Modern laptops promise a kind of magic. Shut the lid or press the sleep button, toss it in a backpack, and hours, days, or weeks later, it should wake up as if nothing happened with little to no battery drain. This sounds like a fairly trivial operation — y’know, you’re literally just asking for the computer to do nothing — but in that quiet moment when the fans whir down, the screen turns dark, and your reflection stares back at you, your computer and all its little components are actually hard at work doing their bedtime routine.
↫ Aymeric Wibo at the FreeBSD Foundation
A look at how suspend and resume works in practice, from the perspective of FreeBSD. Considering FreeBSD’s laptop focus in recent times, not an unimportant subject.
It’s shocking how inconsistently hibernation features work across hardware. Even supported windows manufacturers often seem to struggle at it. My HP laptop provided by work WILL loose charge overnight if I physically close the laptop without telling it to shutdown first. I don’t remember it originally doing this when it was new running windows 8, but it’s absolutely buggy now running windows 11. I haven’t really allocated time to find out why it does this. Who knows, reinstalling everything might fix it, but it’s just a reminder how stuff is just too bloody complicated to just work.
I actually feel that putting computers to sleep is engineering around a problem that modern computers have no excuse for having these days: long boot times. Sleeping computers has become a crutch, not because we couldn’t engineer instant-on computers, but because society has normalized computers performing like molasses. Alas no single entity is really able to fix it because there are so many layers of bloat and antiquated practices from BIOS manufacturers, OS vendors, hardware bootstrap standards, all of which are highly sequential and mostly involve waiting for timeouts.
Engineers could clearly do better given the opportunity and (nearly) instant boot would be a killer feature, however it would imply throwing away antiquated PC standards that have benefited billions of users. I wouldn’t take that lightly; for all the faults, strong PC standards are what enable us to take nearly any off the shelf x86 hardware and run linux on it without asking the manufacturer for permission or support. I do like new innovation and instant boot certainly be a good feature in my book, but I’m always hesitant with non-standard solutions that risk sacrificing the other privileges grandfathered into PCs – like bring your own OS, right to repair/upgrade/service, etc. I’m not sure we’re ever going to see a new platform that’s as open as an x86 PC. I don’t know if there is a corporate will to design a more modern platform that is as open as x86 PCs are….My fear is that the manufactures would see a replacement as an opportunity to lock things down, taking more control away from owners 🙁
It’s as if vertical integration offers immense benefits to user experience. Apple can achieve a long battery life and flawless suspend and sleep on their laptops because they own the stack. Even in the Intel days, they got an off-the-shelf CPU and chipset from Intel, but they had full control over everything else and could optimize their OS just for 2-3 hardware configurations per model year.
At this point, if Microsoft wants to save Windows, they should provide a maximum of 10 hardware configurations that all OEMs have to use. This will also eliminate the driver mayhem, since Windows can ship will all the drivers for those 10 configurations bundled-in (so drivers are only needed for USB peripherals).
But that would take control from OEMs, so I don’t expect Microsoft will do that, which means that OEMs will keep sourcing components and UEFIs from the lowest bidder and the chaos will continue.
kurkosdr,
I find CPUs to be the least of the problem, it’s all the peripherals and chipsets that really end up making things hairy. For better or worse many manufacturers are guilty of not following the standards we have. Even intel are guilty of not consistently implementing energy standards in their own hardware.
https://community.intel.com/t5/Ethernet-Products/On-the-state-of-802-3az-with-Intel-X550/m-p/1364016
We have standards and they work, but its all for naught when manufacturers don’t bother implementing them 🙁
Apple PCs get away with not supporting much hardware because their vision is computers that don’t offer much in the way of customization, but that’s not a very good solution for PCs in general IMHO. I’m glad PCs are more open to hardware innovation, more choice is a good thing! I think we’re so lucky that there’s lots of viable manufacturers to choose from. I do not want my choices to be dictated to me like apple does! I can appreciate that the apple model may be fine for some, but it’s quite regressive for those who value innovation, competition, and openness.
I ultimately think that “more of the same” is the most likely path forward. A company like microsoft does have considerable reach to issue technical mandates to industry partners, however in the past they used their influence to peddle windows-centric interests rather than generic ones that would improve things for a linux user like me. For example, many microsoft certified computers no longer let owners jump into the bios without first having windows release control first.
I’ve already experienced a problem with an unbootable windows OS and needing to go into the BIOS but being unable to get there because windows wasn’t bootable. Fortunately I could remove the SSD to override microsoft’s BIOS lock, But with more computers having soldered storage, there may be fewer support options.
Yes, that was my entire point, by owning everything except the Intel CPU and chipset, Apple could offer better battery life and better suspend and sleep than PCs, since they could optimize their OS for those few hardware configurations they offered.
There is very little innovation in PCs going on (outside the GPU, which Microsoft doesn’t control and will never control anyway), Microsoft should work with PC vendors more closely to standardize the PC in my opinion. Then you can plug things as USB or PCI-E peripherals if you want more innovation or specialized functionalities. But everything on the motherboard should be standardized. Not that they will, of course, which is the reason Macs will keep gaining marketshare at the expense of PCs.
