Ask HN: Why hasn't x86 caught up with Apple M series?

452 points by stephenheron ↗ HN
Hi,

My daily workhorse is a M1 Pro that I purchased on release date, It has been one of the best tech purchases I have made, even now it really deals with anything I throw at it. My daily work load is regularly having a Android emulator, iOS simulator and a number of Dockers containers running simultaneously and I never hear the fans, battery life has taken a bit of a hit but it is still very respectable.

I wanted a new personal laptop, and I was debating between a MacBook Air or going for a Framework 13 with Linux. I wanted to lean into learning something new so went with the Framework and I must admit I am regretting it a bit.

The M1 was released back in 2020 and I bought the Ryzen AI 340 which is one of the newest 2025 chips from AMD, so AMD has 5 years of extra development and I had expected them to get close to the M1 in terms of battery efficiency and thermals.

The Ryzen is using a TSMC N4P process compared to the older N5 process, I managed to find a TSMC press release showing the performance/efficiency gains from the newer process: “When compared to N5, N4P offers users a reported +11% performance boost or a 22% reduction in power consumption. Beyond that, N4P can offer users a 6% increase in transistor density over N5”

I am sorely disappointed, using the Framework feels like using an older Intel based Mac. If I open too many tabs in Chrome I can feel the bottom of the laptop getting hot, open a YouTube video and the fans will often spin up.

Why haven’t AMD/Intel been able to catch up? Is x86 just not able to keep up with the ARM architecture? When can we expect a x86 laptop chip to match the M1 in efficiency/thermals?!

To be fair I haven’t tried Windows on the Framework yet it might be my Linux setup being inefficient.

Cheers, Stephen

138 comments

[ 29.8 ms ] story [ 2512 ms ] thread
All Ryzen mobile chips (so far) use a homogeneous core layout. If heat/power consumption is your concern, AMD simply hasn't caught up to the Big.little architecture Intel and Apple use.

In terms of performance though, those N4P Ryzen chips have knocked it out of the park for my use-cases. It's a great architecture for desktop/datacenter applications, still.

(comment deleted)
RISC vs CISC. Why you think a mainframe is so fast?

ARM is great. Those M are the only thing I could buy used and put Linux on it.

Everything is RISC after it gets decoded. It isn’t 1990 anymore. The decoder costs maybe 1% performance.
I may be out of date or wrong, but I recall when the M1 came out there was some claims that x86 could never catch up, because there is an instruction decoding bottleneck (instructions are all variable size), which the M1 does not have, or can do in parallel. Because of that bottleneck x86 needs to use other tricks to get speed and those run hot.
They can always catch up, it may just take a while. x86's variable size instructions have performance advantages because they fit in cache better.

ARM has better /security/ though - not only does it have more modern features but eg variable length instructions also mean you can reinterpret them by jumping into the middle of one.

I think this is partially down to Framework being a very small and new company that doesn't have the resources to make the best use of every last coulomb, rather than an inherent deficiency of x86. The larger companies like Asus and Lenovo are able to build more efficient laptops (at least under Windows), while Apple (having very few product SKUs and full vertical integration) can push things even further.

notebookcheck.com does pretty comprehensive battery and power efficiency testing - not of every single device, but they usually include a pretty good sample of the popular options.

Framework is a bit behind the others in terms of cooling, apparently due to compromises needed to achieve modularity. However, a well-tuned Ryzen U in the latest ThinkPads is not that far from M chips in terms of computing power per Watt according to some benchmarks.

Most Linux distributions are not well tuned, because this is too device-specific. Spending a few minutes writing custom udev rules, with the aid of powertop, can reduce heat and power usage dramatically. Another factor is Safari, which is significantly more efficient than Firefox and Chromium. To counter that, using a barebones setup with few running services can get you quite far. I can get more than 10 hours of battery from a recent ThinkPad.

There is one positive to all of this. Finally, we can stop listening to people who keep saying that Apple Silicon is ahead of everyone else because they have access to better process. There are now chips on better processes than M1 that still deliver much worse performance per watt.
Because of a random anecdote on hackwrnews?
I considered getting a personal MBP (I have an M3 from work), but picked up a Framework 13 with the AMD 7 7840U. I have Pop!_OS on it, and while it isn't quite as impressive as the MBP, it is radically better than other Windows / Linux laptops I have used lately, battery life is quite good, ~5hr or so, not quite on par with the MBP but still good enough that I don't really have any complaints (and being able to up upgrade RAM / SSD / even mobo is worth some tradeoff to me, where my employers will just throw my MBP away in a few years).
Curious if the suspend / hibernate "just works" when you close the lid?

