Power Consumption of a Typical PC While Gaming

cliffski writes "How much does your PC really draw in terms of power when idle, when in sleep, and when playing a demanding game? I don't trust everything the manufacturers of hardware say, so I thought I'd get myself a watt measuring device and run a few tests on some of the gear I leave on all the time, and the gear I go to the trouble of turning off. The Linksys router drew 8 watts, the monitor drew a fairly noticeable 30-31, but what surprised me was how little power the base unit drew, even when playing Company of Heroes. Also, the variance of power draw for Vista seemed minimal, regardless of what you got the machine to do."

Comments

What about my A/C kicking into overdrive?

[CambodiaSam]

What about the thermal impact? I live in a hot climate, so leaving a PC on seems to have a big impact on the temperature of the room. Sure, I might use a couple hundred Watts to run the gear, but what about the electricity required for the A/C to cool the room back down?

Hey, when it's 100 deg F outside, I notice the difference.

Re:What about my A/C kicking into overdrive?

[AkaKaryuu]

I've had a _very_ notible temperature change in my old apartment when both my friend and I ran our Macbook Pros. Them's hot laptops. Makes you wonder what kind of damage they do to the little friend in your pants.

Re:What about my A/C kicking into overdrive?

[jawtheshark]

Back in the day, I had my server (an AMD Athlon64 2800+), my workstation (2x AMD Athlon MP 2400+), my wifes computer (P-IV 2.6GHz HT) and a huge Colour Laser printer (Ricoh Aficio CL2000) in a 10 square metre office. All the machines were pretty much on all the time.

We never ever needed to turn on the heating in that room. Even when it was -15C outside. In the summer you couldn't stay there for more than half an hour if you dared to close the door. The machines stayed stable though....

We now cut down seriously on the amount of machines we have in our office. (Let's say that getting rid of the MP was already an immense change...)

Re:What about my A/C kicking into overdrive?

[mr_mischief]

Any decent electronics project book will verify that any copper or aluminum wire will gain resistance with increasing temperature.

If you want a quick link, though, how about this article at Dan's Data about power supplies which actually gives some basic theory? It's a little suspect in that it's a review of a particular brand of power supply, and Dan's Data isn't as widely known as Tom's Hardware or Anandtech. What do you want from the very first Google result for the search "warmer power supplies draw more current", though? It also happens that he's right (about the issue, anyway -- I've never reviewed or purchased Topower power supplies).

Re:What about my A/C kicking into overdrive?

[mr_mischief]

I used to work for an ISP which was contacted by the natural gas utility company about canceling service. The gas company decided it wasn't doing us as the customer nor them as the seller any good to keep billing us just for the pipe, as we used about 2 units of gas in the five years at our location. With four offices, a lobby, the call center, and the NOC, we were self-sustaining for heat. Cooling, on the other hand, cost us dearly.

Re:What about my A/C kicking into overdrive?

[wolf12886]

Depending on where you live, that could work the other way as well, ie, offsetting your heating bill.

My tower doesn't produce much heat, but when I'm gaming with my 360, I swear its like having a space heater on, if you've never felt the air coming out of the exhaust fans, its something like putting your hand at the mouth of a hair dryer.

Re:What about my A/C kicking into overdrive?

[SoylentRed]

Was I the only one expected some kind of innuendo there, or do I just have a dirty mind?

Dirty mind if I told you I balanced it on only 1 "table-leg"?

Re:What about my A/C kicking into overdrive?

[bughunter]

No kidding. I bought a MBP almost two years ago, and a few weeks after I bought it, I got into the nightly habit of using it on my lap in front of the TV to browse the web and play games while my wife watched her programs.

A few weeks after that [this is in January mind you] I began waking up in the mornings with an especially acute itchy rash in my groin area, which I had never suffered from before. After a couple of weeks of being unable to cure myself of it, and the rash becoming angry and painful, I finally went to the doctor and was diagnosed with Tinea Cruris . The doctor asked me if I had begun using a sauna or spa regularly, to which I replied negatively...

He then asked me if I had recently purchased a laptop computer. And that's when the lightbulb went on.

After four weeks of twice daily showers, blow drying the affected area, and applying Lotrimin, the rash went away. And I no longer rest my laptop on my lap when in use.

Re:What about my A/C kicking into overdrive?

