM in particular!) for the 13" MacBook Pro and hopefully the mini.


Going into the event, my thinking was that they'd have two aims:

1. "Wow" the audience given the anticipation without a full overhaul of the range. 2. Deliver some solid products that enable the transition while being effective for non-enthusiasts.

From my viewing they hit both. I expect they'll fill in the range next fall with bigger upgrades to the form factor.


I agree, It almost feels like they are going to have 3 main M Series CPUs. This one. One for the iMac and higher end MBPs. And perhaps a third for the high end iMac/ Mac Pro.


RAM limits are pretty easy to explain. 16GB chips cost disproportionately more and use more power.

I wonder if they use 2 4GB chips or 1 8GB chip in the low-end SKU?


It's even easier to explain than that. The RAM is integrated into the CPU. While there are a few SKUs here, Apple only designed and built one CPU with 16GB RAM. The CPUs are binned. The CPUs where all RAM passed testing are sold as 16GB, the 8GB SKUs had a failure on one bank of RAM.

There are no 32 or 64 GB models because Apple isn't making a CPU with 32 or 64GB of RAM yet.


It looked to me that they were placing 2 DDR4 modules beside the chip.

https://images.anandtech.com/doci/16226/2020-11-10%2019_08_4...


Well maybe? Huh. I don't know now, certainly looks like it.


The logic board probably isn't the same, but the SoC [probably] is identical, and with it a lot of the surrounding features/implementation. My own speculation as well of course :)


I doubt the logic board is the same. It’s just that the M1 integrates so much.


Actually that was what I notice in the video as well. Mac mini has huge amount of empty space. And the only difference was the cooling unit fitted on top.


There used to be a DVD player and a hard drive in there, there has to be spare room.

When comparing the mini to other SFF computers be sure to note that the mini has the power supply built in though where most of the others have it external.


If you look at the power supplies it's 30 vs 60 Watts, definitely interested to see what kind of TDP Apple targets with these machines.


They've stated that they target 10W in the Air, the cooling system in the Pro is identical to the Intel so probably 28W ish, and the Mac Mini specs say maximum 150W continuous power, but that probably includes a few tens of watts of USB and peripherals.


2 port Pros usually have 15 watt parts and only a single fan. Of course, PL2 goes far above 15W.


M1 chip page shows that 10 watts is the thermal envelope for the Air


So I get the whole performance-per-watt spiel, but if they're targeting the same 10W as with Ice Lake Y-series[1], it's gonna be hot during continuous workloads like screen sharing, since they've seemingly decided to get rid of even the "chassis fan" they've had on 2019 Air.

[1] https://ark.intel.com/content/www/us/en/ark/products/196596/...


Worth noting that Intel doesn't actually honor those 10W listings, and often boosts above it for significant portions of usage.


This would make sense given the pricing, too.

For example, the $1,249 air is very similar to the $1,299 pro. The MBA has a bigger SSD, but the MBP has a bigger battery, isn't throttled (i.e. has a fan), and also has the touchbar (which may be controversial, but for the sake of comparison, remember that it comes with a manufacturing cost).

It seems reasonable that these are priced similarly. Of course, the machine I want is the MBP with the bigger SSD and no touchbar :)


The max ram in all 3 is only 16gb :(


I noticed the slides kept saying "Up to" 8 GPU cores.

That left me wondering if there are different variants of the M1 with different core counts.

(Note: It always said 8 CPU cores)


Looks like there's two variations of the Air: one with 7 GPU cores and one with 8 GPU cores.

https://www.apple.com/shop/buy-mac/macbook-air


According to Apple's website the Macbook Air appears to only have 7 active GPU cores. I suspect that chips in the Air may be binned separately and may or may not support the higher clock speeds even with the active cooling of the Mac Mini and Macbook Pro.


7 on the base model, 8 on the upgrade. You're probably correct that this is a binning thing.


According to the tech specs on the Apple site, the Air is available with 7 or 8 GPU cores. All other new Macs have 8.


They did the exact same thing with the A12X and the A12Z. 7-core vs 8-core GPU is the only real difference between them.


My guess is maybe this is a yields thing for Apple Silicon? They use the same chips for Air and Pro, but shut off a faulty gpu core that didn't pass QA? Or a temperature thing.


One of the slides mentioned that the air is limited to 10watts though. I wonder if it does have the same soc but its nerfed beyond 10watts.


The two differences are cooling and that the base Air appears to receive binned processors with 7 GPU cores.


The MBP now only has 2 USB-C/Thunderbolt ports which would support this theory.


That's the same as previous low-end MBP.


that's right but the 'regular' one had 4. I've already seen a pro user (in music production) complain about this.

But my point here is that the fact they are both the same supports the theory that the logic board is the same on both models.


They’re continuing to sell the 4 port Intel 13” Pro.


Maybe they'd launch a more expensive 4 USB-C/Thunderbolt ports model with their more powerful chip (and upto 64/128GB memory) like they did with the earlier MBP13s.


"Testing conducted by Apple in October 2020 using preproduction MacBook Air systems with Apple M1 chip and 8-core GPU, as well as production 1.2GHz quad-core Intel Core i7-based MacBook Air systems, all configured with 16GB RAM and 2TB SSD. Tested with prerelease Final Cut Pro 10.5 using a 55-second clip with 4K Apple ProRes RAW media, at 4096x2160 resolution and 59.94 frames per second, transcoded to Apple ProRes 422. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Air."

This is relevant. This means that the performance increase vs Intel is using the extremely throttled 1.2 GHz i7 as the baseline.


According to the Anandtech benchmark the A14 (what's in the latest iPhone) is 8% below the Ryzen 5950X. M1 is likely faster than that.

https://www.anandtech.com/show/16226/apple-silicon-m1-a14-de...

Quote:

"In the overall SPEC2006 chart, the A14 is performing absolutely fantastic, taking the lead in absolute performance only falling short of AMD’s recent Ryzen 5000 series.

The fact that Apple is able to achieve this in a total device power consumption of 5W including the SoC, DRAM, and regulators, versus +21W (1185G7) and 49W (5950X) package power figures, without DRAM or regulation, is absolutely mind-blowing."


This is stupefying to me. Like, when Anand did the A12 or A11 reviews, I wrote:

"The craziest thing is how close [to 1/56th of a Xeon 8176] they get at 3W. That's the part that's staggering. What could Apple do with, say, laptop thermals? Desktop thermals? The mind creeps closer to boggling." -- me, 5 October 2018

Guess we're finding out.


Would this be good a chip for a data warehouse?


I think so, especially with the specialized hardware in there for machine learning (although in a data center you'd probably opt for specialized hardware like discrete GPUs). There are some players on the market for ARM servers as well.

That said, Apple won't make their chips available to the greater market. They might use variations of it in their own data centers, maybe.


> That said, Apple won't make their chips available to the greater market.

I wonder if this could potentially be a good time to bring back Xserve...


They definitely should. Even if the profit on the device wouldn't be decisive, but chip making is all about production numbers. Using the Apple Silicon first for all the Apple owned cloud servers and then offering them in Xserve units would be very profitable and could reduce the power draw of data warehouses a lot.


The question marks for me would be:

1) Is Apple interested (and capable) making macOS into a realistic choice for servers again?

2) If not, is Apple comfortable with selling a box that most people will put Linux on? Would they put resources into a Bootcamp-esque driver package?


The biggest question is: does Apple want to bring back the Xserver? If yes, there are several ways. If not, no good proposal will sway them :) But one thing was interesting during the AS announcement at WWDC. Federighi stressed how good a hypervisor macOS is. That was probably first of all a comment about virtualization on the new Macs, but could also allow to install any operation system as guest on a cloud server. Many cloud services are running inside VMs.


The only reason for that statement is to silence those who would raise a stink over the loss of Bootcamp.

"See, you don't need native Linux on the machine, just use our hypervisor & Docker/VMs"


Yes, that was the context of the statement. But it also would give an answer how customers would use an AS Xserve.


The market for Mac OS specific services would be vanishingly small; everybody else can just use Linux. Apple's not going to make a play for a commodity market.


this is what that boggles my mind. Like are those benchmarks really comparable? I mean Intel and AMD have been doing this far longer than Apple and it seemed as if Apple just came in and bend the laws of physics that Intel and AMD have been limited by


Apple has been at it for 12 years (they bought PA Semi in 2008). The talent they brought onboard with PA Semi had been producing processors since 2003.

Experience counts, but Apple is pretty old-hat at this by now as well.


Apple has been designing chips since the A4 in 2010. Samsung, TMSC, and other manufacturers with more experience actually build the chips for Apple, and this M1 is no exception being manufactured by TMSC.


Isn’t that just the ARM vs x86 difference? They’re completely different architectures, each with their own benefits and drawbacks.


With the size of modern chips the instruction set has very little to do with an overall chip design. Most of it is a balancing act between power, performance and area for things like caches, controllers, optimizations, accelerators and so on. Then there are also factors like verification complexity and the fab technology you are choosing.


I would have thought so too, but why are there no ARM based competitors coming out with similar numbers? Or are there?


Graviton2 I expect would stack up nicely.


One factor seems to be the vertical integration. It seems like these days there are a bunch of large companies succeeding by having the users of the hardware involved tightly with the the designers of that hardware. Tesla is definitely doing it. I think Google is doing it with machine learning. And Apple has definitely been doing it with their phones and tablets. The feedback loop is tight, valuable, and beneficial. No need to make big compromises for backward compatibility, either, just update the software.


Personal opinion, no insider info

I think it's a good CPU, but I don't think it'll be a great CPU. Judging by how heavily they leaned on the efficiency, I am pretty sure that it will be just good enough to not be noticeable to most Mac users.

Can't wait to see actual benchmarks, these are interesting times.