I feel like I've tried several times to get this working in both Linux and Windows on various laptops and have never actually found a reliable solution (often resulting in having a hot and dead laptop in my backpack).

I’m sure it’s great.

As a layman there’s no way I’m running something called “Pop!_OS” versus Mac OS.

5 hours seems a lot worse than the ~10 hours I get on my M4 Air.
> "[...] battery life is quite good, ~5hr or so [...]"

You call five hours good?! Damn... For productivity use, I'd never buy anything below shift-endurance (eight hours or more).

How much do you like the rest of the hardware? What price would seem OK for decent GUI software that runs for a long time on batter?

Am learning x86 in order to build nice software for the Framework 12 i3 13-1315U (raptor lake). Going into the optimization manuals for intel's E-cores (apparently Atom) and AMD's 5c cores. The efficiency cores on the M1 MacBook Pro are awesome. Getting debian or Ubuntu with KDE to run this on a FW12 will be mind-boggling.

I’ve been thinking a lot about getting something from Framework, as I like their ethos around relatability. However, I currently have an M1 Pro which works just fine, so I’ve been kicking the can down the road while worrying that it just won’t be up to par in terms of what I’m used to from Apple. Not just the processor, but everything. Even in the Intel Mac days, I ended up buying a Asus Zephyrus G14, which had nothing but glowing reviews from everyone. I hated it and sold it within 6 months. There is a level of polish that I haven’t seen on any x86 laptop, which makes it really hard for me to venture outside of Apple’s sandbox.
> There is a level of polish

Yeah, those glossy mirror-like displays in which you see yourself much better than the displayed content are polished really well

> "There is a level of polish that I haven’t seen on any x86 laptop, which makes it really hard for me to venture outside of Apple’s sandbox."

Hah, it's exactly the other way around for me; I can't stand Apple's hardware. But then again I never bought anything Asus... let alone gamer laptops.

I suspect the majority of people who recommend particular x86 laptops have only had x86 laptops. There’s a lot of disparity in quality between brands and models.

Apple is just off the side somewhere else.

Most manufacturers just don't give a shit. Had the exact same experience with a well-reviewed Acer laptop a while back, ended up getting rid of it a few months in because of constant annoyances, replaced with a MacBook Air that lasted for many years. A few years back, I got one of the popular Asus NUCs that came without networking drivers installed. I'm guessing those were on the CD that came with it, but not particularly helpful on a PC without a CD drive. The same SKU came with a variety of networking hardware from different manufacturers, without any indication of which combination I had, so trial and error it was. Zero chance non-techy people would get either working on their own.
I recently upgraded from an M1 mac book pro 15", which I was pretty happy with, to the M4 max pro 16". I've been extremely impressed with the new laptop. The key metric I use to judge performance is build speed for our main project. It's a thing I do a few dozen times per day. The M1 took about four minutes to run our integration tests. I should add that those tests run in parallel and make heavy use of docker. There are close to 300 integration tests and a few unit tests. Each of those hit the database, Redis, and Elasticsearch. The M4 Pro dropped that to 40 seconds. Each individual test might take a few seconds. It seems to be benefiting a lot from both the faster CPU with lots of cores and the increased amount of memory and memory bandwidth. Whatever it is, I'm seriously impressed with this machine. It costs a lot new but on a three year lease, it boils down to about 100 euros per month. Totally worth it for me. And I'm kind of kicking myself for not upgrading earlier.

Before the M1, I was stuck using an intel core i5 running arch linux. My intel mac managed to die months before the M1 came out. Let's just say that the M1 really made me appreciate how stupidly slow that intel hardware is. I was losing lots of time doing builds. The laptop would be unusable during those builds.

Life is too short for crappy hardware. From a software point of view, I could live with Linux but not with Windows. But the hardware is a show stopper currently. I need something that runs cool and yet does not compromise on performance. And all the rest (non-crappy trackpad, amazingly good screen, cool to the touch, good battery life, etc.). And manages to look good too. I'm not aware of any windows/linux laptop that does not heavily compromise on at least a few of those things. I'm pretty sure I can get a fast laptop. But it'd be hot and loud and have the unusable synaptics trackpad. And a mediocre screen. Etc. In short, I'd be missing my mac.

Apple is showing some confidence by just designing a laptop that isn't even close to being cheap. This thing was well over 4K euros. Worth every penny. There aren't a lot of intel/amd laptops in that price class. Too much penny pinching happening in that world. People think nothing of buying a really expensive car to commute to work. But they'll cut on the thing that they use the whole day when they get there. That makes no sense whatsoever in my view.

I had a 2020 Zephyrus G14 - also bought it largely because of the reviews.

First two years it was solid, but then weird stuff started happening like the integrated GPU running full throttle at all times and sleep mode meaning "high temperature and fans spinning to do exactly nothing" (that seems to be a Windows problem because my work machine does the same).