[Shivetya]

I moved my office downstairs. Simply put, the lower floor is always cooler in summer and the heat difference in winter isn't justification to have a system upstairs.

As such, the upstairs is left at 82F during the day while not in use, goes down to 78 starting that nearly an hour before expected time to turn in.

Re:What about my A/C kicking into overdrive?

[sm62704]

Laser printers are power hogs by the very way they work. They're space heaters! If you have a laser printer running, it most likely consumes more power than all your other hardware combined, and certainly puts out more heat.

The way a laser printer works is that the laser beam puts an electrostatic charge wherever it lands - which wouold be where you want the paper to be black.

The charge on the paper attracts the toner, which is black plastic ground into fine powder. A heater in the unit, at 1800 degrees f, melts the black plastic on to the paper.

If you care about your wallet (let alone global warming from the coal they have to burn for its electricity), you'll keep the laser turned off most of the time.

Re:What about my A/C kicking into overdrive?

[Kingrames]

You might be a geek, if your laptop gave you the itch.

Re:What about my A/C kicking into overdrive?

[OglinTatas]

you're lucky it was just crotch rot. Dell laptops have the dubious distinction of being the first such devices to cause penis burns.

http://www.theregister.co.uk/2002/11/22/man_burns_penis_with_laptop/

I _always_ use a lap "desk" (a flat board on top, some padding on the bottom) when I use my macbook pro, to prevent just such a thing. As you know, it gets pretty warm.

Re:What about my A/C kicking into overdrive?

[camperslo]

Any decent electronics project book will verify that any copper or aluminum wire will gain resistance with increasing temperature.

While what you say is true, there is no reason to believe that resistance losses are a significant portion of the total losses in our power supplies or that it that those losses increase by a significant percentage over the temperature range seen. Without proper analysis, facts can be used to jump to the wrong conclusions. (compare with dangers/effects of high power microwaves and discussion of WiFi for example)

In power supplies I've built that were similar but not identical to PC supplies, most of the losses were switching/conduction losses in the power transistor(s) and in the rectifiers. In the case of the rectifiers the conducting voltage drop actually gets SMALLER at higher temperatures.
In practice, the main concern about elevated temperatures in a PC is an increase in the failure rates of components. Some simply fail if too hot. Thermal cycling can also cause cracks in solder connections over time. That means high temperature operation elevates failure rates both when it occurs and to a lesser extent later.

Re:What about my A/C kicking into overdrive?

[Macman408]

Typical air conditioning can remove heat from your house with a 30-45% penalty; eg running a 100W appliance might cost another 35W in air conditioning. Incidentally, at least for air conditioning planning, I've seen a human listed as producing about 600 BTUs per hour, or 175 Watts. So your room might be warming up as much from you using the computer as it is from the computer itself.

Also, some people seem surprised that their computer has a 450W power supply even though it is only drawing 150 Watts. This is because a power supply needs to supply the peak power for all accessories that might ever be installed. If you buy a computer from Apple or Dell, the power supply needs to have enough capacity to handle not only what you're getting, plus power for extra hard drives, PCIe cards, USB devices, FireWire devices, and anything else that might be added later. Furthermore, the power budget that they work with during design likely takes into account the maximum power specified for every single chip on the motherboard, even though it is unlikely that any one of them could reach that limit, much less all of them at the same time. The CPU's specifications might require 80W, even if it's only for a few milliseconds, and for the worst combination of operating temperature, manufacturing variance, and CPU load. There's a large margin built in to the design to ensure that your computer's power needs won't exceed what can be supplied.

In my case, my Dual 2.7 GHz PowerMac G5 has a 600W power supply, even though the peak usage I've measured is around 250W. Another 90W or so is reserved for PCI/PCI-X slots that I don't use, plus there's capacity needed for another hard drive, and 4 more sticks of RAM. Add in 15W available for bus power on FireWire, 2.5W for USB bus power... Then there's the difference between the actual sustained peak usage and the specification's instantaneous peak usage, which increases the requirements significantly. Pretty quickly, it adds up to something pretty close to 600 Watts.

I borrowed a Watts Up meter from the local library (the local power company supplies them to area libraries). I'd suggest that those interested in learning about their power consumption check if there is a similar program in your area, or ask if the power company, library, environmental group, or other organization would be interested in starting one. Or, offer to buy one and donate it when you're done, and encourage others to do the same.

