"Hibernate doesn't use power"... or does it?

Does hibernation use more power than a total shutdown?

The oft-given answer is no, but I can't help but wonder if that's a lie: hibernation seems to be an ACPI feature, and a completely different sleep state (S4) than, well, "off" (S5).

So, does hibernation use more power than shutdown?
If not, then why is it a different hardware state than shut down?

Edit:

I forgot to mention the last part of the question, sorry:
If so, then what makes it different from shutdown, from a hardware (not driver) perspective?
Shouldn't devices not be receiving power either way?
i.e. How does a device behave differently when it is "shut down" versus when it is hibernated?

user541686

Posted 2013-03-05T18:28:41.290

Reputation: 21 330

7Modern PCs use power even when shutdown. – Celeritas – 2013-03-05T18:31:10.223

@Celeritas: Yes, which is why I asked if "does hibernation use more power than shutdown?" – user541686 – 2013-03-05T18:35:44.433

Depends on the system, and what you mean by a "full shutdown". Full "Off" (S5) uses more power than unplugging it from the wall. Check out Intel's All about System Power States (S0-S5) for some good info.

– Ƭᴇcʜιᴇ007 – 2013-03-05T18:54:08.873

@techie007: Is that link wrong, too? It says: "Power State S4: The system is almost at an OFF state, expect for a trickle power." [...] "Note that in power state S4 the system can restart from the context data stored on the disk, but in S5 the system requires a reboot." That makes no sense to me -- it would seem to imply that if I remove my laptop's batteries (or if it runs out of power) when it is hibernated, then the OS won't be able to resume from hibernation. That's not true is, it? – user541686 – 2013-03-05T18:59:37.253

@Downvoters: Care to comment? – user541686 – 2013-03-05T19:00:28.550

On modern computers S4 and S5 both have trickle power (or else the soft power button on the front of your chassis wouldn't work to turn it on). The difference is that S4 saved RAM to disk before shutting down, and will restore it on boot. S5 just dumps RAM when it shuts down. – Ƭᴇcʜιᴇ007 – 2013-03-05T19:01:15.440

3If you yanked the cord while it was in S4, it will still resume (from disk) once you give it power again and turn it on. The fact that most computers draw SOME power in both S4 and S5 has nothing to do with a difference between S4 and S5. :) – Ƭᴇcʜιᴇ007 – 2013-03-05T19:03:02.910

@techie007: How is S4 different from S5 then? Wouldn't it do exactly the same thing in either case? Why would S4 use more power than S5? – user541686 – 2013-03-05T19:07:08.377

The difference is that S4 saved RAM to disk before shutting down, and will restore it on boot. S5 just dumps RAM when it shuts down :) – Ƭᴇcʜιᴇ007 – 2013-03-05T19:07:33.533

@techie007: Uh what? So S4 dumps the RAM and restores it on boot, whereas S5 dumps the RAM but doesn't restore it on boot?! What's the point of dumping RAM then, just for fun? – user541686 – 2013-03-05T19:08:04.210

I think your lack of understanding and amount of questions go beyond what SU can provide due to what the site is for. Perhaps jump into chat, or hit a forum where you can have multiple questions answered to help you understand. – Ƭᴇcʜιᴇ007 – 2013-03-05T19:10:19.943

S4 dump the contents of RAM to DISK. S5 just turns off the RAM (and therefore losing it's contents). – Ƭᴇcʜιᴇ007 – 2013-03-05T19:11:15.357

3@techie007: Ahh, it's so confusing when you say S5 "dumps RAM", because a "dump" means writing contents to disk (crash dumps, core dumps, etc.), but when you said "dump" you meant discard, which confused the heck out of me. Okay, I see what you're saying now, but it still doesn't make sense: S5 is a state, not an action. It can't do anything, can it? The action is a transition, and it seems like going from S4 to S1 is equivalent to going from S5 to S1 -- either way, the system restores the contents of RAM and powers all devices back on, so what makes S4 use more power than S5? – user541686 – 2013-03-05T19:12:44.533

@Mehrdad - Your conclusion would be correct. – Ramhound – 2013-03-05T19:54:42.290

5My computer will react to keyboard presses when in S4 (and will power up), it will not when in S5. So there is a behavioral difference. The OS doesn't care though; if you unplug/replug the PC, it will go to S5, but the system will restore itself just fine. – avakar – 2013-03-05T22:16:15.310

@avakar: Please post that as an answer! :) – user541686 – 2013-03-05T23:23:09.670

3How in the world is this question "not constructive"? I don't see how these votes to close are justified... – bwDraco – 2013-03-05T23:34:37.353

Answers

It depends.

On some computers, S4 and S5 will use the same amount of power. I believe this is the more common scenario.