You skipped the keynote parts where they say faster 2x, 4x, 8x... and they say faster than every other laptop CPU and faster than 98% of laptops/PCs in their class sold last year? They said it several times. But I agree... let's wait for the benchmarks (btw. iPads already overturns nearly all laptop CPUs)


> You skipped the keynote parts where they say faster 2x, 4x, 8x... and they say faster than every other laptop CPU and faster than 98% of laptops/PCs in their class sold last year? They said it several times.

The important part there is "in their class".

I'm sure the Apple silicon will impress in the future, but there's a reason they've only switched their lowest power laptops to M1 at launch.

The higher end laptops are still being sold with Intel CPUs.


That's not what they said during the keynote.

They said the Macbook Air (specifically) is "3x faster than the best selling PC laptop in its class" and that its "faster than 98% of PC laptops sold in the last year".

There was no "in its class" designation on the 98% figure. If they're taken at their word, its among every PC laptop sold in the past year, period.

Frankly, given what we saw today, and the leaked A14x benchmarks a few days ago (which may be this M1 chip or a different, lower power chip for the upcoming iPad Pro, either way); there is almost no chance that the 16" MBPs still being sold with Intel chips will be able to match the 13". They probably could have released a 16" model today with the M1 and it would still be an upgrade. But, they're probably holding back and waiting for a better graphics solution in an upcoming M1x-like chip.


This probably says more about the volume of low end PCs being sold than about the performance of the air.


If you believe that, then you've either been accidentally ignoring Apple's chip R&D over the past three years, or intentionally spinning it against them out of some more general dislike of the company.

The most powerful Macbook Pro, with a Core i9-9980HK, posts a Geekbench 5 of 1096/6869. The A12z in the iPad Pro posts a 1120/4648. This is a relatively fair comparison because both of these chips were released in ~2018-2019; Apple was winning in single-core performance at least a year ago, at a lower TDP, with no fan.

The A14, released this year, posts a 1584/4181 @ 6 watts. This is, frankly, incomprehensible. The most powerful single core mark ever submitted to Geekbench is the brand-spanking-new Ryzen 9 5950X; a 1627/15427 @ 105 watts & $800. Apple is close to matching the best AMD has, on a DESKTOP, at 5% the power draw, and with passive cooling.

We need to wait for M1 benchmarks, but this is an architecture that the PC market needs to be scared of. There is no evidence that they aren't capable of scaling multicore performance when provided a higher power envelope, especially given how freakin low the A14's TDP is already. What of the power envelope of the 16" Macbook Pro? If they can put 8 cores in the MBA, will they do 16 in the MBP16? God forbid, 24? Zen 3 is the only other architecture that approaches A14 performance, and it demands 10x the power to do it.


Not all geekbench scores are created equal. Comparing ARM and x86 scores is an exercise in futility as there are simply too many factors to work through. It also doesn't include all workload types.

For example, I can say with 100% confidence that M1 has nowhere near 32MB of onboard cache. Once it starts hitting cache limits, it's performance will drop off a cliff as fast cores that can't be fed are just slow cores. It's also worth noting that around 30% of the AMD power budget is just Infinity Fabric (hypertransport 4.0). When things get wide and you have to manage that much wider complexity, the resulting control circuitry has a big effect on power consumption too.

All that said, I do wonder how much of a part the ISA plays here.


M1 has 16MB of L2 cache; 12MB dedicated to the HP cores and 4MB dedicated to the LP cores.

Another important consideration is the on-SOC DRAM. This is really incomparable to anything else on the market, x86 or ARM, so its hard to say how this will impact performance, but it may help alleviate the need for a larger cache.

I think its pretty clear that Apple has something special here when we're quibbling about the cache and power draw per core differences of a 10 watt chip versus a 100 watt one; its missing the bigger picture that Apple did this at 10 watts. They're so far beyond their own class, and the next two above it, that we're frantically trying to explain it as anything except alien technology by drawing comparisons to chips which require power supplies the size of sixteen iPhones. Even if they were just short of mobile i9 performance (they're not), this would still be a massive feat of engineering worthy of an upgrade.


AMD's Smart Memory Access was recently announced. In unoptimized games, they're projecting a 5% performance boost between their stock overclock and SMA (rumors put the overclock at only around 1%).

The bigger issue here is bandwidth. AMD hasn't increased their APU graphics much because the slow DDR4 128-bit bus isn't sufficient (let alone when the CPU and GPU are both needing to use that bandwidth).

I also didn't mention PCIe lanes. They are notoriously power hungry and that higher TDP chip not only has way more, but also has PCIe 4 lanes which have twice the bandwidth and a big increase in power consumption (why they stuck with PCIe 3 on mobile).

It's also notable that even equal cache sizes are not created equal. Lowering the latency requires more sophisticated designs which also use more power.

https://www.amd.com/en/technologies/smart-access-memory


Really interesting analysis.

So you expect the M1 MBP will outpace the more expensive Intel 13 MBP they are selling for more money!!

How would that not destroy sales of their “higher end” intel MBP 13?

If so do you think it will be better across most workload types?


I don't see why they would care that they are destroying sales of the intel MBP 13. Let consumers buy what they want - the M1 chip is likely far higher profit margins than the intel variant, and encouraging consumers to the Apple chip model is definitely a profit driver.

Some people especially developers may be skeptical of leaving x86 at this stage. I think the smart ones would just delay a laptop purchase until ARM is proven with docker and other developer workflows.

Another consideration - companies buying Apple machines will likely stay on Intel for a longer time, as supporting both Intel and ARM from an enterprise IT perspective just sounds like a PITA.


Docker runs in a VM on MacOS anyway, it’s not running on bare metal. Is instruction-set even relevant in that world?


That uses hardware virtualization which is very much dependent on the architecture. Running an x86 docker image on a M1 would take a significant performance penalty.


Thank you for posting this! I was so confused when folks were saying 16GB of ram was too little.

I run Linux on all my machines, and even running many (5-10) containers, 16GB was plenty. I now understand a bit better.


My Linux laptop locks up every now and then with swapping when I'm running our app in k3s; three database servers (2 mysql, 1 clickhouse), 4 JVMs, node, rails, IntelliJ, Chrome, Firefox and Slack, and you're starting to hit the buffers. I was contemplating adding more ram; 64 GB looks appealing.

I would not buy a new machine today for work with less than 32 GB.


Apple is notorious for releasing products that cannibalize their own products, I don’t think that concern would dissuade them.

There are likely many people who are not ready to switch yet either.

It has been a while since the ppc/x86 transition, but I want to say it was a similar situation then


The first 20” Intel iMac was released in the same chassis as the G5 iMac it was replacing with 2-3x the CPU speed. I beleive they continued to sell the G5 model for a brief while though for those that needed a PPC machine.


I think that’s the point? They want to show that M1 decimates comparable Intel chips even at lower price points.

This release is entirely within Apples control, why would they risk damaging their brand releasing a chip with lower performance than the current Intel chips they are shipping. They would only do this at a time when they would completely dominate the competition.


Impossible to know until we get the hard numbers.

But, just looking at A14 performance and extrapolating its big/little 2/4 cores to M1's 4/4; In the shortest tldr possible; Yes.

M1 should have stronger single-core CPU performance than any Mac Apple currently sells, including the Mac Pro. I think Apple's statement that they've produced the "world's fastest CPU core" is overall a valid statement to make, just from the info we independent third-parties have, but only because AMD Zen 3 is so new. Essentially no third parties have Zen 3, Apple probably doesn't for comparison, but just going on the information we know about Zen 3 and M1, its very likely that Zen 3 will trade blows in single core perf with the Firestorm cores in A14/M1. Likely very workload dependent, and it'll be difficult to say who is faster; they're both real marvels of technology.

Multicore is harder to make any definitive conclusions about.

The real issue in comparison before we get M1 samples is that its a big/little 4/4. If we agree that Firestorm is god-powerful, then can say pretty accurately say that its faster than any other four-core CPU (there are no four-core Zen 3 CPUs yet). There's other tertiary factors of course, but I think its safe enough; so that covers the Intel MBP13. Apple has never had an issue cannibalizing their own sales, so I don't think they really care if Intel MBP13 sales drop.

But, the Intel MBP16 runs 6 & 8 core processors, and trying to theorycraft what performance the Icestorm cores in M1 will contribute gets difficult. My gut says that M1 w/ active cooling will outperform the six core i7 in every way, but will trade blows with the eight core i9. A major part of this is that the MBP16 still runs on 9th gen Intel chips. Another part is that cooling the i7/i9 has always been problematic, and those things hit a thermal limit under sustained load (then again, maybe the M1 will as well even with the fan, we'll see).

But, also to be clear: Apple is not putting the M1 in the MBP16. Most likely, they'll be revving it similar to how they do A14/A14x; think M1/M1x. This will probably come with more cores and a more powerful GPU, not to mention more memory, so I think the M1 and i9 comparisons, while interesting, are purely academic. They've got the thermal envelope to put more Firestorm cores inside this hypothetical M1x, and in that scenario, Intel has nothing that compares.


> But, the Intel MBP16 runs 6 & 8 core processors, and trying to theorycraft what performance the Icestorm cores in M1 will contribute gets difficult.

Anandtech's spec2006 benchmarks of the A14 [0] suggest the little cores are 1/3 of the performance of the big ones on integer, and 1/4 on floating point. (It was closer to 1/4 and 1/5 for the A13.) If that trend holds for the M1's cores, then that might help your estimates.

[0] https://www.anandtech.com/show/16226/apple-silicon-m1-a14-de...


Many people will get the other model because they specifically want an Intel machine. I would expect it to be replaced in the coming months.


> Not all geekbench scores are created equal. Comparing ARM and x86 scores is an exercise in futility as there are simply too many factors to work through. It also doesn't include all workload types.

