|
Patch reliability is unclear, but widespread attacks make patching prudent. Go ahead and patch, but watch out for potential problems. |
Do you put your computer to sleep? Do you enable hibernation? I’m not convinced that it’s bug free in either Windows 11 or Windows 10. Now while at t
[See the full post at: Do you put your computer to sleep?]
Susan Bradley Patch Lady/Prudent patcher
Yes, I set to turn off screen and sleep after 20 minutes (10 minutes if on battery). Hibernation enabled but rarely used. Never had any trouble with sleep or fast startup, but I do restart at least every 3 days. In an office setting, I think everyone should lock with Win+L when they walk away.
The default fast startup can become an issue in an office if computers are rarely restarted, because minor problems can linger. I persuaded staff to occasionally use Shift+Shutdown at the end of their day, instead of just Shutdown; so that fast startup is not used on the next startup, which is as good as a restart.
Talked with my IT department last week. Complained that sleep didn’t work on my ultra portable Dell laptop. They explained to me that they disabled sleep to ensure the computer would be awake for updates. I told them that the $3,000 computer became excessively HOT in my briefcase when moving from place to place. They replied that I needed to shut the computer off every time I moved it. So…Microsoft wants it on and tied to the internet 24/7/365 and my IT department wants me to turn it off every time it moves. Go figure.
I always shut my computers down at night. One never knows who or what is creeping around the Internet at night.
This argument has been going on for a very long time, even when there wasn’t an Internet as we know it now.
Mark
Always shut down at night, always have done. I update when I’m around to monitor what’s going on
I have hibernation and sleep completely disable on my computers. If I’m using them, they’re on. If I’m not using them, they’re off – including the UPS or power stips they’re plugged into.
Three reasons:
First, I’m retired so have a near zero need to lock or sleep my computers if I get up to make a cup of coffee or answer the phone.
Second, I do not want my computers to be connected 24/7 via Intel’s Management Engine – which is three rings below the kernel, on a seperate microprocessor, runs as long as the computer has battery or wall power connected, has direct access to the ethernet controller, has shown security vulnerabilities, etc. (https://en.wikipedia.org/wiki/Intel_Management_Engine) I prefer not to have MS waking up my computer at night and doing updates or whatever. “I’d rather do it myself.”
Third, I prefer to disable as many “energy vampires” in my home as possible. Keeps money in my pocket and less carbon in the air. https://www.electricrate.com/data-center/what-is-phantom-energy/
Win10 Pro x64 22H2, Win10 Home 22H2, Linux Mint + a cat with 'tortitude'.
Over 23 years and MS still has not fixed sleep/hibernation, this shows how clueless MS is.
I have start to disable sleep/hibernation since around 2000 or so. It has always been a problem. MS never care to fix it. It was an issue when had 3000-4000 PC not working correctly and getting calls from non-techy savvy people that computer not working. It became a normal part to disable it and have 1 less thing to deal with it. At work, have a script to turn off all PC after none using it for 1 hour. This cause people that work at home and did not work to call IT and say my computer turn off. Well genius you were not working since it was. Boss forced us to increase it to 2 hours but this shows how many people do not work when they say they work.
My home computer is on when I am using it and off when I am not using it.
I have two desktops at home, and have never used sleep or hibernate because I’ve always read that they don’t work reliably. Also, I turn both computers and monitors off at night or if I’m going out for more than say half an hour. I was brought up not to leave unnecessary appliances on when they are not being attended because of the fire risk, and that continues to represent the UK Fire Service’s advice today. I’m certainly glad that I was around when my computer’s liquid cooler blew a few years ago.
I used to use screensavers but they caused problems sometimes and I haven’t used them for years. Similarly I used to turn the monitors off when they were the kind that resulted in static images being burnt onto the screen, but these days I leave them on all the time I have the computers on.
The one issue I have coming up is the forced switchover from a landline phone plugged into the wall socket with Virgin Media’s form of VOIP which will mean plugging the phone into the modem which will then have to be left on 24/7 whereas I have always switched it off with the computers. This switchover is due to take place throughout the UK by the end of next year (meaning the end of local telephone exchanges) and I am gradually getting used to leaving the modem on if I go out during the day in an attempt to wean me off my old habits – but it still goes off at night!
I’d been using my main computer for gaming and had just switched it off but left it on at the mains. I was sat at my second computer alongside it when there was a loud bang and sticky brown goo emerged from under the main computer and spread across the desk towards me. It wrecked the motherboard, memory chips and processor. The machine was 3 weeks inside the warranty so I was lucky in that respect, although I think it also contributed to the failure of the graphics card just a few months later.