Some computers, however, can be configured to (for example) leave the network card powered up when in S4 but turn it off when in S5. In that scenario, power consumption will be higher in S4 than in S5.

Harry Johnston

Posted 2013-03-05T18:28:41.290

Reputation: 5 054

Can you provide any links to sources with more information? – ruakh – 2013-03-07T18:58:15.977

2@ruakh: see the "Deep Sleep Control" option, page 49 in ftp://ftp.dell.com/Manuals/all-products/esuprt_desktop/esuprt_optiplex_desktop/optiplex-9010-aio_Owner%27s%20Manual_en-us.pdf – Harry Johnston – 2013-03-07T20:40:48.313

According to this article hibernate is about the same as powering off:

Hibernate powers down your monitor to about 5 watts of energy and your PC to 2.3 watts—virtually the same as turning your PC off

Celeritas

Posted 2013-03-05T18:28:41.290

Reputation: 7 487

Virtually the same? On a laptop with a battery with a capacity of 46 Wh, that means it loses 50% in 10 hours. So better not hibernate before you go to bed... – AndyO – 2018-08-28T12:06:54.487

Who's monitor gets power from their PC? IDK about you but my monitor has get's power from a wall outlet. – Supercereal – 2013-03-05T18:56:55.707

@Kyle there must be a little bit because if you unplug the monitor from the PC then it gives a message like "no signal detected" so their must be power over that cable, even when the machine is shutdown. But I like your idea of getting power directly from the PC :P – Celeritas – 2013-03-05T18:59:03.193

An easy test for you to do: take a VGA/DVI cable that isn't plugged into a computer and plug it into a monitor, it will do the same thing. The monitor has power from the wall and can sense when the metal pins are connected as it completes a circuit, in fact one of those pins is a ground. – Supercereal – 2013-03-05T19:08:27.330

2Once upon a time I had a power supply that included a switched plug for attaching your monitor, so it would power down automatically with the PC (this was from before the days of "green" monitors that would know when to shut down). Either way, in this case my monitor was being powered by the PC. Not the standard today, by any means... – techturtle – 2013-03-05T19:15:24.390

A monitor that didn't take power from the PC wouldn't comply with the current VESA EDID or HDMI standards. For example, it wouldn't be able to supply identification data when unplugged, which is required by the standard. (See, for example, section 8.4.4 of the HDMI standard, version 1.3a) – David Schwartz – 2013-03-05T20:24:26.817

When your computer hibernates, the contents of physical memory (RAM) are stored onto the hard drive and the computer is physically turned off. As a result, the system does not consume any more power than if it was shut down (but see exception below).

When you turn your computer back on, early in the boot process, the system will detect that it went into hibernation and will restore from the memory image on the hard drive. A system with legacy BIOS will likely go through POST as if it is starting up from a full shutdown. A system with UEFI will often be aware that the system had hibernated (having been notified so by the operating system) and bypass POST, instead proceeding directly to the bootloader.

Do not confuse sleep with hibernate. Sleep, or standby, mode places the computer into a low-power state to allow you to quickly resume your session. If power is lost, data may be lost as this is equivalent to a hard shutdown.

Newer versions of Windows support hybrid sleep, where the contents of physical memory are written to the hard drive as with hibernation, but the system remains in a low-power state rather than turning off altogether. This allows for quick resuming while preventing data loss should power be lost--the computer will simply load from the memory image stored on the hard drive.

On Windows, the file hiberfil.sys in the root of the system volume is used to reserve space for the contents of physical memory for hibernation.

Note, however, that if wake-on-LAN (WOL) is enabled, the system will keep the network adapter on to permit the system to resume automatically on a WOL message. Depending on BIOS configuration, the network adapter may or may not be on when the system is fully shut down, and it may be possible to configure WOL in the BIOS setup so that WOL is enabled in hibernate but not when fully shut down. Otherwise, there would be no power difference.

On some systems, and once again depending on BIOS configuration, there may indeed be a nontrivial difference between a full shutdown and hibernate. For example, a computer may keep USB ports powered in hibernation to permit resuming by pressing a key or moving the mouse. In any case, a power failure will not cause a system in hibernation to lose data.

bwDraco

Posted 2013-03-05T18:28:41.290

Reputation: 41 701

They have to be different sleep states because the transition rules are different (what happens when you hit a key or press the power button). So even if the power consumption was precisely the same, they'd still have to be different sleep states. The documentation for S4 says, "The system consumes the least power compared to all other sleep states."