The meme of saying that Geekbench is not a useful metric across cores, or that it is not representative of real-world usage, and therefore cannot be used needs to die. It’s not perfect, it can never be perfect. But it’s not like it will be randomly off by a factor of two. I’ve been running extremely compute-bound workloads on both Intel and Apple’s chips for quite a while and these chips are 100% what Geekbench says about them. Yes, they are just that good.


In the case of Infinity Fabric: high speed I/O links generally consume gobs of power just to wiggle the pins, long before any control plane gets involved.

In this case, it's high speed differential signaling, and that's going to have a /lot/ of active power. There's a lot of C*dv/dt going on there!


I’m pretty sure the amount of cache on the chip was on one of the slides, according to Wikipedia it’s 192 kb for instructions and 128 kb for data.

To me it seems pretty unlikely to be that important because if you can have 16gb of memory in the chip, how hard can it be to increase the caches a fraction of that?


A normal SRAM cell takes 6 transistors. 6 * 8 * 1024 * 1024 * <total_mb> is a big number.

Next, SRAM doesn't scale like normal transistors. TSMC N7 cells are 0.027 nanometers while MY cells are 0.021 (1.35x). meanwhile, normal transistors got a 1.85x shrink.

I-cache is also different across architectures. x86 uses 15-20% less instruction memory for the same program (on average). This means for the same size cache that x86 can store more code and have a higher hit rate.

The next issue is latency. Larger cache sizes mean larger latencies. AMD and Intel have both used 64kb L1 and then move back to 32kb because of latencies. The fact that x86 chips get such good performance with a fraction of the L1 cache points more to since kind of inefficiency in Apples design. I'd guess AMD/Intel have much better prefetcher designs.


> 6 * 8 * 1024 * 1024 * <total_mb> is a big number.

No, when your chip has many billions of transistors that’s not a big number. For 1 mb that’s about 0.2%, a tiny number, also when multiplied with 1.85.

Next the argument is that x64 chips are better because they have less cache while before the Apple chips couldn’t compete because Intel had more. That doesn’t make sense. And how are you drawing conclusions on the design and performance of a chip that’s not even on the market yet anyway?


With 16MB of cache, that's nearly 1 billion transistors out of 16 billion -- and that's without including all the cache control circuitry.

Maybe I'm misunderstanding, but the 1.85x number does not apply to SRAM.

I've said for a long time that the x86 ISA has a bigger impact on chip design and performance (esp per watt) than Intel or AMD would like to admit. You'll not find an over-the-top fan here.

My point is that x86 can do more with less cache than aarch64. If you're interested, RISC-V with compact instructions enabled (100% of production implementations to my knowledge) is around 15% more dense than x86 and around 30-35% more dense than aarch64.

This cache usage matters because of all the downsides of needing larger cache and because cache density is scaling at a fraction of normal transistors.

Anandtech puts A14 between Intel and AMD for int performance and less than both in float performance. The fact that Intel and AMD fare so well while Apple has over 6x the cache means they're doing something very fancy and efficient to make up the difference (though I'd still hypothesize that if Apple did similar optimizations, it would still wind up being more efficient due to using a better ISA).


I’m sure you are well versed in this matter, definitely better than I am. You don’t do a great job of explaining though, the story is all over the place.

I’ll just wait for the independent test results.


That's 16GB of DRAM. Caches are SRAM, which has a very different set of design tradeoffs. SRAM cells use more transistors so they can't be packed as tightly as DRAM.


16gb is 500 times as much as 32mb. Cache memory is not 500 times as large as normal memory.

If 32mb would be too hard they could have easily went for 1mb.

But they didn’t and that’s a pretty good indicator it doesn’t make a lot of difference.


The Ryzen 9 is a 16 (32 hyperthreads) core CPU, the A14 is a 2 big / 4 little core CPU. Power draw on an integer workload per thread seems roughly equivalent.


Last time I checked the TDP/frequency curve of Apple's chips was unimpressive [0]. If you crank it up to 4Ghz it's going to be as "bad" or even worse than a Ryzen CPU in terms of power consumption per core. There is no free lunch. Mobile chips simply run at a lower frequency and ARM SoCs usually have heterogenous cores.

>but this is an architecture that the PC market needs to be scared of.

This just means you know nothing about processor performance or benchmarks. If it was that easy to increase performance by a factor of 3x why hasn't AMD done so? Why did they only manage a 20% increase in IPC instead of your predicted 200%?

[0] https://images.anandtech.com/doci/14892/a12-fvcurve_575px.pn...


Isn’t this a RISC vs CISC thing though? AMD and Intel use a lot of complicated stuff to emulate the x86 instructions with a VM running on a RISC core. Apple controls hardware and software including LLVM so they compile efficient RISC code right from the get go. It’s their vertical integration thing again.


You got a point in the optimization part, it's difficult to compare both chips when you're running on completely different OSs, specially when one of them ruins a specially optimized OS like ios

The closest it could get I think would be running a variant of Unix optimized for the Ryzen.



Don't just downvote me, explain. I'm asking because I'm interested. Why is my hypothesis wrong?


My point was this is marketing spin.

There's no way to say what "faster" means or what 98% they used. Is it faster than 98/100 models? or faster than 98% of the 261M laptops sold in 2019?

"faster than 98% of laptops sold last year" is a nice quotable soundbyte that will be spread without any of those details.

I'm actually a big fan of Apple's products. I'm sure all the numbers, R&D, and benchmarking will prove the M1 to be impressive.


I don't know much about this stuff but if all this is true then I don't see what's holding AMD / Intel back. Can't they just reverse-engineer whatever magic Apple comes up with, and that will lead to similar leaps in performance?


This wording is very hand-wavey. They can and should give a direct comparison; as consumers we don't know what the top 2% of sold laptops are, or which is the "best selling PC laptop in its class". Notice it's "best selling" and not most powerful.


> "faster than 98% of PC laptops sold in the last year"

Keep in mind that the $3k MBP is probably part of the 2% in the above quote. The large majority of laptops sold are ~$1k, and not the super high end machines.


"98% of laptops sold" is not a useful metric.

If you buy a Ferrari, they won't market saying its latest car is better than 98% of cars sold last year.

it is not a good metric to compare performance of budget laptops where sales is going to be higher to premium laptop.


Just awhile ago Apple said the AIR was in a class of it's own so....


So 2% of the airs sold last year are equal to or better than this.


The vast majority of PC laptops sold are cheap low-end volume computers.


Yeah, I just thought about all those 499/599 Euro (factory-new) laptops that are being sold in Germany. They are okayish but nothing I'd like to compare great systems against.


Really good observations - I suspect that the 16" MBPs may be in the 2% though.

Plus, given use of Rosetta 2 they probably need 2x or more improvement over existing models to be viable for existing x86 software. Interesting to speculate what the M chip in the 16" will look like - convert efficency cores to performance cores?


My guess is that they'd still keep the efficiency cores around, but provide more performance cores. So likely a 12 or 16 core processor, with 4 or 6 of those dedicated to efficiency cores.

The M1 supposedly has a 10w TDP (at least in the MBA; it may be speced higher in the MBP13). If that's the case, there's a ton of power envelope headroom to scale to more cores, given the i9 9980HK in the current MBP16 is speced at 45 watts.

I'm very scared of this architecture once it gets up to Mac Pro levels of power envelope. If it doesn't scale, then it doesn't scale, but assuming it does this is so far beyond Xeon/Zen 3 performance it'd be unfair to even compare them.

This is the effect of focusing first on efficiency, not raw power. Intel and AMD never did this; its why they lost horribly in mobile. Their bread and butter is desktops and servers, where it doesn't matter. But, long term, it does; higher efficiency means you can pack more transistors into the same die without melting them. And its far easier to scale a 10 watt chip up to use 50 watts than it is to do the opposite.


I'm old enough to remember the MacBook Pro intro (seems a long time ago!) when Steve Jobs said it's all about performance per watt.

My only worry about the systems with more cores (Mac Pro etc) are about the economics for Apple of making these chips in such small volume.

PS Interesting that from Anandtech the M1 has a smaller die area than the i5/i7 in the Intel Airs so plenty of room for more cores!


>This is the effect of focusing first on efficiency, not raw power. Intel and AMD never did this; its why they lost horribly in mobile.

If you want a more efficient processor you can just reduce the frequency. You can't do that in the other direction. If your processor wasn't designed for 4Ghz+ then you can't clock it that high, so the real challenge is making the highest clocked CPU. AMD and Intel care a lot about efficiency and they use efficiency improvements to increase clock speeds and add more cores just like everyone else. What you are talking about is like semiconductor 101. It's so obvious nobody has to talk about it. If you think this is a competitive edge then you should read up more about this industry.

>Their bread and butter is desktops and servers, where it doesn't matter.

Efficiency matters a lot in the server and desktop market. Higher efficiency means more cores and a higher frequency.

>But, long term, it does; higher efficiency means you can pack more transistors into the same die without melting them.

No shit? Semiconductor 101??

>And its far easier to scale a 10 watt chip up to use 50 watts than it is to do the opposite.

You mean... like those 64 core EPYC server processors that AMD has been producing for years...? Aren't you lacking a little in imagination?


Have no considered being less snarky as you careen your way through that comment without really understanding any of it? The efficiency cores on Apple’s processors perform tasks that the main, more power hungry processors aren’t necessary for, which a profoundly different situation from a typical server that is plugged in and typically run with capacity all the time. Honestly, besides being against the rules, the swipes you make against the commenter “not knowing about the industry” are just shocking to see considering how much you missed the point that was being made.


Without actual real world performance tests from neutral 3rd parties all that really really need to be taken with a grain of salt. Never trust the maker of a product, test for yourself or watch for neutral reviewing parties who don't have a financial incentive.


Remember that the Intel CPUs have shitty thermals. 1.4 GHz with turbo up to 4, etc.

Smacking that with single cores on a 5nm process is probably pretty easy.