There had been a recall of machines with a specific batch of the coolers used at that time but when I had asked if they wanted my machine back I was assured it wasn’t in the batch – I suspect that was wrong! The recall was due to a manufacturing fault in that when the pipe was made some filings weren’t properly cleaned out and with the machine in use they moved around the cooler until they got caught up together and blocked it.
In fairness, this was some 12 years ago, and while it has put me off liquid coolers for life I have been assured here (by our esteemed Editor I think) that they have improved a lot since then!
Thanks for details. “12 years ago” is a long time in tech years, especially for mainstream liquid cooling. I’d also agree that liquid cooling has improved a lot. I’ve six trouble free years with liquid cooling on my desktop and would get it again. Very little fan noise except when the CPU is under momentary high usage. However, if I had your experience, I’d likely feel differently. Recency bias is hard to shake.
Custom desktop Asus TUF X299 Mark 1 16GB RAM i7-7820X
4 27" 1080p screens 2 over 2.
Laptop Clevo/Sager i7-9750H - 17.3" Full HD 1080p 144Hz, 16GB RAM Win 10 Pro 22H2 all
I’ve had a few HP desktop PCs over the last couple decades, but the most recent one, rather than using the legacy S3 standby (BIOS controlled), uses the “Modern Standby”/”Instant On”/S0 low power state for “Sleep” controlled by the OS in conjunction with a System-on-Chip (SOC). Apparently with pressure from MS to comply, the hardware vendors no longer provide BIOS support for the legacy (S3) Sleep state, so standby has to be managed through the OS/SOC (S0).
I loved the old legacy standby – I never had to restart my HP machines except for monthly updates or an occasional new hardware driver. Because it takes me 10 minutes or so to setup my preferred environment when I start my machine from boot, it was *so* much better to instead return from standby in a few seconds ready to work, or put it back to Sleep in seconds knowing I could quickly pick up where I left off. Reliably, every time – I NEVER had any problems returning from standby. Even if you didn’t have reliable electric or at least a UPS, you could configure it for Hybrid standby – which would normally return from standby quickly, but if there was a power outage the machine would recover using the contingency hibernation file.
One related thing I’d do is make sure Scheduled Tasks weren’t configured to “Wake the computer to run this task” – no waketimers, pet peave, so they’d stay sleeping until *I* woke them.
Now with “Modern Standby” (S0) there is no Hybrid method. And every time the monitor is soft-powered off, the system goes into S0 by design. (With legacy S3 I’d often soft-power-off the monitor while keeping the system running tasks or if I wouldn’t be using it for a while. Now I have to use a jerry-rig to keep the system running when the monitor is off: either hard-power off the monitor, or run the UAP Media Player with a song on loop & the app’s sound muted – it keeps the system from going low-power.) While in S0 deep sleep (DRIPS), the System-on-CHIP (SOC) frequently polls to see whether the system should run various tasks. This is on a FAST system that could run these tasks in a split second whether it’s idle or not, in the background without breaking a sweat. What I DON’T want is the system coming out of low power state to run these tasks all day & night. I have a hybrid SSD/HDD storage; wear & tear on the HDD from all the power-cycling is really racking up in the SMART stats. Alternatively to S0, Hibernation would put wear on the SSD & be slow. Worse, new-to-S0 users don’t initially understand why their PC doesn’t go into low power standby on request (indicated by the power-light flashing or color changed, etc.) It’s often because it’s running maintenance & myriad other tasks the moment it goes “idle” before it will eventually reach low power “DRIPS”, maybe.
By configuring specific Scheduled Tasks to not run unless there’s an internet connection, combined with disconnecting my Ethernet cable before *every* standby, I’ve been able to tame it to behave fairly similarly to legacy S3. (There’s a “Network Disconnected” setting, but it’s ignored for most purposes.) And the system by design wants to check for updates, emails, telemetry, VoIP, “news” & other distractions, etc., *frequently*. So my work-arounds usually only work until the next Windows Update or driver update causes the power cycling to go back into over-drive or worse – fail completely in reaching low power standby. Even when working well, it can vary anywhere from 10s to 5+ minutes to not-at-all reaching low power. (You can sometimes determine what blocked it by running Microsoft’s Sleep Study reporting utility & trying to decipher the cryptic codes that can’t be found anywhere online.) The only thing it attains reliably is turning the monitor off, so at least the “perception” is that the system is sleeping.