David Schwartz

Posted 2013-03-05T18:28:41.290

Reputation: 58 310

1States and transitions are completely orthogonal though. What makes it an ACPI concept instead of an OS/driver concept? They could just make different transitions to the same state; what makes the states themselves different? – user541686 – 2013-03-05T18:37:44.333

@Mehrdad: A state is defined by what happens when you're in that state, how you enter that state, and how you leave that state. If two things are left differently, then they are different states. – David Schwartz – 2013-03-05T20:23:21.297

How are S4 and S5 left differently, from the hardware's perspective? – user541686 – 2013-03-05T20:25:48.873

2On many systems, the BIOS attempts to resume as quickly as possible when leaving the S4 state (to meet rapid resume requirements). In order to do this, it has to be a different state. Otherwise, the BIOS would have no way to know which behavior to provide. (The OS can just check for a valid hibernation file, so it doesn't need a different hardware state. But the BIOS can't.) Also, some firmwares clear memory when leaving the S4 state to ensure a new OS instance can't steal information left in memory from the previous instance. (I don't understand why that's needed, but it is a fact.) – David Schwartz – 2013-03-05T20:38:03.017

Ahh, +1 that explains a lot, thanks for the great info! – user541686 – 2013-03-05T20:44:53.080

According to my knowledge, Hibernate saves your work to a file on your HDD and then turns off your PC as shutdown would do. So it's basically the same as power off.

I rarely shutdown my pc (only on occasions when i actually have to reboot in order to load some new files or when windows is acting crazy)

rzr

Posted 2013-03-05T18:28:41.290

Reputation: 216

My best measurement of how much power is consumed by by Toshiba Laptop (Intel Core i5, purchased in 2011) in Hibernation mode is that it uses 1.7 Watts. Note, this includes any standby power consumption of the power pack! I.e. the 1.7W may be just what is wasted in the power pack and have nothing to do with what is going on in the computer.

I cannot tell you the accuracy of the measurement... so instead I will tell you my method of measurement. I allowed the computer and battery to achieve a hibernating state and full battery charge by leaving it plugged in for a day, then allowed it to enter hibernation mode prior to starting the test. between the wall plug and the computer power adaptor I wired in a household tariff meter (Enermet E130-BS).

It recorded that 0.040kW hours of electricity were consumed during 24 hours. In the tariff meter documentation it states that the meter itself consumes =< 0.2W. I cannot tell if the 0.040 kW hrs is inclusive or exclusive of the power consumed by the tariff meter. It has a very technical description of the accuracy of the meter but I cannot understand it.

user380547

Posted 2013-03-05T18:28:41.290

Reputation: 1

-1

It uses no power, I've numerous times put my computer into hibernate brought it to a friends house and start up as if I never unplugged it.

Mike L

Posted 2013-03-05T18:28:41.290

Reputation: 1

This doesn't really add anything to what's already been posted, and your supporting evidence appears to be anecdotal. – fixer1234 – 2016-10-19T01:50:32.117

-3

Hibernation does use power. Its not the same as Shutdown but can be related closely to Sleep mode. When a computer goes into hibernate mode, it saves the contents in its RAM to the harddrive and goes to sleep.

Hibernation uses much less power than sleep mode.

FAM

Posted 2013-03-05T18:28:41.290

Reputation: 7

"It's not the same as shutdown"... well, that's basically what I said in my original question, where I said it's S4 instead of S5. The question is, well, how? Isn't everything turned off either way? What makes it different from shutdown? And yes, it uses power, but does it use *more power* than shutdown? (Why should it?) – user541686 – 2013-03-05T18:48:56.613

Accoding to Microsoft this is incorrect - hibernate stores the contents of the ram to disk AND THEN SHUTS DOWN THE PC, using no power while its off - http://windows.microsoft.com/en-us/windows7/sleep-and-hibernation-frequently-asked-questions states "[hibernation] puts your open documents and programs on your hard disk, and then turns off your computer."

– davidgo – 2013-03-05T18:50:09.267

1@davidgo: That's what everyone says, and that's why I asked this question: it seems to be an outright lie (though not a very wrong one) as far as I can see -- just look up ACPI states S4 and S5. – user541686 – 2013-03-05T18:51:45.597

@davidgo Is correct. Sleep states (like S3) save RAM connects and go into low power mode, with enough power to keep the RAM live. Hibernation dumps to disk and shuts off, and requires no power to retain the state. – Ƭᴇcʜιᴇ007 – 2013-03-05T19:04:48.823

1My reading on S4 vs S5 states all say the same thing - that the system is powered down after dumping memory to disk. Can you point to anything which implies otherwise that I can consider ? – davidgo – 2013-03-05T19:10:35.730

-1: The computer is physically turned off in hibernation. It is no different in power consumption than shutting down the computer. – bwDraco – 2013-03-05T23:07:20.547

@DragonLord: Are you sure this is wrong? According to the other answer, some devices (such as some network cards) are not as fully turned off in hibernation as when they are shut down... – user541686 – 2013-03-06T05:23:16.997

@Mehrdad: See the edit to my answer. – bwDraco – 2013-03-06T06:13:30.563