Meanwhile the manufacturer, having released a new model, lost interest, so no firmware updates to address those issues.

I currently have the Framework 16 and I'm happy with it, but I wouldn't recommend it by default.

I for one bought it because I tend to damage stuff like screens and ports and it also enables me to have unusual arrangements like a left-handed numpad - not exactly mainstream requirements.

Most probably it is not impacting on Microsoft sales?
My M1 Macbook Pro I used at work for several months until the Ubuntu Ryzen 7 7840U P14s w/32GB RAM arrived didn't seem particularly amazing.

The only real annoying thing I've found with the P14s is the Crowdstrike junk killing battery life when it pins several cores at 100% for an hour. That never happened in MacOS. These are corporate managed devices I have no say in, and the Ubuntu flavor of the corporate malware is obviously far worse implemented in terms of efficiency and impact on battery life.

I recently built myself a 7970X Threadripper and it's quite good perf/$ even for a Threadripper. If you build a gaming-oriented 16c ryzen the perf/$ is ridiculously good.

No personal experience here with Frameworks, but I'm pretty sure Jon Blow had a modern Framework laptop he was ranting a bunch about on his coding live streams. I don't have the impression that Framework should be held as the optimal performing x86 laptop vendor.

Can you share your entire Threadripper configuration, please? I am looking to build a Linux workstation in the next 12-18 months and started weighing my options.

Weirdly, the choice pool seems to have shrunken in the last few years.

M1’s efficiency/thermals performance comes from having hardware-accelerated core system libraries.

Imagine that you made an FPGA do x86 work, and then you wanted to optimize libopenssl, or libgl, or libc. Would you restrict yourself to only modifying the source code of the libraries but not the FPGA, or would you modify the processor to take advantage of new capabilities?

For made-up example, when the iPhone 27 comes out, it won’t support booting on iOS 26 or earlier, because the drivers necessary to light it up aren’t yet published; and, similarly, it can have 3% less battery weight because they optimized the display controller to DMA more efficiently through changes to its M6 processor and the XNU/Darwin 26 DisplayController dylib.

Neither Linux, Windows, nor Intel have shown any capability to plan and execute such a strategy outside of video codecs and network I/O cards. GPU hardware acceleration is tightly controlled and defended by AMD and Nvidia who want nothing to do with any shared strategy, and neither Microsoft nor Linux generally have shown any interest whatsoever in hardware-accelerating the core system to date — though one could theorize that the Xbox is exempt from that, especially given the Proton chip.

I imagine Valve will eventually do this, most likely working with AMD to get custom silicon that implements custom hardware accelerations inside the Linux kernel that are both open source for anyone to use, and utterly useless since their correct operation hinges on custom silicon. I suspect Microsoft, Nintendo, and Sony already do this with their gaming consoles, but I can’t offer any certainty on this paragraph of speculation.

x86 isn’t able to keep up because x86 isn’t updated annually across software and hardware alike. M1 is what x86 could have been if it was versioned and updated without backwards compatibility as often as Arm was. it would be like saying “Intel’s 2026 processors all ship with AVX-1024 and hardware-accelerated DMA, and the OS kernel (and apps that want the full performance gains) must be compiled for its new ABI to boot on it”. The wreckage across the x86 ecosystem would be immense, and Microsoft would boycott them outright to try and protect itself from having to work harder to keep up — just like Adobe did with Apple M1, at least until their userbase starting canceling subscriptions en masse.

That’s why there are so many Arm Linux architectures: for Arm, this is just a fact of everyday life, and that’s what gave the M1 such a leg up in x86: not having to support anything older than your release date means you can focus on the sort of boring incremental optimizations that wouldn’t be permissible in a “must run assembly code written twenty years ago” environment assumed by Lin/Win today.

No incentive. x86 users come to the table with a heatsink in one hand and a fan in the other, ready to consume some watts.
One downside of Framework is they use DDR instead of LPDDR. This means you can upgrade or replace the RAM, but it also means memory is much slower and more power hungry.

Its also probably worth putting the laptop in "efficiency" mode (15W sustained, 25W boost per Framework). The difference in performance should be fairly negligible compared to balanced mode for most tasks and it will use less energy.

(comment deleted)
In my opinion AMD is on a good way having at least comparable performance to MacBooks copying Apples architectural decisions. Unfortunately their jump on the latest AI Hype Train did not suit them well for efficiency. Ryzen 7840U was significantly more efficient than Ryzen AI 7 350 [1]

However, with AMD Strix Halo aka AMD Ryzen AI Max+ 395 (PRO) there are Notebooks like the ZBook Ultra G1a and Tablets like the Asus ROG Flow Z13, that come close to the MacBook power / performance ratio[2] due to the fact, that they used high bandwidth soldered on memory, which allows for GPUs with shared VRAM similar to Apple's strategy.