I often end up using my slow old equipment because S3 standby on them is so much faster & more reliable. Modern Standby, for me, severely dampens my enthusiasm for an otherwise great HP desktop system. (I won’t even get into laptop systems overheating or draining the battery from Modern Standby.) I understand why S0 would be great for *some* users, but it’s not ready for prime time and the old BIOS alternative has been removed…
I don’t think Modern Standby would suit me either, but I found this recent post encouraging. It covers both disabling Modern Standby and enabling the legacy S3 sleep state. If you decide to try it, please post about the outcome.
A year ago, I was shocked when I found out what had been substituted for my good old S3 standby on my shiny new machine, and that method was the first I soon tried to “revert” to S3 from S0. Afterward, the “powercfg /a” command even indicated that I had successfully activated S3. Unfortunately, S3 standby requires support from the BIOS/UEFI, and most of the manufacturers are no longer providing that code in their latest systems. When I attempted to place the system into S3 standby:
The system immediately went into 0-watt “sleep” (rather than typically taking a few seconds.) But the power LED was still lit full time instead of its typical blinking for standby. When I attempted to return from standby, power resumed to *some* components, but not to others, and video didn’t resume. I was forced to hard-power off to restart. So back to Modern Standby. I’ve done so many “Sleep Studies” and even “Windows Performance Analyzer” event tracing, trying to resolve the issues, but the only consistencies are inconsistencies. I hate it for many reasons and am currently resigned to just powering down most of the time. Giant step backward for me, as the legacy S3 had worked so perfectly for my prior systems – I hope S0 works better by the time I acquire my next system.
By the way, some hackers developed a BIOS/EUFI supplement (required disabling Secure Boot, etc.) to add back S3 support, but it doesn’t seem advisable – for security purposes and possibility of bricking the system.
I don’t agree that “modern standby” is not ready for prime time as far as a hardware feature. It’s how Windows uses it that needs work.
I use Linux, which refers to “modern standby” as S2Idle (short for Suspend to Idle). Intel calls it S0ix. I usually refer to it as S2Idle, as that is the term I see the most in my OS.
In Linux, S2idle is quite decent. When I close my Dell XPS 13’s lid, it goes into S2idle instantly, and remains there until I open the lid. No ifs, ands, or buts about it. I’ve left it in S2idle for more than a day on battery, and it is still ready to go in a moment’s notice when I open the lid. All told, it would take about two days to run a full charge down to zero.
If Windows feels like downloading updates, it will attempt to do so, and there have been scores of reports of the laptop sitting in the bag waking, attempting to perform updates, then cooking itself and running down the battery. Laptop bags are padded, so they’re excellent insulators, keeping all that heat in, and there’s no ventilation no matter how fast the fans spin.
When you get to where you are going, you find your laptop very hot (hopefully not damaged… at the very least it takes a toll on the battery to get that hot) and very not ready for usage like you thought it would be, with a drained battery.
The whole point of Modern Standby, from Microsoft’s perspective, is that it can things it cannot in S3. What good is your PC to Microsoft when it is just sitting there unused?
Modern Intel CPUs and SoCs still do support S3 for now, but the OEMs often disable it, as Dell has in the XPS 13 (9310). My Acer and Xenia both still have S3 available, and I used to select it on both on principle, but both use S2idle now. You can switch this in Windows, but it is not as simple as simply switching the option from A to B as in Linux. According to what I have read, that is supposed to require a reinstallation of Windows, but I think there are workarounds to avoid that.
Dell XPS 13/9310, i5-1135G7/16GB, KDE Neon
XPG Xenia 15, i7-9750H/16GB & GTX1660ti, KDE Neon
Acer Swift Go 14, i5-1335U/16GB, KDE Neon (and Win 11 for maintenance)
The “traditional” sleep state is known as S3, or “suspend to RAM.” When you initiate this sleep state, it copies the system state (that is not already stored in RAM) into some of the free RAM, then turns off power for the whole system (more or less), except for the RAM. The RAM on a PC is dynamic RAM, so it has to continuously be refreshed or it will “forget” its data. That takes some power, and other bits of the motherboard need to be alive to deliver that power, but nearly everything is powered down.
The CPU is not active in S3, so it cannot be processing anything. The system has to wake up (resume) from the sleep state in order to process any data. It can take a few seconds to resume from S3, as the system state is copied back from RAM into wherever it needs to be, and the various system devices are brought back to full power.
For this reason, S3 would not be a good choice for a device like a phone that has to always be on, listening for calls, incoming text messages, incoming notifications from social media apps, and all kinds of stuff. If it can process that stuff while still in a partial low power state, then go back into a full low power state when that is finished, then it can have battery run times that are decent while also functioning in a way that people want.