The reason is almost certainly that its the lowest power laptops is where the volume is and because it makes absolute sense to go for that part of the market first.

I'm really impressed by what they've done in this part of the market - which would have been impossible with Intel CPUs.


I think the main reason is discrete GPUs. I’m guessing they are currently working on PCI Express for M1. That or a much faster GPU on a different version of the SoC.


Based on what Apple says. Is there an independent cross platform benchmark which is actually relevant and that says the same?


No, because it hasn't shipped yet.


For how often they repeated "faster", I was surprised to see them announce no benchmarks with some sort of baseline or context. Fast than what at doing what?


A number of times when they said that I though, "but, that is a GPU heavy workload". Its not clear if they were comparing to existing mac's or more recent ryzen/intel xe machines. Both of which have very significant GPU uplifts with their integrated GPUs.

AKA, I suspect a big part of those perf uplifts evaporate when compared with the laptops that are appearing in stores now.

And the ipad benchmarks also remain a wait and see, because the perf uplifts I've seen either are either safari based, or synthetic benchmarks like geekbench. Both of which seem to heavily favor apple/arm cores when compared with more standard benchmarks. Apples perf claims have been dubious for years, particularly back when the mac was ppc. They would optimize or cherry pick some special case which allowed them to make claims like "8x" faster than the fastest PC, which never reflected overall machine perf.

I want to see how fast it does with the gcc parts of spec..


They are comparing to the MacBook processors which they kept as ancient 8th gen Intel quad cores.


Never rely on first party claims. Always wait for 3rd party benchmarks.


Did they say what they meant when they said faster?

I saw some graphs that looked unrealistically smooth with two unlabeled axis


The literally benchmarked the slowest 10th Gen i7 CPU on the market.

And I got so excited about it...


Which means that, according to Apple, it's better than the CPUs in the other laptops it's competing against. That slow i7 is typical for thin+light 13" machines. I fail to see how this is a bad thing or even sneaky marketing.


Intels marketing is more at fault here. i7 cpus range from slower 4 core 1.8ghz thin and light cpus to high end desktop 8 core 3.5ghz ones.

When the name i7 is mentioned people are more likely to think it's one of the high end desktop processors rather than the mobile processors.


If they targeted exclusively the consumer or media editor market - sure.

But then they throw in code compilation... that's where they stop being honest.


They also said it's the fastest laptop CPU?


"in its class"


Classic Apple claim


So you're disappointed that the MacBook Air doesn't beat the CPU and GPU performance of a hulking "desktop replacement" style notebook?


The complaint is that Apple interprets “in its class” to mean literally the slowest mobile processor of the prior generation made by Intel. It’s not even “fastest in the class of low-priced value notebooks for everyday use” which is what a reasonable person might interpret that to mean.


The Pro is quoted as 2.8x the 1.7 GHz i7.

Not Intel's fastest chip but, if these benchmarks are correct, that's a big speed bump and the higher end Pro's are to come. How fast does it have to be in this thermal envelope to get excited about?


I'd wait for actual benchmarks just judging by the amount of parachute clauses in the disclosures. Power consumption numbers should be great but I'm guessing it falls flat on some compute tasks that can't leverage custom instructions.


From Anandtech:

"Apple claims the M1 to be the fastest CPU in the world. Given our data on the A14, beating all of Intel’s designs, and just falling short of AMD’s newest Zen3 chips – a higher clocked Firestorm above 3GHz, the 50% larger L2 cache, and an unleashed TDP, we can certainly believe Apple and the M1 to be able to achieve that claim."

Not sure what the custom instructions you're referring to are.


How many cores does that have, and how many cores does the i7 have though?


Why is this a mystery or rhetorical question? Look on the website? It's public:

M1

> Apple M1 chip

> 8-core CPU with 4 performance cores and 4 efficiency cores

> 8-core GPU

> 16-core Neural Engine

Intel

> 1.7GHz quad-core Intel Core i7, Turbo Boost up to 4.5GHz, with 128MB of eDRAM

https://www.apple.com/macbook-pro-13/specs/

https://www.apple.com/shop/product/G0W42LL/A/refurbished-133...

Apple's claim:

> With an 8‑core CPU and 8‑core GPU, M1 on MacBook Pro delivers up to 2.8x faster CPU performance¹ and up to 5x faster graphics² than the previous generation.

https://www.apple.com/shop/buy-mac/macbook-pro/13-inch-space...

Fine print:

> Testing conducted by Apple in October 2020 using preproduction 13‑inch MacBook Pro systems with Apple M1 chip, as well as production 1.7GHz quad‑core Intel Core i7‑based 13‑inch MacBook Pro systems, all configured with 16GB RAM and 2TB SSD. Open source project built with prerelease Xcode 12.2 with Apple Clang 12.0.0, Ninja 1.10.0.git, and CMake 3.16.5. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.

> Testing conducted by Apple in October 2020 using preproduction 13‑inch MacBook Pro systems with Apple M1 chip, as well as production 1.7GHz quad‑core Intel Core i7‑based 13‑inch MacBook Pro systems with Intel Iris Plus Graphics 645, all configured with 16GB RAM and 2TB SSD. Tested with prerelease Final Cut Pro 10.5 using a 10‑second project with Apple ProRes 422 video at 3840x2160 resolution and 30 frames per second. Performance tests are conducted using specific computer systems and reflect the approximate performance of MacBook Pro.


The 2019 MBP 13 supposedly uses an 8th gen, 14nm Intel part (14nm is 6 year-old technology).

A more fair comparison would be Tiger Lake (20% IPC improvement) on Intel's terrible 10nm process. The most fair comparison would be zen 3 on 7nm, but even that is still a whole node behind.


> The 2019 MBP 13 supposedly uses an 8th gen, 14nm Intel part (14nm is 6 year-old technology). A more fair comparison would be Tiger Lake (20% IPC improvement) on Intel's terrible 10nm process. The most fair comparison would be zen 3 on 7nm, but even that is still a whole node behind.

I'm not sure exactly what your point is. The fact it is IT IS on a better node. That's probably a big part of the performance picture. But, as a user I don't care if the improvement in performance is architectural or node. I just care that it's better than the competition.


I agree, but the real question becomes whether users would prefer 11th gen Intel processors or zen 2/3 over what they're shipping now. Would you take AMD seriously if they started talking about how their 5950 beat an ancient 8th gen Intel processor? I expect a higher standard from hardware makers and don't give a free pass to anyone over that kind of shenanigan.


We'll see third party benchmarks in less than a week and then this discussion is moot.


Do you really? It’s nice if it’s faster (and a hard sell if it’s slower) but really most people aren’t going to do the same things on a Mac they are going to do on a Windows laptop. You can’t run Windows or Windows software on these Macs. And it’s really no use to compare for instance Office for Windows to Office for Mac.


It depends on how you define fair. The base level M1 MBP is the same price as the previous generation Intel MBP and has the same amount of ram and storage. So in terms of a generation-on-generation comparison, it's absolutely fair for Apple to position the benchmarks that way.


Yes, Wikipedia says it's i7-8557U which has 8 MB cache.

So comparing a

4-Core, 5nm, 16 MB cache, 10 watts TDP vs

4-Core, 14nm, 8 MB cache, 15 watts TDP.

https://ark.intel.com/content/www/us/en/ark/products/192996/...


4 + 4 vs 4 hyperthreaded I think. Not sure I follow the point though.


The new Apple chip has more cores, so faster performance doesn't necessarily mean faster single-core performance. Faster on a parallel benchmark doesn't necessarily translate into faster performance for most real-world scenarios. That is all. That being said, I do believe Apple has very competent chip designers.


From Anandtech

Apple claims the M1 to be the fastest CPU in the world. Given our data on the A14, beating all of Intel’s designs, and just falling short of AMD’s newest Zen3 chips – a higher clocked Firestorm above 3GHz, the 50% larger L2 cache, and an unleashed TDP, we can certainly believe Apple and the M1 to be able to achieve that claim.

Thats per core performance.


Also unclear what they even meant when comparing performance to the “highest selling PC”, I’m not even sure if analysts could tell you what the top PC on the market is


Only for the Air, the M1 Pro is tested against the equivalent i7 Pro so I assume not throttled as much.


Sorry to ask, what is the "Fabric" module in their architecture? Also, is the "Neural Engine" some neural network specific processor?


The "Fabric" is the interconnect between different components on the chip. It's what the different units use to communicate and transport data.

The "Neural Engine" is a special compute unit specifically for machine learning. The iPhone Camera app uses it for semantic segmentation, for example.


Fabric is a common term for describing on-chip interconnects in a system-on-a-chip.

It's all the wires and buffers and pipeline registers and arbiters and decoders and muxes that connect the components together.

It's dominated by wires, typically, which is probably how it came to be known as a fabric. (Wild speculation on my part.) I've been hearing that term for a long time. Maybe 20 years?


The Neural Engine in iPhones is used for things ranging from photography enhancement to facial recognition for the unlock and Siri.


I wondered this too. In my experience, "fabric" typically refers to the programmable logic region in a FPGA-based SoC (such as the Xilinx Zynq). I'm assuming the M1 does not have this, so I'm not sure what it means in this context.


I think Apple has done an amazing job of pulling off their own silicon. At Sun I got to be involved peripherally in the effort that created SPARC and it was way more work than I had suspected. So kudos to Apple's design, engineering, and QA teams!

I am also noting that they have killed the "ARM SoC's will never match what Intel can do" story line. That always reminded me of the "Microcomputers will never be able to do what minicomputers can do" story line that everyone except DEC and HP seemed to realize wouldn't hold up over time. "Never say never" is a good motto in the computer business.

