Why Is Apple’s M1 Chip So Fast?
debugger.medium.com
debugger.medium.com
What about XBox and Playstation?
I gather AMD makes equivalent integrated silicon for the leading game systems, so that might be the most interesting comparison.
I’m kind of sick of video encode/decode perf comparisons though as it is invariably just software vs hardware, and unsurprisingly hardware is vastly faster.
WebKit compile time:
"2020 13-inch MacBook Pro with Intel CPU: 46 minutes and 10 seconds
M1 MacBook Pro: 20 minutes and 43 seconds"
And even more fascinating:
"2019 Mac Pro (Intel 12-Core 3.3GHz CPU): 20 minutes and 11 seconds
M1 Mac mini: 19 minutes 32 seconds"
Some other measurements of some other compilations show the advantage of many core Intels but with more cores and much more electricity spent:
https://build2.org/blog/apple-m1-compilation.xhtml
"time sh ./build2-install-0.13.0.sh --local --yes ~/install
163s"
"As the first point of reference, let's compare it to my workstation with an 8-core Intel Xeon E-2288G (essentially i9-9900K plus ECC). It does the same build using vanilla Clang in 131s. While E-2288 is faster, M1's performance is nevertheless impressive. Especially when you consider that during the build the workstation is belching hot air and screaming like an airplane about to take off while M1 is whisper-quiet with barely warm air coming from its exhaust."
It seems that in the comparable power envelope M1 remains better:
"CPU CORES/THREADS TIME
Apple M1 4+4 163s
AMD Ryzen 4700U 8/8 222s
Intel Core 1165G 4/8 295s"
IMO this bit right here is worthy of highlighting again:
> Especially when you consider that during the build the workstation is belching hot air and screaming like an airplane about to take off while M1 is whisper-quiet with barely warm air coming from its exhaust.
I have never run a significant amount of compilation on any machine that didn't hit heat issues. So either the M1 is doing very well at managing heat, or Clang is doing incredibly poorly at exploiting the full system. In either case this makes the M1 look like something special.
I don't believe Apple processors are "game changing" as the hype warrants or that AMD and Intel or DOOMED, because the comparison with AMD and Intel CPU itself is flawed. It's only game changing like the lightening connector of Apple - it was based on existing technology that Apple was bold enough to customise and innovate for its specific needs. Their current processor is the same.
People are forgetting that this is not the first time that different CPU architectures are competing with one other! Apple themselves are historical proof of this with the many switches they have made over the last 3+ decade.
Apple has made a system-on-chip that is obviously quite different from the current Intel or AMD processors. Moreover their software is finely optimized for it. Kudos to them for doing so - they certainly have been bolder in pushing for these obvious existing innovations that no one was willing to spend some money to do so.
While Intel and AMD also offer integrated GPU with some of their CPU, they are nothing like the current spread of ARM SoCs that offer many other units integrated with the CPU - as this article itself points out, Apple's SoC contains a GPU, an image processing unit, a digital signal processing unit, Neural processing unit, video encoder / decoder, a "secure enclave" for encryption / decryption and unified memory (RAM integrated). (Note that this is not a unique innovation as claimed by this article, and some others - many ARM SoCs like these already exist in different variations. In fact, it's what made ARM popular.)
Obviously, when a system software or application uses these specific units of an SoC to process specific data, they may be faster than a processor that doesn't have these units. And Intel and AMD processors currently don't have these specific units integrated with their CPU. That is why comparing the Apple SoCs with Intel or AMDs processors and calling them "better" is bullshit to begin with as they aren't the same thing. If Intel and AMD also had these units along with RAM integrated within it, than it would be a reasonably fair comparison. (Assuming all the other hardware and software remain the same, which is quite difficult to do so with Apple computers as they make custom hardware that doesn't easily accept other parts.)
Yes, Apple system-on-chips for its desktop PC do appear to give a comparable performance to the current Intel and AMD CPUs (and even surpass it in some specific computing) through macOS. But calling it an AMD or Intel killer however is just illogical and exaggerated marketing - there is nothing that prevents either AMD or Intel to innovate even beyond what Apple has done, and integrate the RAM or build similar SoCs, if the market demands it.
I am betting that while Apple will certainly increase their profits and customer base with their own chips, they will also lose many "insignificant" customers like me who have no interest in buying an even less customisable and more closed system from Apple. Intel and AMD CPUs offer me more computing freedom than an Apple Silicon ever will.
The way it's similar is that you yourself make a clear case in the number of ways apple's new silicon is actually groundbreaking but then you really are trying to concoct it in a way to find a flaw. The only real problem that you point out here, which also exists with the iPhone, is the anticompetitive and customisable nature of the future Apple is creating. Someone taking a stand on that in principle is commendable though questionable practically. Ignoring that for a moment, it is not really up to argument except by biased "old timers" that this is truly amazing work in chip and software design. It's not like no one else wants or cares about any of these metrics, it's that no one has been able to get their sh*t together in a way that Apple has.
Which is interesting because just a few years back MacBooks became a very bad choice for anyone for a number of hardware issues (outrageous keyboard, finicky motherboards, etc). So this is a very interesting 180 indeed.
So true. The iPhone was by far not the first smartphone but it appeared with optimal timing as it was possible to create something really polished while still having enough features. Also now low energy is a trend and having a fast laptop with passive cooling is just surreal.