That is what S0ix/S2Idle/Modern Standby does. It’s a software based sleep mode (though supported by hardware based power saving), described as “Suspend to idle.” That’s what S2Idle means… it does not mean S2 in the sense of S3, which might confuse some people. S2idle actually uses S0, the fully running state of the PC.
When suspended to idle, the system achieves a very low power state (C10, on my XPS) and puts the various system devices into a similar form of very low power operation, but without turning them off fully. They don’t need to reinitialize and have their state loaded into them after sleep, because they never went fully off. The system is still technically running unsuspended, but most of the CPU is powered off until it is needed.
Certain processes (designated by the OS) are permitted to keep running while suspended to idle. On a phone, these are the listeners for the various notifications and such that I mentioned, as well as network connectivity. A phone, of course, has to remain connected to the cellular network in order to receive an incoming call or text.
PCs with S2Idle can also remain connected to a wifi access point in a low power state, so that as soon as the PC needs to use that connection, it is instantly available. No connection is possible during S3, and even if it was, the CPU is powered down and unable to process any of it. It will have to handshake and re-associate with the access point after it comes out of S3.
According to Intel, S2idle (which they call S0ix) can save as much power as S3, while allowing the PC to wake and resume internet functionality nearly instantly. If that is truly the case, there is not much reason for S3 anymore, as it takes much longer to wake and become functional again.
The reason people still want S3, though, is that Microsoft’s implementation of S2idle is not making them happy. There are many tales of annoyance with “modern standby” across the web, with many people reporting the kinds of things that have been mentioned in this thread. When you wake the laptop to use it once again, it could wake up and have only used a small bit of its battery charge, as expected, but it could also have used a big chunk of it and roasted the PC to boot while it was in the bag.
The S3 sleep mode, if it is available on a given model, solves this, because when the PC is sleeping, it’s not doing anything else. It can’t self-wake from a process running on the CPU, because the CPU isn’t active. The odds of the thing remaining asleep when you are not using it are much better, so it won’t be likely to waste your battery charge and cook itself while in the laptop bag. (I personally would recommend using hibernate rather than sleep when bagging the laptop, but lots of people prefer sleep… and if it works right, it should be fine.)
In Linux setups like mine, the way the system handles S2Idle is much like S3. While the system can allow certain processes to run while suspended to idle, it doesn’t. When I put it to sleep, I am confident that it will enter sleep mode quickly and remain that way until I open the screen in order to use it again. It has yet to disappoint me in the nearly 3 years I have had the XPS (which only does S2idle sleep).
The one exception I know about as far as remaining in the sleep state is that it will wake when it reaches the low battery threshold, at which point it will do whatever it it set to do in the power settings… in my case, to give a warning sound, then wait one minute, and if power is not connected by that point, to initiate hibernation.
I do not know about anyone else, but I much prefer a laptop that does nothing but conserve power over a semi-sleeping laptop that is still doing wakey-wakey stuff after I told it to sleep. I do not need it listening for anything (especially if it means keeping the net connection active) while it is supposed to be sleeping and conserving my battery charge for when I next decide to use the unit. I certainly do not need it trying to install updates on its own when it is supposed to be sleeping!
Dell XPS 13/9310, i5-1135G7/16GB, KDE Neon
XPG Xenia 15, i7-9750H/16GB & GTX1660ti, KDE Neon
Acer Swift Go 14, i5-1335U/16GB, KDE Neon (and Win 11 for maintenance)
Hibernation writes the current system state to the drive (hidden file hyberfil.sys) This can entail a VERY large file, depending on RAM capacity and everything that’s currently open prior to hibernation. Writing to SSD’s over previously written storage locations many times (more likely as the drive is filled closer to capacity, and when very large files are being written frequently) can eventually wear it out. Not sure how long that would take, but it seems like the hibernation process would be prone to SSD wear over time. That’s the idea anyway.
I disabled the hibernation feature to reduce writes to the SSD. The hibernation file equals the amount of RAM memory, so 32 GB of RAM = 32 GB of writes to the SSD for each hibernation.
Hibernation = system state saved to storage device. Uses very little power.
Sleep = system state saved to memory. Uses more power than hibernation as the memory must be kept alive.
Donations from Plus members keep this site going. You can identify the people who support AskWoody by the Plus badge on their avatars.
AskWoody Plus members not only get access to all of the contents of this site -- including Susan Bradley's frequently updated Patch Watch listing -- they also receive weekly AskWoody Plus Newsletters (formerly Windows Secrets Newsletter) and AskWoody Plus Alerts, emails when there are important breaking developments.