Framework did not manage to put this thing in notebook yet, but shipped a Desktop variant. They also pointed out, that there was no way to use LPCAMM2 or any other modular RAM tech with that machine, because it would have slowed it down / increased latencies to an unusable state.

So I'm pretty sure the main reason for Apple's success is the deeply integrated architecture and I'm hopeful that AMD's next generation STRIX Halo APUs might provide this with higher efficiency and hopefully Framework adapts these chips in their notebooks. Maybe they just did in the 16?! Let's wait for this announcement: https://www.youtube.com/watch?v=OZRG7Og61mw

Regarding the deeply thought through integration there is a story I often tell: Apple used to make iPods. These had support for audio playback control with their headphone remotes (e.g. EarPods), which are still available today. These had a proprietary ultra sonic chirp protocol[3] to identify Apple devices and supported volume control and complex playback control actions. You could even navigate through menus via voiceover with longpress and then using the volume buttons to navigate. Until today with their USB-C-to-AudioJack Adapters these still work on nearly every apple device published after 2013 and the wireless earbuds also support parts of this. Android has tried to copy this tiny little engineering wonder, but until today they did not manage to get it working[4]. They instead focus on their proprietary "longpress" should work in our favour and start "hey google" thing, which is ridiculously hard to intercept / override in officially published Android apps... what a shame ;)

1: https://youtu.be/51W0eq7-xrY?t=773

2: https://youtu.be/oyrAur5yYrA

3: https://tinymicros.com/wiki/Apple_iPod_Remote_Protocol

4: https://github.com/androidx/media/issues/2637

> If I open too many tabs in Chrome I can feel the bottom of the laptop getting hot, open a YouTube video and the fans will often spin up.

Is that your metric of performance? If so...

  $ sudo cpufreq-set -u 50MHz
done!
They're big, expensive chips with a focus on power efficiency. AMD and Intel's chips that are on the big and expensive side tend toward being optimized for higher power ranges, so they don't compete well on efficiency, while their more power efficient chips tend toward being optimized for size/cost.

If you're willing to spend a bunch of die area (which directly translates into cost) you can get good numbers on the other two legs of the Power-Performance-Area triangle. The issue is that the market position of Apple's competitors is such that it doesn't make as much sense for them to make such big and expensive chips (particularly CPU cores) in a mobile-friendly power envelope.

I tend to think its putting the memory on the package. Putting the memory on the package has given the M1 over 400GB/s which is a good 4x that on a usual dual channel x64 CPU and the latency is half that of going out to a DRAM slot. That is drastic and I remember when the northbrige was first folded into the CPU by AMD with the Athlon and it had a similarly big improvements in performance. It also reduces power consumption a lot.

The cost is flexibility and I think for now they don't want to move to fixed RAM configurations. The X3D approach from AMD gets a good bunch of the benefits by just putting lots of cache on board.

Apple got a lot of performance out of not a lot of watts.

One other possibility on power saving is the way Apple ramps the clockspeed. Its quite slow to increase from its 1Ghz idle to 3.2Ghz, about 100ms and it doesn't even start for 40ms. With tiny little bursts of activity like web browsing and such this slow transition likely saves a lot of power at a cost of absolute responsiveness.

The manufacturing process they use for memory is not a good choice, actually. It is a tradeoff.
x86 has long been the industry standard and can’t be remove, but Apple could move away from it because they control both hardware and software.
Apple tailors their software to run optimally on their hardware. Other OSs have to work on a variety of platforms. Therefore limiting the amount of hardware specific optimizations.
I think the Ryzen ai max 395+ gets really close in terms of performance per watt.
I don't think there is a single thing you can point to. But overall Apple's hardware/software is highly optimized, closely knit, and each component is in general the best the industry has to offer. It is sold cheap as they make money on volume and an optimized supply chain.

Framework does not have the volume, it is optimized for modularity, and the software is not as optimized for the hardware.

As a general purpose computer Apple is impossible to beat and it will take a paradigm shift for that for to change (completely new platform - similar to the introduction of the smart phone). Framework has its place as a specialized device for people who enjoy flexible hardware and custom operating systems.

A lot of insightful comments already, but there are two other tricks I think Apple is using: (1) the laptops can get really hot before the fans turn on audibly and (2) the fans are engineered to be super quiet. So even if they run on low RPM, you won't hear them. This makes the M-series seem even more efficient than they are.

Also, especially the MacBook Pros have really large batteries, on average larger than the competition. This increases the battery runtime.