The MacBooks seem inferior in many ways but if you compare just the fact sheets, they are just more close to what a mobile user needs. (Battery life, generally solid components, low weight, it just works - and if not you can go to the next Apple store)
Intel definitely went a few times in the wrong direction with Itanium and also all the CPU bugs (although ARM had them as well). But I'm sure they'll come up with something new if needed, this wouldn't be the first time
Lol. Your comment reminded me of another sarcastic comment I came across on this topic, on Slashdot (https://m.slashdot.org/story/379156#comments)
> I've been seeing these articles for years. Oh no, Apple has made AirPods, who will buy anyone else's wireless earbuds now?. Oh no,Apple has created iPadOS, who will buy a PC laptop now?. Oh no, Apple made the trash can Mac with AMD "server" cards for so cheap, who will buy a PC workstation now?. Oh no, Apple made the iPad, who will buy a normal laptop now?. Oh no, Apple has made the iPhone, who will buy any other smartphone now?.
And that's the point you are missing.
Nobody denies that the Apple processor is a technological innovation. But like the iPhone, it is definitely an innovation limited to the APPLE ECOSYSTEM.
The new Apple processor will definitely make them more profit as it will attract more consumers. But it will neither kill Qualcomm or Nvidia or Intel or AMD. And whether you like it or not, Intel and AMD processors biggest advantage is the softwares they can run on it. And that's a huge advantage over Apple Processor that can only run the highly controlled and closed macOS BigSur where an application Firewall isn't even allowed to block Apple apps because Apple doesn't want you to.
Furthermore, the biggest casualty here is probably going to be Intel: graviton and M1 clearly show ARM is the future, so someone on the PC side is going to step up and create the M1 competition. It's probably not going to be as good but comparable, but it'll go circles around x86 again. This will then accelerate the transformation of more and more software to ARM compatibility and intel will probably play a similar song to Nokia or IBM and play a sad slow tune of death while it's only customers would be lazy monoliths sticking to their old server architectures.
Absolutely no. Nobody with a conscience should recommend closed systems with less and less options for us consumers. The biggest advantage that AMD and Intel have over ARM is the software support they have and that means more freedom and choice for us than any similar ARM system. I've said it before - I bought an Intel Mac only because it gave me the option to not be trapped in Apple's clutches. And I certainly won't buy an Apple computer anymore unless I know for sure that it doesn't restricting my computing freedom anymore. (But there is no indication to that as Big Sur shows, Apple has tightened its control on macOS by preventing application firewalls from blocking Apple apps - an indication of how it further moves to turn macOS to a closed system like the ios.)
And it is precisely for this reason that I don't see Intel or AMD dying out anytime soon.
However, don't underestimate Apple's ability to take incremental improvements and make great products out of them:
> No wireless. Less space than a Nomad. Lame.
MacBook Air has a desirable form factor, 2-day battery life, reasonable price, and is actually powerful enough to stream and edit video, play some games, and doesn't sound like a helicopter while doing it.
This is a major breakthrough right now. You literally can't get any other laptop that is fanless, and has this price, battery life, and performance. It used to be a "choose two" situation.
That said, I do completely agree with you that the new Apple processors with its associated products will increase Apple's profit and even its consumer base. So much that the loss of "insignificant" consumers like me, who care more about computing freedom and avoiding closed systems won't really matter to them. Apple's target consumers are no longer the tech savvy.
Actually I do wonder if they have a real issue here. Can the chip makers (AMD/Intel), the PC manufacturers (Lenovo/Dell/HP/etc) and Microsoft cooperate enough to compete.
I'm not sure Microsoft is that committed to Windows anymore (mobile first/cloud first). The PC manufacturers have been complacent for a long time, happy to suckle on the WinTel teat. Intel have been struggling to deliver (this is probably what drove Apple to switch).
Perhaps Microsoft could step up but my gut feeling is their heart isn't in it. What little moat Win32 still has disappears with a switch to ARM.
"Competing" in this sense is delivering similar user experience (battery life, performance, seamless hardware interactions) that Apple is achieving through their top-to-bottom control of the hardware and software.
It's not enough to show off Intel/AMD SoCs and call that good when the other components and software force subpar UX.
As far as performance, MacOS has been slower than both Windows and Linux on the same hardware for as long as I can remember. I see no reason why both operating systems won't achieve better utilization of these upcoming systems than Apple does with their M1.
https://www.phoronix.com/scan.php?page=article&item=macos101...
The M1 operates in a unique and valuable space right now, but the claim that it will be effectively impossible for the rest of the industry to move into that space doesn't hold water.
AMD is rumoured to be building a CPU that has both ARM and x86 cores - hope that is true as that will be truly game changing.
What's been impressive is that emulated code optimized for x86 runs faster on the M1. If more developer mindshare focuses on ARM optimizations and specifically writing software for the M1, the performance gains likely will even be more significant. That said, these chips are designed for a single computer, which hasn't been much of a focus area for Intel/AMD the past few years.
It'd be interesting to know how close the code quality is to what you'd get from generating ARM code from the compiler. The jump from compiling to the M1 may not be too big if the translation is good.
Another question is whether it's equally able to take advantage of the specialized execution units. It'd be interesting to know whether the native M1 compiler can target those units directly or if they're mostly used by Apple's libraries. I'd bet on the latter, although since there's no overhead of transferring data to the GPU, I wonder if there's any opportunity there to expand usage there.
The fact that little more than a recompile is needed to make native binaries (obviously lots of exceptions to that) is also impressive and a testament to the work done until now on iOS/iPhone/iPad ecosystem. Of course, some minor complaints that old 32-bit apps don't work, even on intel systems.
Meanwhile we're on year 8 of false starts for Windows on ARM. In contrast to apple, a 10 year old 32-bit x86 app might work, but a recent app probably won't.
It definitely isn't! The point was that we should consider both software + hardware together when praising the performance of the M1. That is why comparing the M1 with an Intel / AMD processor is like comparing Apples to Oranges. :)