Let me clarify what I meant by saying “everything on the motherboard should be standardized”. I mean not just the chipset and UEFI, but everything a modern motherboard has on it, for example WiFi, Ethernet, sound, Thunderbolt/USB controller, brightness and volume controls on laptops etc
No manufacturers writing their own crappy UEFIs and sourcing WiFi and sound chips from the bottom of the parts bin of the worst vendors (Realtek and Broadcom), just define 10 good motherboard configurations per year using components that have the least terrible drivers and make Windows support them well.
Again, not that this will happen. It will require Microsoft either going vertical integration or making OEMs sad, and that won’t happen.
kurkosdr,
That’s not exclusive to apple as some PC manufacturers do get sleep working right on their PCs too. Power consumption is obviously a well known con of x86 computers and Apple’s advantage there is largely the product of running on a more efficient ARM CPUs.
I could understand you saying most consumers aren’t taking advantage of any special hardware, however that’s not to say there is no innovation happening here. There are so many peripherals and extender boards that it’s impossible to cover them all. The PC is the undeniable leader in peripheral innovation, not only at home but in hospitals, industry, etc. Apple’s excessive control and weak commitment to long term support make them much less desirable to target.
When I was doing SDR work, the company that contracted me to do the work was actually an apple shop. Low and behind they bought standard x86 PCs to run the backend hardware because apple didn’t have any compatible hardware. Apple’s x86 product lines didn’t support the hardware running macos.
For all your praise of apple for not supporting esoteric hardware, that is a genuine weakness for innovation both in theory and in practice.
Depends on requirements of course. Apple have some very weak spots in terms of competing with discrete GPUs, serviceability, upgradablitiy, etc. Personally I only want to run linux and apple’s proprietary hardware kind of makes it a second class citizen. In the past you’ve criticized linux users for wanting to run linux on more ARM computers including apple’s, yet we seem to agree that ARM computers have some advantages over x86.
That would make things easier, but this idea of a standard PC that everybody clones kind of died in the 90s. In many ways it’s unfortunate, but microsoft saw to it that peripherals would be standardized through windows drivers, not the hardware interfaces themselves. It means a lot more drivers are needed in every operating system….yeah it can get messy, but realistically I don’t think this problem is going away.
Sure, but there is no way for customers to easily tell which PC manufacturers are those.
Those can be added as PCI-E or USB accessories. Even deeply proprietary stuff like the stereoscopic controller for the Nvidia 3D Vision glasses is just a standard USB device. Apple slowly restricting third-party drivers for USB and PCI-E accessories is another rant for another day. But the essential stuff every computer has (WiFi, Ethernet, sound, Thunderbolt/USB controller, brightness and volume controls on laptops etc) should be standardized in PCs IMO. So much grief with modern PCs (on Windows and Desktop Linux) stems from the bewildering array of different models for essential components. I mean, if Qualcomm can present a standardized open interface for their WiFi hardware to the OS and hide the complexity inside DSPs in the hardware, others can do the same. Again, that’s what it takes to save the PC, I am not saying it will happen.
kurkosdr,
I have no problem cheering for standards, though in reality not all manufacturers are on board with them. Some are guilty of going the proprietary route. Apple are guilty too.
I don’t mind agreeing on the merits of more ubiquitous standards though IMHO ARM manufacturers deserve even more criticism than x86. Your argument that x86 is going to fail over standards feels odd considering that hardware standards is actually one of the strengths of x86’s over just about everything else. Apple’s m? computers are FOSS unfriendly at least in part due to apple going the proprietary route, leaving them partially/poorly supported under linux.
If Microsoft defines a standard, they’ll have to follow it, otherwise Windows won’t work with the particular hardware. Much like USB sticks and CD/DVD drives have to follow a specific standard, because Windows uses a standardized driver to talk to those.
Just because x86 isn’t the steaming pile of turd that ARM is when it comes to standardization, it doesn’t mean it’s good. It’s not as bad as ARM (where you have to rely on the OS vendor to create a “ROM” specific to your particular device) since at least x86 comes with a standardized bootloader (in PCs), but things like WiFi, Ethernet, sound, Thunderbolt/USB controller, brightness and volume controls on laptops etc are very much not standardized, and that’s a big pain point. Not that it will be resolved, I am saying how it could be resolved: standardized interfaces and Windows supporting only those for those classes of devices.
Apple Silicon is a vertically-integrated stack, it only has to work on MacOS, still, Apple has complete control over the standardisation of Apple Silicon Macs (no reliance on buggy inscrutable proprietary drivers from the Realteks and Broadcoms that plague the PC ecosystem), and that’s a strength of Apple Silicon Macs.
Yes, I know Apple Silicon Macs are technically ARM, but in practice they are their own ecosystem, they have nothing to do with the Snapdragons and the like.
kurkosdr,
Not “IF”…microsoft HAS defined standards, but the problem for you is that microsoft’s standards revolve around windows driver interfaces.
I understand that compatible hardware would make everything easier, but Microsoft doesn’t care whether ethernet hardware uses the same programming interface as long as all manufactures provide a compatible driver.