Re:What about my A/C kicking into overdrive?

[CodeBuster]

You could have negotiated to keep the gas line open and use the gas to run sterling heat pumps to cool the server room rather than using electricity from the grid to run conventional air conditioning units. It probably was or is worth looking into.

Re:What about my A/C kicking into overdrive?

[mollymoo]

So does your computer, powersupplies get less efficeint the warmer the room is. So while your useing only 200 watts, at 70 degrees, at 85 degrees, it's probably past 250.

Resistance does increase with temperature and a thermally controlled fan will spin faster and draw more current. But enough for a 25% rise in consumption from a 15 degree (in unspecified units, I guess you mean Fahrenheit) temperature rise? That's seems like a hell of a lot for a fairly modest rise in temperature.

For an 80% efficient power supply, an increase of 25% overall consumption is more than double the power loss. The reality is very complex, but we can pick out a few relevant numbers to get a feel for the magnitudes involved. Empirical testing would be easier than an analysis, but here's some food for thought:

For copper, the resistance rises by about 0.4% per degree Celsius rise. Your roughly 7 Celsius rise would increase it by a whopping 2.8%. You'll have melted the insulation well before even a 50% rise in the resistance of your copper wire.

If you look inside a power supply, you'll see a big fat heatsink. Attached to that are rectifiers and switches - diodes and FETs. That's where a big proportion of your power supply's inefficiency comes from. Looking at the first power FET datasheet I have to hand (for a Fairchild HUF75337P3), the on resistance increases by something like 1% per degree Celsius rise. For diodes because the forward voltage drop actually decreases with increasing junction temperature - they get more efficient. For an International Rectifier 12CWQ03FN it looks to be about 0.2% lower per degree Celsius rise.

The YS-Tech 80mm fans in this box next to me consume 0.84W at full speed. That's a slow fan though, I wouldn't be surprised if more typical ones used 2-3W at full speed.

Hardly a complete analysis, but just can't see where you're getting this additional 50W from. I think you're out by an order of magnitude.

Re:What about my A/C kicking into overdrive?

[billcopc]

Actually, Dell uses the term "laptop" on official communication, including the front page of their web site.

The whole "don't call it a laptop" scare is pure nonsense dreamt up by some ignorant middle-manager or escalation expert. I didn't work there long, just long enough to realize I didn't want to be that escalation expert, so I turned down the promotion and left shortly thereafter. Like any call center, the rules are mostly made-up on-the-fly, and the people making the rules are usually the least technical in the building.

Re:What about my A/C kicking into overdrive?

[billcopc]

Dan's Data was around back when Tom's Hardware was just a bunch of teenagers chipping at the Berlin wall with claw hammers.

He's got an odd personality at times, and is anything but objective, but the man tends to know his facts.

And yes, a power supply's efficiency and load capacity are adversely affected by temperature. That's partly why the better ones weigh a ton : big bad-ass heatsinks.

Re:FP

[Kuroji]

I don't suppose you would be willing to tell us WHICH ONE IT IS, now, would you?

I love kill-a-watt

[Bandman]

My wife is huge into low-energy tools, and she got us a kill-a-watt to play with.

In my server racks, I've got the PDU equivalent of this from APC. They've helped me many times in load balancing the power draw across our circuits

Re:I love kill-a-watt

[penguin_dance]

My wife is huge into low-energy tools, and she got us a kill-a-watt to play with.

Hey pervert, keep your smutty bedroom toys to yourself! :P

Re:I love kill-a-watt

[gEvil (beta)]

My wife is huge into low-energy tools

Obviously. She married you.

Re:I love kill-a-watt

[bucky0]

>>I can hardly believe that the router mentioned was using 8 watts, what is the time period there?

8 Watts is probably about 8 Joules / Second. You know, in that ballpark.

Re:I love kill-a-watt

[ivan256]

I can hardly believe that the router mentioned was using 8 watts, what is the time period there? I know the power supply on my linksys router is in the milliamps so, basic conversion would indicate to me that is not possible. I am probably thinking of something backwards though. Time is a component of the "watts" unit. One watt is one joule per second. So the time period is irrelevant.

8 watts at 120 volts (simplistically speaking) would only be about 66 milliamps.

Re:I love kill-a-watt

[compro01]

450W at idle?!