That said, as the integration continues apace re-usability and repairablilty reach all time lows. When you consider that you could take a 1970's minicomputer, disassemble it to the component level, and create an entirely different (and functional) computer out of those components, you get a sense of how quaint that notion seems these days. I doubt anyone will ever desolder an M1 chip and put it onto a new board of their own design. And that reminds me more of the IBM "die on substrate" modules that they started using in mainframes which made them basically nothing more than low grade gold ore if they no longer worked as a mainframe.


I agree about the upgradability/repairability. However, if you look at like a cell phone maybe it makes more sense. No one has ever complained that you couldn't update the RAM or the CPU in a iPhone. In order to upgrade the hardware of your cell phone you have to purchase another one. I'm guessing Apple is looking at the Mac hardware in the same light. "Want more power? Buy a new laptop!"

The more interesting question is how does this affect the Mac desktops? It would be a shame to throw out a whole Mac Pro machine just because it only has 64GB RAM. Then again, this event did release the Mac Mini, which is a desktop and does not have upgradable RAM or storage or CPU. Hmmmm...


> I'm guessing Apple is looking at the Mac hardware in the same light. "Want more power? Buy a new laptop!"

How often do phone users actually upgrade their phones for more power? I mostly use my computer to develop software, play video games, and edit media, and stream videos. I mostly use my phone to text, play music, and access the internet.

Almost every PC purchase I've ever made was driven by a need for better performance, but I've never upgraded my phone for that reason. I've upgrade my phone for a better screen, a better camera, a bigger battery, software support, new hardware features (4G, biometrics, USB-C, etc)... but never because I felt like I needed a faster processor or more system memory.


The iPhone has a very smooth process of upgrading the phone. You turn on the new iPhone next to your old one, it prompts and voila. It's like upgrading hardware of your current phone.

I'm unsure if we're at a point where these things are even feasible. e.g I have configured my MBP a lot, the .zsh file, home directory and endless configuration.

I would not be surprised if they ditch the kernal completely, and make everything an App (much like chromebook)


You can do something similar with Macs. Using Migration Assistant I was able to transfer all my data (.zsh, git repos, configs, Docker containers, everything.) Took less than an hour. In fact, I did this because I needed to upgrade my RAM


Yup, did this when I upgraded from a 2013 MacBook Air to a 2018 MacBook Pro. Everything worked great except for homebrew, which unfortunately had to be nuked and re-installed; but I can't really blame Migration Assistant for that.


I thought, the parent comment was talking about hardware upgrade on the same laptop.:)


I don't know about the new model but the 2018/2020 MacMini does have upgradeable memory.


iMac as well.

Also, the previous Mac Mini has 64G max memory and not the laughable 16G.


I think that is a good observation, perhaps at some point the motherboard becomes the swappable part?


It could be done but companies rather want you to buy a complete computer again and again. Same with cars. That way they sell more stuff and earn more. With proper environmental marketing tricks like the recent one with chargers and recycling robot lies, environmental impact is hidden.


Unlikely, it'll be as expensive as just buying a new one and it's unlikely Apple would offer full motherboard replacements easily.


I think that the future of computer customization is in virtualization. Instead of having a single very expensive machine to upgrade, we just go grab 30 servers from a bulk seller and use something like Xen or Proxmox to use them all as a single machine with 1 virtual CPU, 1 virtual disk and 1 virtual network interface. Got another computer? Just add it to the pool. So while computers are getting harder to disassemble, it's also getting easier to pool them together to share compute, network and storage.


Apropos SPARC, one of the most interesting features of that ISA (apart from mainstreaming register windows) was the tagged arithmetic instructions. Does anybody know why they didn't have much success in later ISAs?


Apple mentions TensorFlow explicitly in the ongoing presentation due to the new 16-core "Neural Engine" embedded in the M1 chip. Now that's an angle I did not expect on this release. Sounds exciting!

Edit: just to clarify, the Neural Engine itself is not really "new":

> The A11 also includes dedicated neural network hardware that Apple calls a "Neural Engine". This neural network hardware can perform up to 600 billion operations per second and is used for Face ID, Animoji and other machine learning tasks.[9] The neural engine allows Apple to implement neural network and machine learning in a more energy-efficient manner than using either the main CPU or the GPU.[14][15] However, third party apps cannot use the Neural Engine, leading to similar neural network performance to older iPhones.

Source: https://en.wikipedia.org/wiki/Apple_A11#Neural_Engine


I can't see how something that tiny can compete in any meaningful way with a giant nVidia type card for training. I'd imagine it's more for running models that have been trained already, like all the stuff they mentioned with Final Cut.


Not all NN models are behometh BERTs, U-Nets or ResNets. Person detection, keyword spotting, anomaly detection... there are lots of smaller neural nets that can be accelerated by a wide range of hardware.


For an 18-hour battery life computer (Macbook Air) that now doesn't even have a fan, it's for a complete different market segment from where nvidia cards dwell.


Yeah I would imagine it's intended for similar use-cases as they use for iOS - for instance image/voice/video processing using ML models, and maybe for playing around with training, but it's not going to compete with a discreet GPU for heavy-duty training tasks


It is just not meant to be a training device so comparing with data Center or developer GPUs is useless. Faster inference for end-users is what is mentioned by Apple and the only use case where this hardware makes sense.


Isn't it better to rent a cloud with as many GPUs as necessary for a time needed to train the model? I don't know state of things in ML.


Not necessarily.

It can be surprisingly cost-effective to invest a few $k in a hefty machine(s) with some high-end GPU's to train with due to the exceedingly hefty price of cloud GPU compute. The money invested up-front in the machine(s) pays itself off in (approximately) a couple of months.

The "neural" chips in these machines are for accelerating inference. I.e. you already have a trained model, you quantise and shrink it, export it to ONNX or whatever Apple's CoreML requires, ship it to the client, and then it runs extra-fast, with relatively small power draw on the client machine due to the dedicated/specialised hardware.


For productionizing/training massive models, yes.

But in the development phase, when you are testing on a smaller corpus of data, to make sure your code works, the on-laptop dedicated chip could expedite the development process.


I agree with the parent poster that it's probably more about inference, not training.

If ML developers can assume that consumer machines (at least "proper consumer machines, like those made by Apple") will have support to do small-scale ML calculations efficiently, then that enables including various ML-based thingies in random consumer apps.


Cloud GPU instances are very expensive. If you get consumer GPUs not only do you save money, you can sell them afterwards for 50% of the purchasing price.


Curious to hear responses to this too..


Sure, inferencing doesn't need floating point instructions, so NVIDIA will stay the only real solution for desktop/laptop based model training for a long time.


Can anyone who knows about machine learning hardware comment on how much faster dedicated hardware is as opposed to, say, a vulkan compute shader?


On the NVidia A100, the standard FP32 performance is 20 TFLOPs, but if you use the tensor cores and all the ML features available then it peaks out at 300+ TFLOPs. Not exactly your question, but a simple reference point.

Now the accelerator in the M1 is only 11 TFLOPs. So it’s definitely not trying to compete as an accelerator for training.


That depends entirely on the hardware of both the ML accelerator and the GPU in question, as well as model architecture, -data and -size.

Unfortunately Apple was very vague when they described the method that yielded the claimed "9x faster ML" performance.

They compared the results using an "Action Classification Model" (size? data types? dataset- and batch size?) between an 8-core i7 and their M1 SoC. It isn't clear whether they're referring to training or inference and if it took place on the CPU or the SoC's iGPU and no GPU was mentioned anywhere either.

So until an independent 3rd party review is available, your question cannot be answered. 9x with dedicated hardware over a thermally- and power constrained CPU is no surprise, though.

Even the notoriously weak previous generation Intel SoCs could deliver up to 7.73x improvement when using the iGPU [1] with certain models. As you can see in the source, some models don't even benefit from GPU acceleration (at least as far as Intel's previous gen SoCs are concerned).

In the end, Apple's hardware isn't magic (even if they will say otherwise;) and more power will translate into higher performance so their SoC will be inferior to high-power GPUs running compute shaders.

[1] https://software.intel.com/content/www/us/en/develop/article...


It's kind of sad that AMD spent years to get mediocre TensorFlow support and Apple walks in with this. It really shows how huge Apple is.


In fairness, it's been possible to convert a TensorFlow model to a CoreML model for a while, and in April TensorFlow Lite added a CoreML delegate to run models on the Neural Engine.

https://blog.tensorflow.org/2020/04/tensorflow-lite-core-ml-...

So don't think of it as Apple walked right in with this so much as Apple has been shipping the neural engine for years and now they're finally making it available on macOS.


But they needed to, MacBooks are simply no option if you want to train models. I dont expect crazy performance but would be great if MacBooks would be an option again for prototyping / model development at least


Who mentioned training? Most of these chips are only any good for inference. A wonderful symphony with the Apple computing mantra, of course.


Totally agree.

Besides, can the neural engine be used to speed up other tasks?


Could this also be remedied by Apple supporting Nvidia GPUs again? And then you plug in a beefy eGPU?


I think it is interesting that Apple is at Nvidias mercy again with the Nvidia ARM deal. Hopefully they get their shit together


I think this is for inference not learning, even though they use the term machine learning. They seem to just mean running models based on machine learning approaches.

Tensorflow includes stuff for inference.


My thinking was along the same line as yours, but the way apple framed it seems to suggest that the M1 accelerates model training and not just inference. Here's the actual quote "It also makes Mac Mini a great machine for developers, scientist and engineers. Utilizing deep learning technologies like tensorflow ... which are now accelerated by M1". It should be pretty straightforward to test this though: installing tensorflow-gpu on a mac mini and seeing the result. I suspect, TF's latest branch should also indicate which GPUs are supported. Curious to hear more thoughts on this.


They do not have any hardware combination which can actually support even modest GPU intensive training sadly, so much touting running models instead of training.


Yeah, I was confused at that implication. I don't think these are designed for training!

