Standby Efficiency

Riffraff

Deity
Joined
Jan 2, 2006
Messages
2,059
Location
Munich
Ok I have a question concerning the efficiency of the standby modus vs turning the computer on and off.

Quite simply, I am interested in how much energy is required to power a computer down and then up compared with letting it run in standby mode. I really can't tell where the break even point for switching off computers is - could be anywhere between 10 mins and 5 days. I'm quite sure this is all strongly influenced by the computer aswell as the OS - personally I use a pretty new/fast PC and Vista 32.

So anyone got some numbers? I'm really not intersted in uneducated guesses, cause I've got those myself...
 
Well, at least for notebooks, I know the numbers for standby. In standby, the computer uses power only for memory*. Each stick of DDR2 notebook memory uses 1.8 Watts of power, and since most notebooks use 2 sticks of RAM, that means it uses 3.6 Watts of power per hour in standby. Note that some newer notebooks - a small percentage of those that have come out since mid-July - use DDR3 RAM, which uses 1.5 Watts per stick, and some use older types of RAM. Desktop numbers are probably similar, but my guess would be slightly more power-draining. Resume time from standby is typically near-instant, so any power used there is quite little. If it takes more than a couple seconds to resume from standby, you probably aren't actually using true standby - rather "hybrid sleep" or something of the sort that saves data to disk.

How much is used to start the computer depends a lot more on the hardware and software. The more powerful the computer is in general, the more power would be used in starting it up, and the more programs that start on a boot, the longer the startup and the more power will be used. The likely exception here is the hard drive - a faster hard drive uses minimally more power, but also means a quicker start-up. If I remember, my notebook uses about 2% of its battery to start up - thus about 1.5 Watts. Going by that, if you're gone for anything more than 25 minutes, turning your computer off would be more power-efficient.

Since I don't have many programs up, I'll test this in just a minute...

What I recommend most is hibernate; it uses no power and resumes quicker than restarting. But there are times when standby or a full restart are preferable. It takes 1/3 to 1/2 the time to resume that restarting does - less if restarting takes an unusually long time on your computer - thus using less power.

*Disclaimer: There's a little bit being used up if it is plugged in, but I'm assuming you aren't going to unplug it after you turn it off, or that it's an unplugged laptop, in which case this is irrelevant.

---------

Test results: A full boot, from pushing the power button to when the computer was fully responsive (including the Start Menu and All Programs therein) and startup programs had begun, took 999 Microwatts of power (0.999 Watts). It would take 16 minutes, 39 seconds for Standby to use the same amount of power. Specifications are as follows:

Hard drive: Hitachi Travelstar, 7200 RPM, 149 GB (binary; 160 digital), 16 MB cache, 2.5-inch
OS: Windows XP SP3
Anti-Virus/Firewall: Sophos/Windows
Screen brightness: Maximum.
Peripherals: Wireless card OFF, Ethernet connected, optical Microsoft Mouse connected via USB, headphones connected, CD ROM in drive, Bluetooth off, touchpad on, all else off.
Processor: Intel Core 2 Duo T7500 (power rating: 35 Watts - uses less most of the time)
Graphics card: GeForce 8600M GT (maximum power usage: 22 Watts - uses less most of the time)
Other startup programs: AIM 6.5, Windows Defender, a couple other very small ones
RAM: 2x1 GB laptop DDR2

Oh and I should note that I unplugged my laptop after it had shut down; it'll take a bit more power (but probably less than double) if you intend to count the time it takes to shut down as well - putting you back around the 25 minute range. Also that newer may or may not mean more power-efficient - that kind of goes in cycles. And usually the faster a computer is for the time it was built, the more power it uses.

Power usage measured with RightMark CPU Clock Utility. Your results will differ of course, but this can be used at least as a general indication.
 
I just stick my laptop in hibernate when not in use. It just dumps the contents from the RAM into the HDD then goes into hibernate.
 
Wow Quintillus, thank's alot for your replay, this was exactley the kind of data I was looking for. I had expected the break even point to be signifacantly later to be honest, so it looks like I really should change my habits.

What I dont quite understand is this:
What I recommend most is hibernate; it uses no power and resumes quicker than restarting. But there are times when standby or a full restart are preferable. It takes 1/3 to 1/2 the time to resume that restarting does - less if restarting takes an unusually long time on your computer - thus using less power.

I'm not sure I have two different settings hibernate/standby in my system. I'm actually assuming now though, that Vista goes into hibernate - what I assumed it was standby. So if the hibernate mode uses no power, where is the point in powering the PC down at all? The only thing I could think of besides some system readjustment requiring a restart would be the durabiltity of the components..
 
You're confusing units in your post. Watts are already a measure of energy per time (joules per second), you can't have a watt per time period.

After being now in my fourth year of studying physics - how the hell did I miss that?

At least it seems consistent though ;)
 
Back
Top Bottom