The power rating on the power supply indicates how much it is capable of supplying. it will not draw more than is needed, plus some for the thermal losses in the power supply (a good many supplies are better than 80% efficient).

My fairly powerful system (3ghz core2 duo, 2GB ram, 500GB hard drive, 8800GTS, 700W generic brand power supply, and a 17" CRT from 7 years ago,) draws about 370W running flat out according to my UPS (a 780W/1200VA APC unit) and sits at about 200-250W at idle (for various definitions of idle).

Re:I love kill-a-watt

[Amouth]

on top of that the ratings you see on the power blocks is normaly the dc output - the kill-a-watt meausres the power draw on the AC side before conversion - there is no doubt that router is running on less than 8watts as you lose in the conversion and heat. and most bricks arn't what you would call effecient devices they are cheap.

but as the power total isn't much the effeciency isn't that big of a deal

if you are only 50% effecient in conversion but that loss is only say 10 watts no big deal compared to a loss of several hundrad watts with an old AT non switching powersupply

Re:I love kill-a-watt

[Endo13]

This is why some posts should be modded to +10.

Kicking and screaming

[gearloos]

Personally my I^2 R losses are always better if I'm kicking the box and screaming after a good wow gank.

Little information

[Kelz]

All he talked about with regards to the PC was the processor and video card. What power supply was he using? A super efficient one or a super stable one? How many peripherals were hooked in? Hard drives? Fans?

Re:Little information

[cliffski]

Hi. Theres a single hard drive in there. a typical 'shipped with the box' PSU, no wireless card or anything fancy.
The PC is from mesh Computers, about a year old.

Worth the cost

[wolf12886]

Where I live we pay $.08 per kilowatt hour, so running my computer 24/7 costs me (assuming 190 watts at idle) $11 a month, which is not nothing, but is certainly worth the convenience.

Also, I doubt the leds on any of the devices mentioned account for nearly any of the measured current draw, 20ma at 5v is .1w, so you'd have to have to have a hell of alot of leds to make a noticeable difference in power usage. most of that current is probably being burned up in the wall wart, linear regulators or transmitters, if I had to guess.

Re:Worth the cost

[RoverDaddy]

Wow, what pompous, righteous indignation over 190W! You have no idea what this person's lifestyle is, whether they do lots of other things to help the environment and keep their energy costs low. My PC is on 24/7 too, for several reasons:

1. I frequently access it from work over SSH.
2. When I'm home, somebody in the family is using it nearly all the time (instead of TV which is constantly in use in plenty of other homes)
3. When nobody is using it, it's running BOINC on behalf of World Community Grid doing useful things like cancer/AIDS research.
4. My life experience with computers indicates that a computer running all the time will live longer than a computer switched on and off.

Maybe my opinion on PC life is bunk, but overall I don't think I'm killing off the ecosystem by keeping my PC on. I've replaced every incandescent bulb I can with compact fluorescents. I actually turn lights OFF when I leave a room. I use LCDs instead of CRTs. I recycle. I drive a small car, and keep the speed down and I've cut down my driving considerably.

There are much better targets for your ranting and raving than $11 worth of electricity.

BTW, I'm very jealous of the $0.08/KWH price. Where I live (Massachusetts, US) the cost of electricity is more like $0.20/KWH.

I need one of those

[legoman666]

I have 4 monitors on my machine. 3 19" LCD's and 1 22" LCD. The PC itself is a Q6600 @ 3.1ghz and 2 HD3870's, also overclocked. The CPU and 2 GPUs are watercooled. There are also 4 hard drives and a sound card.

I've think I've estimated the power draw at around 450w under full load (not including the monitors. 3 of them are turned off when I play games).

Luckily I don't pay my electric bill.

Re:I need one of those

[Holi]

Luckily I don't pay my electric bill.

A true slashdot stereotype.

Let me guess, late 20's and parents basement.

been wondering about this also

[800DeadCCs]

been wondering about this also.
mainly due to having only one 20A outlet, and the building is old enough I don't want to risk that much.
looking to build a new system, I want to make a strong but low power-draw system (gonna use a 45nm intel chip). Looking at specs on various parts suppliers sites, I come across numbers like "total thermal dissipation", or things like the notes on Intel's ATOM board: "fully populated board with accessories uses 75W max"

Where do I find out exactly how much wattage I need?
A lot of the calculator sites seem to be either a tad old, or just give info on a few select parts.

on a note about the article,
I'd rather see what the power usage is while starting up (seems that's when the biggest drain usually is).
as for the printer, OK, it's just a deskjet, but show the drain on a laser warming up (for B/W, you're better with one of those).