(If they are or can be, I'm interested)


> (If they are or can be, I’m interested)

Exactly. Currently I am training my models using Google Colab and then exporting the model to run on my MBP. Would be interesting if I could do it locally

Another interesting thing is that ( if this is for training ) this will become the only accelerated version of Tensorflow for macOS as: - No CUDA drivers for latest macOS - AMD ROCHm only supports Linux runtime


I’m hoping the M1 can be used for prototyping with small data sets, then final training on Google Colab with complete data sets.


Is this just opinion? Maybe they are designed ALSO for training. I wonder if this things can replace nVidia graphic cards on training? The neural core has a LARGE area on the chip design similar to the GPU area.


Yes, need to see more strong evidence that the new MBP's can handle large amounts of ML Training using TF or CreateML so we don't have to get NVIDIA machines/laptops.


Would love to get GPU accelerated learning in Lobe (http://www.lobe.ai).


No 32GB options on any of the models announced today is a shame. Or not I guess as it means I won't be buying anything today

M1 looks impressive, hopefully it scales as well to higher TDPs.

Pretty much what I expected today with all the low power models getting updated.

With no fan I expect the Air will throttle the CPU quickly when using the high performance cores. Calling it an "8-core CPU" is a little cheeky IMHO but I guess everyone does it.

Looking at the config options it seems the CPU/SoC is literally the "M1" and that's it. No different speeds like before so apparently the same CPU/SoC in the Pro and the Air? I guess the fans in the Pro will allow the high performance more headroom but still kinda odd. I wonder if that will carry over to the 16" MBP next year?

I am disappointed to see the design is identical to what we currently have. I was hoping we would see a little refinement with thinner screen bezels, perhaps a slightly thicker lid to fit in a better web cam (rather than just better "image processing") and FaceID.

Overall I am disappointed due to 16GB max RAM and the same old design. I guess we will see the bigger machines get updates sometimes between March and July 2021.

Also are we calling it the M1 CPU or the M1 SoC?


> I am very disappointed to see the design is identical to what we currently have

Yeah one possible explanation I can think of is that they're not 100% confident in how rollout will go, and they want to avoid the situation where people in the wild are showing off their "brand new macs" which are still going through teething issues. There's less chance of souring the brand impression of M1 if they blend into the product line better.

Alternatively, maybe they want to sell twice to early adopters: once for the new chip, and again in 12-18 months when they release a refreshed design.


I think it could also be that a lot of people will be 'calmed' by the design they're used to when you're trying to convince them of a new chip the average consumer might not understand


I don't know about that, I think the confident move would have been to release the new chips along with a big design update.

As you say, I think the average consumer doesn't understand the difference between an Intel chip and an Apple chip, and will probably not understand what if anything has changed with these new products.

I would say developers would be the group which would be most anxious about an architecture change (which is probably why this announcement was very technically-oriented), and developers on average are probably going to understand that design changes and architecture changes basically orthogonal, and thus won't be comforted that much by a familiar design.

On the other side, average consumers probably aren't all that anxious due to the arch change, and would be more convinced that something new and exciting was happening if it actually looked different.


> developers on average are probably going to understand that design changes and architecture changes basically orthogonal

If the reaction here to the touchbar is representative, perhaps they didn't want the M1 announcement to be overshadowed by the new feature related whining.


Probably a good mix of both. No need to send out a completely new design along with the new chips, plus the designs are really not that outdated. In any case these are squarely targeted at the "average consumer" market - lower budget, keeps computers for a very long time, places significant value on the higher battery life.

Less reliant on perfect compatibility of enterprise software, which I am sure is something they want a little more time to sort out before committing their higher end lineup.


> No 32GB options on any of the models announced today is a shame. Or not I guess as it means I won't be buying anything today

I think this was the fastest way to get something out quickly and cover as many SKUs as possible.

We'll need to wait for the next/larger SKU models to get out of the 16B / 2TB / 2-port limitations. Maybe they should call those devices the Macbook Pro Max.

I wonder how many more SKUs will be needed. Obviously they are going to minimize them, but will they be able to cover Macbook Pro 16", iMac, iMac Pro and Mac Pro with only one (unlikely) or two (likely) more SKUs?


When Apple does eventually produce larger than 16GB / 2TB / 2-port Macs I hope they don’t name them MacBook Pro Max. As long as they are called Macs this naming convention seems silly. I’d prefer that we don’t have any Max Macs.


> I am very disappointed to see the design is identical to what we currently have

Lot's of moving parts in a single generation is risky for scheduling and changing CPU architecture is huge enough for a generation change.


Maybe they really really don't want professionals to buy the 1st gen M1 Macs, because we have the toughest use cases and will shit on it the hardest if something goes wrong. Lack of 32GB seems like a pretty clear signal.

The Mac has turned into a very nice luxury-consumer business. It is a status signal even if its just used to check email and watch YouTube. It's a pretty smart idea to let these users be the gen 1 guinea pigs.


Also just two ports on a PRO model. C‘mon there is plenty of room on both sides. Maybe lets it look elegant but then the cable mess will be on the desk, thanks.


This MBP refresh replaces the lower-end MBP which also only had 2 ports.

This is a small detail that is easy to miss.

The "higher-end" MBPs with 4 ports are only availabe with the intel CPUs for now


Those 8 extra PCIe lanes eat a lot of power. It's worth noting that Intel's processor they used in the 2019 MBP 13 had 16 PCIe lanes (I think the chipset added more too), but this chip probably has 10-12 at the absolute most.

The "we integrated thunderbolt" bit is also implying an untruth that nobody else does that. Intel did that last year with Ice Lake. AMD will do that next year when they upgrade their IO die.


It was just mentioned in passing, and they did not imply that the others did not. It’s just the answer to an obvious question about these devices: “Does it support Thunderbolt?” “Yes, it’s integrated into the M1”.

You are reading too much into this.


Are there any other ARM devices with Thunderbolt? I am not aware of any.


No, but that's not relevant. Their x86 to ME graphic shows all the things merging together.

The point is that it's mostly Apple's fault to begin with. They choose to use outdated processors without integrated thunderbolt. They chose to add their massive A10 die instead of a normal, optimized SSD controller (not to mention this screws over data recovery and SSD replacement). They are exaggerating at best about RAM as the chips merely sit beside the core. The only thing they should really claim is the IO die and even that's nothing too special.


> Also are we calling it the M1 CPU or the M1 SoC?

The M1 Chip: https://www.apple.com/mac/m1/


> Or not I guess as it means I won't be buying anything today

Honestly, people who don't use their computers as Facebook machines or have very specific things they want to do with the ML engine should probably stay away from Arm Macs for a couple years. There's bound to be plenty of software that doesn't work or doesn't work well, Docker will have ~limited usefulness until more and more people publish Arm images, etc. Plus I'm waiting for the typical Apple move of screwing early adopters by announcing that next year's model is 3x as fast, $200 cheaper, etc.


Hmmm... What is the day 1 ability for, say, a systems programmer to run VSCode/golang development locally? It'd be neat to find out. The fact that they don't say doesn't tell me much, either it's "not a problem" or it's "such a problem we won't talk about it".


The Visual Studio Code team tweeted that they are targeting a release for the Insiders channel by the end of November.

https://twitter.com/code/status/1326237772860981249?ref_src=...


Both should work in Rosetta for now and I expect them to become native in the next few months.


> Plus I'm waiting for the typical Apple move of screwing early adopters by announcing that next year's model is 3x as fast, $200 cheaper, etc.

They aren’t screwed, the early adopters adopt the new model early by definition. They get the new model and the move on to the next new model.

I know several of these bold individuals.


I didn't watch the event, but if you have an Intel MacBook, will everything keep working as it is now? Even when upgrading to Big Sur?


Yeah, there’s no change for Intel Mac owners really. Given Apple will still be selling Intel Macs through 2021 and likely 2022, I’d expect them to be fully supported through something like 2026/7 (though your exact model may vary depending on its current age).


Such a good point. I wonder how many folks don't actually realize there's a different ecosystem for Docker x86 and ARM images.


Will Apple to transparent emulation (x86-on-ARM) or just not run x86 binaries?


Rosetta 2 is a x86_64 to ARM64 translation layer.

https://developer.apple.com/documentation/apple_silicon/abou...

Some things still won't run though. Among other issues, there's a page size difference (16 kb vs 4 kb) that bit Chrome (since fixed): https://bugs.chromium.org/p/chromium/issues/detail?id=110219...


> Some things still won't run though. Among other issues, there's a page size difference (16 kb vs 4 kb) that bit Chrome (since fixed): https://bugs.chromium.org/p/chromium/issues/detail?id=110219...

Rosetta 2 on shipping hardware uses 4kb pages; it was only the DTK that was 16kb-only.


Especially since it’s unified memory between the CPU and GPU - no wonder they were showing off mobile games on a desktop device


13" has always had unified memory.


I'm very curious to see some real benchmarks. I have no doubt that they've achieved state of the art efficiency using 5nm, but I have a harder time believing that they can outperform the highest end of AMD and Intel's mobile offerings by so much (especially without active cooling). Their completely unit-less graph axes are really something...


Read the fine print at the bottom, the claims they are making seem very..... not backed up by their testing. They chip they say they beat that makes them the best in the world is the last gen mackbook with an i7


Their claimed 3x faster is massive even if the baseline is low. And they also achieved that vs. the Mac Mini's 3.6GHz i3. So I don't think fastest singlethread performance is an outlandish claim.

Historically speaking, Apple doesn't underdelivier on their claimed numbers. I'm excited to see what they'll do with a full desktop power budget.


Which was severely underpowered because the cooling solution was knee-capped. This is partly Intel's fault, but they made their last-gen so underwhelming that this _had_ to be better


If they could outperform AMD/Intel today, they would've offered 16" MBPs / iMacs / Mac Pros. Today, they have power-efficient chips, so they converted their low-end machines, and Mini so that companies porting their software can start migrating their build farms.