Again, I’m not disagreeing with you, but for better or worse, this is how things have been done going back to the 90s and it seems unlikely to change.
Apple are just as bad when it comes to not following industry standards. I see it as hypocritical to fault PC manufacturers for rolling their own proprietary hardware interface and then turn around and not fault apple. Apple are deserving of the same criticism and are arguably have been even worse with moves like dropping standard m.2 drives so that consumers don’t have the opportunity to fix/upgrade using standard hardware.
To the extent that industry standards were actually important, then logically apple deserves to be on the naughty list.
Problem is those “compatible drivers” are sometimes crap and awful, and this is an important piece of the Windows PC experience that Microsoft has no control over.
Not that Microsoft cares, in fact, they perceive all those proprietary and inscrutable “compatible drivers” talking to equally proprietary and inscrutable hardware interfaces to be their moat against Desktop Linux, not realizing that the OS they are losing massive chunks of marketshare to isn’t Desktop Linux, but MacOS.
This is what you don’t get, when it comes to MacOS, Apple doesn’t have to make sure MacOS provides a decent experience in laptops and desktops made by a dozen different OEMs (built using components from a dozen different suppliers), they only have to make sure MacOS runs well on Macs, and they control the Mac experience down to the transistors now, so, when it comes to MacOS specifically, industry standards (for hardware interfaces) are NOT important, because it’s a vertically integrated platform.
In even plainer English, people buy a Mac to run MacOS on it (yes, I know, you want to buy a Mac to run Desktop Linux on it, but you don’t matter statistically or financially), so, as long as Apple makes sure MacOS and Macs play nice with each other, MacOS users are happy.
Instead, Microsoft cannot make sure Windows plays nice with all or even most Windows PCs out there, they control neither the hardware not the drivers that make the hardware work. A standardized hardware interface could fix this by providing a “reference platform” OEMs have to meet (and judged by how well they meet it, with the OS complaining when the hardware interface is not standard) instead of OEMs and their awful component suppliers hiding behind crappy buggy drivers and promising/issuing driver updates all the time that don’t fix the issues.
kurkosdr,
I can appreciate that not everyone will be satisfied with all PC manufactures, and that’s fair enough. But you CAN buy windows certified hardware, also wholly microsoft branded PCs as well. So I find the point that you can’t get this with PCs to be unfounded.
Apple seems to keep self sabotage their own market for business/gaming/compute/servers…. Opinions will vary and I’ll admit Apple used to attract a certain kind of power user. But over time apple have been turning their backs on prosumers with products engineered to restrict and screw over advanced owners who can no longer service their own machines. Even in apple forums there are apple owners complaining about about apple making hardware less serviceable.
Also, don’t underestimate how much interest there is in running linux on apple hardware.
I get the feeling you just want to complain about any manufactuer that’s not apple. I suppose that’s your prerogative, you don’t need to justify your preference for apple to me, I’m happy for you whatever you choose! But do understand that not everyone else is going to be a fan, especially when it comes to serviceability, hardware customization, expand-ability, right to repair, etc. I’m glad for my sake that I have more choice than the limited products that apple has to offer.
@Kurkosdr
It’s mostly a HW issue.
Apple intel laptops would drain battery in sleep quickly compared to Apple Silicon counterparts. And Snapdragon Windows laptops have very similar sleep/resume behavior to M-series.
It comes down to AMD/INTL just having really poor mobile limits engine/system controller compared to vendors with more mobile experience (APPL, QCOM, MTEK, etc).
The thing is, even back in the Intel Mac days, Macbooks had better battery life even compared to same-spec PCs (even when including battery Wh in the specs), precisely because Apple could optimize their OS for the few hardware configurations they offered, and also the hardware components they chose were less crap.
Meanwhile on PCs, you have a choice between Windows trying to deliver a good battery life and working suspend-sleep by relying on buggy third-party drivers written by the lowest bidder, or Desktop Linux trying to deliver a good battery life and working suspend-sleep by relying on kernel developers reverse-engineering undocumented hardware from proprietary Windows drivers, including reverse-engineering any software workarounds the manufacturer has added in the proprietary Windows drivers to work around hardware bugs.
There is a reason Intel Macbooks were the best laptop to run Desktop Linux on.
kurkosdr,
Out of curiosity do you have data supporting this claim?
I am skeptical of claims that apple had much of an advantage on intel compared to other intel manufacturers. Apple’s own PR material was always unreliable because they use non-scientific methods, cherry pick data, and cite non-reproducible claims, as apple typically does. I am more open to believing battery life claims if and only if tests are conducted by reputable 3rd parties.
Alas, although I tried searching for this historic data for comparable hardware, I didn’t find it. Let me know if you have data.
@Kurkosdr
FWIW In my own experience, for the same specs (CPU/GPU/Battery) I didn’t see any massive difference between x86 Macbooks and the equivalent Thinkpad (the laptops I have more experience in corporate land). Linux has always seem to have performed worse than either (again from personal experience).
The vertical integration for Apple pays more dividends in term of their own M-series onward as battery/performance differentiator against x86.
I think bitlife is a one-of-a-kind game because it really shows how complicated and unpredictable life is.