I don't see why they need to rush into the higher end. Let ARM chips prove themselves for consumer scale. Enterprise is going to be weary of the new tech and potential compatibility issues. Proving it in the consumer market makes far more sense, and they really get to flex the benefits in something like a macbook air. Fanless, silent, and huge battery life increases are really tangible upgrades.

A beefy chip in an MBP 16 will be exciting, and I'm sure they are working on that, but it will make far more sense once ARM is established in the "normal" consumer models.


They probably beat AMD/Intel on perf/power efficiency, which is why it makes sense for MBA and 13" MBP.

The smaller machines are also likely held back by cooling solutions, so if you have Intel beat on power efficiency in a tiny form factor, you can boost your clock speed too.


Considering how AMD also beats intel on power/perf by a wide margin and they compared their results against Intel CPUs I wouldn't be surprised if their power/performance was close to AMD (they do have heterogeneous CPU cores which of course is not the case in traditional x86)


Apple's perf/watt is much higher than AMD. Anandtech has the A14 within a few percent of the 5950X on single threaded workloads. Power consumption is 5 watts (entire iPhone 12) versus 49 watts (5950X package power).

Normalizing for performance (say A14 vs 5950X at 4.5GHz instead of 5GHz) would close the gap somewhat, but it's still huge. Perhaps 4x instead of 10x - those last 500MHz on the AMD chip cost a ton of power.

Of course, none of this is particularly surprising considering Apple is using both a newer process and gets nearly 60% higher IPC.

[1]: https://www.anandtech.com/show/16226/apple-silicon-m1-a14-de...


You are comparing a desktop processor that's been tuned to overboost as much as thermals will allow it to a mobile phone CPU ?!?

AMD has a laptop series of CPUs, M1 compared to that is the only relevant comparison - the rest is apples to oranges.


AMD's Smart Memory Access would like to have a word. I'd note that in unoptimized games, they're projecting a 5% performance boost between their stock overclock and SMA (rumors put the overclock at only around 1%).

https://www.amd.com/en/technologies/smart-access-memory


I don't understand what that has to do with power/performance ? Did you mean to reply to someone else ?


Yes.

I would say on your topic though that Intel has a hybrid chip that might interest you. Lakefield pairs 4 modified Tremont cores with 1 Sunny Cove core.

https://www.anandtech.com/show/15877/intel-hybrid-cpu-lakefi...


I think that we will see how this pans out in practice - but power efficiency is not something I associate with Intel - they just seem to far behind in manufacturing process to be competitive.


I find it interesting, but I'm not really interested in that market segment. Atom competes with the A53 on most things. I want a little more power.

I'd like to see 2 zen cores paired with 4 updated jaguar cores.


We don't even have new Zen 3 mobile chips yet if they are anything like the New Desktop parts we just got last week then everything Apple just put out is FUBAR.


I'm hoping Zen 3 delivers because 4000 mobile chips shows great promise but is not available in anything premium (MacBook pro level) and has supply issues.

That being said Apple has an advantage in that they control the entire stack and on the power efficiency side of things they have experience with mobile CPU design, they can do stuff like heterogeneous cores and can optimise for it throughout the stack, I think it would be challenging for x86 to do the same.


They can probably out perform them, but only for a few seconds/minutes. That's why they put a fan in the pro.


Which actually sounds like a good fit for a dev machine, where you only compile once in a while, but do not sustain high loads


looks at 1-hour compile-test runs on a threadripper yes, of course, great fit.


It’s a great option for office work.

Need a day off? Have some errands? Hit the compile button.


lolwut. you mean let your CI/CD on a real computer/server compile it much faster so you can keep working?


Seems silly to do that compile/test on a laptop instead of a server.


Well they're advertising XCode so


if you have a server for on demand compiles why would you buy a macbook over el cheapo linux machine? besides vanity ofc


OSX and overall better dev experience?


Where's dmesg, kvm, namespaces, cgroups, and package management built in?

What about inodes and cpu spikes even running node on osx? Let alone doing that inside a container.

Nah. OSX is long gone for "better dev experience". Linux always wins.


Not at the company I work for.


On the one hand I'm kind of jealous, but on the other it's really fun :D


> better dev experience

what constitutes a better dev experience?


From just off the top of my head

- Same text editing shortcuts that work consistently throughout all apps in the system including your editor/IDE, browser, and terminal.

- iTerm2 having so called "legacy full screen mode" that I use exclusively. (I've been searching for something similar for windows/linux for quite some time).

- The general system polish and usability. This is not directly "dev experience", but it's something you interact with while doing development throughout the day and it's just hands down a lot better than anything linux has atm.


> - iTerm2 having so called "legacy full screen mode" that I use exclusively. (I've been searching for something similar for windows/linux for quite some time).

Is that about disabling animations? That can be done globally on windows and linux desktop environments.


No, it's about window being in a non-exclusive fullscreen mode. It's fullscreen, but at the same time can be on the background underneath other windows. Looks like this:

https://imgur.com/a/YNzcA0w

Essentially, I always have terminal in the background even when I'm not working acting as my wallpaper.


Ah, that's basically a frameless maximized window. It requires some tinkering but there are tools to force that behavior on other applications; for both windows and various linux window managers.


A mac server?


Which don't exist?


Exactly.


My compiles take like 8 minutes at worst.


You'd hit thermal throttling way before 8 minutes I'd guess.

A lot of my iPhone app builds are pretty quick though, < 1 minute.


I did a lot of development on a pixelbook. Things were mostly fine for development, but compiling took 2-3x as long due to throttling.

Our unit tests took around 30 minutes on a fast system though and were simply not worth running outside of the specific tests I was writing.


So if I understand they have the exact same chip in the Air as the Pro? Will better thermal make that much of a difference? Is there really that big of a reason to get Pro over Air at this point?


Better thermal will make a big difference for anything where you need high power for a long period. Like editing video for example. Anything where you need only shorter bursts of power won't make as big of a difference.

But better thermals will probably mean they can run the CPU in the Pro at higher base clock speed anyway, so it will probably be faster than the air all around - we'll have to wait for benchmarks to know for sure though.


The thing is that even new Ryzen cores start shining at >20W, while ARM SOCs (especially with 5nm) can do very well at these low power envelopes.


> (especially without active cooling)

It seems likely the higher end of the performance curves are only attainable on the systems with active cooling. Or at least only sustainable with active cooling. So the MacBook Air would still realize the efficiency gains, but never the peak performance on that chart.


Someone will probably Geekbench it by next week.


Apple's A series chips have been decimating the competition in phones for years now, what makes you think they can't do the same in this space?


The fact that, like in most industries, it's easy to catch up and very hard to beat once you do. It's very telling that Apple didn't compare their M1 with something like AMD Zen3.


They don't use AMD processors in their current products. It would have been a meaningless comparison to their customers.


Because Zen 3 has been out for a week


Max 16GB of RAM on these new machines is really not that great to be honest. The mini supported up to 64GB before.


Wow, that's actually a pretty big limitation. I guess it's tough to do 64 GB with their on-package unified memory.

I wonder if they're working on a version with discrete memory and GPU for the high end? They'll need it if they ever want to get Intel out of the Mac Pro.


This would seem to point toward a tiered RAM configuration that acts somewhat like Apple's old Fusion Drives: On-package RAM would be reserved for rapid and frequent read/write, while the OS would page to discrete RAM for lower priority. Discrete RAM would act as a sort of middle ground between on-package RAM and paging to the SSD.

Then again, maybe their in-house SSD controller is so blazing fast that the performance gains from this would, for most applications, be minimal.


Apple doesn't like writing to their SSDs much, to prevent wear-out.


Let's think about that a little bit. If the RAM is fast and the SSD is fast and the virtualization options are limited, then this is good enough?

Or, inspire me. Which processes really require occupying and allocating bigger blocks of RAM?

I personally don't want to purchase another machine with 16gb RAM but that's mainly because I want the option of having a powerful Windows guest or two running at the same time. But if you take out that possibility, for now, what if the paradigm has changed just a tad.


SSD latency is still several orders of magnitude higher than RAM latency. Having similar magnitudes of total throughput (bandwidth) isn't enough to make RAM and SSDs comparable and thus remove the need for more RAM. Random access latency with basically no queue depth is very important for total performance. Certainly, SSDs are far better at this than hard drives were... but, SSDs still serve a different purpose.

Intel's Optane SSDs are based on a completely different technology than other SSDs, and their low queue depth latency is significantly better than any other SSDs out there, but good luck talking Apple into using more Intel stuff just when they're trying to switch away, and even then... just having some more real RAM would be better for most of these creator/pro workloads.


I have a project that won’t compile on systems with less than 32 GiB of RAM, and I refuse to refactor the hideously overgrown C++ template magic that landed me here.


I suspect "Apple silicon" will not really be very suitable for software engineering.


Their performance claims are the very essence of vague, but Apple sure seems certain it will be great for software engineering. I'm curious. I won't be convinced until we get some real data, but signs seem to point that way. What makes you strongly suspect it won't be great?

    Build code in Xcode up to 2.8x faster.
    
    [...]

    Compile four times as much code on a single charge, thanks to the game-changing performance per watt of the M1 chip.
source: https://www.apple.com/newsroom/2020/11/introducing-the-next-...

I have a hunch it will be adequate for single-threaded tasks and the real gains will come for multithreaded compilation, since its superior thermals should enable it to run all cores at full blast for longer periods of time without throttling, relative to the intel silicon it replaces.


Not everyone's codebase is an over-bloated mess


I've been building an over-bloated mess on Apple silicon for months now; it's been quite good at it actually.


For now, most developers use MacBooks and tools like vscode already have apple silicon build.


Saying that "most developers use MacBooks" requires a very different understanding from mine of what the words "most" or "developers" mean.


You are downvoted, but you are right. In Asia (esp India, Indonesia, Phillipines & China) but also in the EU, I see 'most developers' walking around with PC (Linux or Windows=>mostly Windows of course) laptops. I would say that by a very large margin 'most developers' on earth use Windows machines.

The most vocal and visible (and rich) ones have Macbook's though, i guess that's where the idea comes from.


I'm guessing some took issue with the possibly poorly phrased "our understanding [...] of what [...] 'developers' mean [differ]". It can be read as me dismissing people that use Macs as "not real developers", where my intent was to counteract the opposite: people dismissing others that might be using Windows as "not real developers" because otherwise they would be using Macs, which is circular logic that I have heard expressed in the past. And I say that as someone who has used Macs at work for the past decade.


According to the Stackoverflow Survey 2019, 45% use Windows, 29% use MacOS, 25% use GNU+Linux


It's actually from the 2020 survey in 2019 windows was at 47,5% and in 2018 at 49,9%. Unfortunately this metric apparently wasn't tracked in 2017 so we'll never be sure if they were above 50% back then.

So currently the majority of responding dev use a POSIX (mostly)-compliant operating system. That is actual food for thought.

Sources: https://insights.stackoverflow.com/survey/2018#technology-_-... https://insights.stackoverflow.com/survey/2019#technology-_-... https://insights.stackoverflow.com/survey/2020#technology-de...


I suspect your suspicion is going to be very wrong.


    Which processes really require occupying and 
    allocating bigger blocks of RAM?
It's not uncommon to work with e.g. 50GB+ databases these days.

They don't always need to be in RAM, particularly with modern SSD performance, but if you're using a laptop for your day job and you work with largeish data...


Virtualization and containers. Especially if you want to run an Electron based code editor next to it.


Containers on Mac rely on virtualization, don't they still? Will the new CPU arch have a native virtualization SW? Because if not, I suspect that the virtualization layer might break with the translations to and from X86, and/or might take pretty significant performance penalty.

A wild unsubstantiated guess of course, at this point (or rather, a worry of mine).


Containers on Mac still rely on virtualization, but Apple said at WWDC (and showed a demo) of Docker for Mac running on Apple Silicon, and of Parallels running a full Debian VM. Both were running ARM builds of Linux to run properly, and Apple added virtualization extensions to the new SoC so it doesn't have to resort to emulation (if the software running is ARM and not x86).

https://developer.apple.com/documentation/hypervisor


They must be... there's no chance they're wiping out their Intel lineup with machines that max at 16GB of RAM. Especially not for the Mac Pro.


I suspect/hope they are for the 16” MacBook Pro, which is still Intel-based.


They only launched their lower performance machines today. Air, mini, two port Pro.

So that’s the context to interpret the Ram they offer.


The context is the Mac Mini previously supported 64GB of RAM. In fact it was a pretty powerful machine, with options for a six core i7, 10Gb Ethernet, and as I said 64GB RAM. Now it's neutered with GigE and 16GB RAM, despite having much better CPU and GPU performance.


You can still buy an Intel iMac Mini with 64GB RAM and 10GB ethernet. If you don’t like the M1 mini, buy the Intel one.


But then you get worse CPU and GPU performance. What a choice.


SSD performance with their integrated IO controller might close the gap here, the same way that pretty fast storage on their iPhones makes the lack of RAM there not so debilitating.

But yeah, agreed that not having a 32GB option is somewhat disappointing.

(sent from my MacBook Pro with 32GB of RAM)


This is an interesting line of inquiry. In the gaming console world, a similar principle has been explicitly espoused by Sony in the decision to only go with 16GB of RAM (a smaller factor increase than any previous generation) for the Playstation 5, as the speed of its custom SSD architecture should in theory make that 16GB go a lot further than it would have with a more traditional design.


PS5 shares memory with the GPU, doesn't it? Which is different to just using RAM to load games.

I play modern games with RX 5700XT 8GB video card and 8GB RAM machine without issues.... So maybe there's not much need for more memory...


And then PS5 load times are worse than Series X. Whoops.


That's when playing backwards-compatible games (PS4 games on PS5, and Xbox / 360 / One games on Xbox Series X) as Microsoft spent more time optimizing their console to play old games. PS5 loads next-gen titles significantly faster than the Xbox Series X due to the 2.5x faster SSD.


"significantly faster than the Xbox Series X"

You have every right to support your Sony stocks, but blatant lies though ?


That's for reasons that I would expect the technical readership of this site to understand.

Please take this shallow retort back to Kotaku.


mmmmm yeah, I will let you all be the guinea pigs here and check back next month.

I am very curious about the performance.

If the macbook air is editing 6k videos and scrubbing smoothly, I am optimistic.


Heh, I have no intention of being in the initial round of beta testing! I explicitly upgraded to an ice lake machine this summer in order to get the best Intel Mac while waiting for the dust to settle.

But I do believe the dust will settle, and I'm optimistic about the future of Apple Silicon.


While a valid workload for a segment, I would be a bit skeptical of generalizing the extraordinary video editing capability to general purpose computing, as those things are facilitated heavily by hardware decoders.


Agreed, but they did mention what sounded like impressive code compilation times in the presentation too, which is pretty salient for most of us here.

Of course, we'll need to wait for third party verification in the end.


Totally--was not diminishing <strike>A14Z</strike>M1 at all--just to moderate expectations from the quite astonishing video editing capability to "measly" ~2x. Even two year old iPad Pro A12X kicks ass on Clang benchmarks (see Geekbench), although I am not sure how sustained the performance is in that thermal envelope.


out of curiosity, were we saying thermal envelope before the Apple presentation today


I don't want my SSD being killed by being used as a swap device, especially as it's not replaceable anymore. My 16GB MBP can write 200-300GB a day if I'm pushing it hard. It doesn't bother me too much because I can just replace it with a bog standard NVMe drive. But if I'm going to buy a new Mac, I want lots of RAM.


I’ll believe that when I see it. For like ten years now marketing gremlins have been breathlessly saying “these SSDs will be so fast they can replace RAM!” and it’s never even close.


For reference, I have a maxed 2018 mini I've upgraded twice -> 32GB then to 64GB.

Amazon price for 2x32gb modules is lower than it ever has been as of today ($218.58) [1] and I have had no problem making full use of that memory in MacOS. [2]

[1] https://camelcamelcamel.com/product/B07ZLCC6J8

[2] https://i.imgur.com/BDWrGw3.png


A bit more: I also run the blackmagic rx580 egpu to the XDR pro display. The system outperforms the current entry Mac Pro in geekbench 5.

So I def have eyes on the new mini. My suspicion is the rx580 will still provide 2x or more graphical power than any of these machines.


And it costs an extra €224 to go from 8 to 16gb. This is ridiculously expensive.


Any option in Euros is about 30% more expensive than buying the same in the US or Hong Kong (which is pretty crummy, but possibly related to taxes and the cost of doing business in the EU).

EDIT: I don't mean VAT/sales tax, I've considered sales taxes in the comparison, but also exchange rates of $1.18/1€. The difference is almost exactly 30%. It looks like the cost of doing business in the EU is much higher, and/or Apple chooses to price their products however they want by region.


Most of the US priced don't include tax... Eu prices include vat...


I did include CA sales tax (in Los Angeles) of 9.5%, vs the VAT-inclusive price in France/Germany/Italy, and the difference _after_ including those was about 30% higher in the EU. Germany was the lowest, probably because they temp. reduced their sales tax from 19% to 16%.

I can actually get a round-trip economy flight (pre-COVID and now) to LA just to buy a Mac mini, and save about $400. It's really.. unfortunate.


It's not just sales tax, there's also import taxes no?


That's true, I'm comparing local purchase prices with currency conversion rates. The rest depends on your personal situation and tax jurisdiction.


No what I mean is that Apple pays import duties to bring the computer into the EU from its point of manufacture. This further increases the cost compared with the US or a duty-free zone like Hong Kong. 30% overall increase of price is what I would expect for import duties + VAT, so Apple isn’t overcharging here.


Oh ok, gotcha. I’m not trying to blame any entity here, just pointing out the price difference that’s larger than the sales tax difference between the EU and CA. Interestingly HK prices for iPhones are almost exactly California prices (within a couple of dollars) even though there is no sales tax in HK - that’s probably on Apple.


No import duty either. HK is considered an international transshipment zone with no tarrifs.


Pop down to Switzerland. More than USA, less than Eurozone, and 7.7% VAT.


Not for computers into Germany.


Maxed out Mac Mini is $1699 pre tax in US and €1917 in Ireland. That's €1558 pre VAT price, or $1838 including mandatory 2 year warranty.

$139 is NOT 30% difference.


1917€ would be about $2262, and $1699 pre-tax would be about $1699*1.095=$1860 with tax, or about $402 more expensive.

I'm comparing retail pricing in France (with a 2-year EU warranty) to business pricing in the US (with a 1-year warranty). That comes out to 1949€ including VAT in France, or about $2299 USD, vs $1749 USD including Los Angeles/CA sales tax with business pricing.

About 31% more.


IMHO it’s misleading to compare the gross prices due to difference tax rates. JAlexoid’s comparison is a much more honest representation of how much Apple marks up prices in Europe. I wouldn’t expect Apple - or anyone - to eat the 10% sales tax/VAT difference.


Right, but I’m not trying to determine Apple’s markups. I just want to know where to buy a new Mac for myself. I don’t care who gets the markup, Apple, the gov’t, whatever.

Since you use words like “misleading” and “honest”, I wonder what your agenda is. Maybe you’re an Apple shareholder who’s allergic to any perceived criticism.


Then your argument is with the typical EU VAT tax model, not Apple.

It's not Apple's fault, that our governments decided to tax the crap out of everything.


Don't forget VAT. I think US prices don't include taxes as it's state dependent.


It's not quite as insulting as their $800 2TB SSD.