Accuracy

[Anonymous Coward]

I wonder how accurate those readings are. Cheap power meters often only produce correct values for purely resistive loads. Everything uses switching power supplies nowdays (they're very efficient compared to linear power supplies), and they draw current in peculiar ways (some have power factor correction which improves things). Does anyone have a Kill-a-Watt vs Oscilloscope accuracy comparison for different kinds of loads?

Re:Accuracy

[Trogre]

My cheap power meter displays power factor, but not very accurately. It will tell me an AC fan has a power factor of 33% (correct) and an early-model switch-mode PSU has a power factor of 100% (wrong).

After seeing the effects of several hundred inductive loads on an AC grid, I now only buy PSUs with Active Power Factor Correction. It costs less in the long run.

Re:Accuracy

[Quelain]

I don't think they are accurate at all on switch mode power supplies. I have one which is definitely wrong when measuring a PC PSU.

I think they expect to see peak current at the peaks of AC voltage, but a switch mode PSU will take small bites of current which may or may not coincide with the voltage peaks.

Vista increases the wattage

[mariushm]

Note that he uses Vista and he says his computer doesn't need more watts when playing games compared to normal usage.
Maybe this is because Vista's 3D interface already taxes the video card and forces it to draw a lot of power?

Provisioned power vs. used energy

[tepples]

Time is NOT a component of the "watts" unit. [...] The joule is is the SI unit of energy measuring heat, electricity and mechanical work. The watt is a unit of power, and power is the rate of energy transfer. A watt is one joule per second (1 J/s); there's a time component on the bottom of the fraction.

A watt is current times voltage; e.g. at 100 volts one ampre, you have 100 watts. Right, but voltage itself is derived from power. A volt is a watt per ampere, or a joule per ampere-second, or a joule per coulomb.

but if you check your electric bill you'll see that you aren't billed for watts used, but kilowatt hours used. I seem to remember reading that electric companies bill for both provisioning and usage: one rate for peak power (in watts) that could be used and another for energy (in kilowatt hours or megajoules) that is used. I will grant that some electric companies don't show provisioning as a separate line item on single-family residential accounts.

Re:Consoles

[CastrTroy]

Here's a comparison that shows XBox 360 vs. PS3 vs. Wii vs. PC in many different areas including standby, idle, gaming, and movies (Wii not included in movies).

Power consumption of PS3 / Wii / XBOX 360

[IYagami]

You should take a look at http://www.hardcoreware.net/reviews/review-356-1.htm

This is the main information:
Power Consumption in Games
PS3: 185.9 Watt average
XBox360: 176,54
PC (see link for more information): 156,6
Wii: 16.8

Bad Latitude

[argent]

Talk about your latitude adjustment.

Re:Bad Latitude

[SacredByte]

I wonder what gave you the Inspiron for that comment?

Re:No sources needed

[pushing-robot]

So, if you have a 200 watt power supply, making 200 the 80%, you would be drawing around 250 watts of power.

A very common fallacy is that a PSU always draws as many watts as it's rated for; in other words, a 500-watt PSU constantly draws 500 watts or more. This is incorrect; your PSU only supplies (and draws) as many watts as your computer currently needs.

"80-plus certified" means the PSU was tested to be 80% efficient at 25%, 50%, and 100% load. Assuming you have a fairly low-end system, your 200-watt PSU may never supply more than 100 watts, and therefore (being 80% efficient) never draw more than 125 watts. If you added a component to your system that consumed an extra 20 watts, your PSU would supply an additional 20 watts, and draw an additional 25 watts (again, 80% efficiency). Simple as that.

Useful, relevant links... for once:

[Ihmhi]

After you figure out your kWh usage for your respective devices, this kWh cost calculator is useful for finding out how much it costs to run it during a period of time.

I leave my computer on all the time. I highballed its power usage at 200 watts to factor in the speakers, monitor, and computer itself (the monitor is not on all the time, but the computer is nowhere near 200 watts - so I just did a rough estimate.

You would also want to find out much a kWh costs in your state to plug in the correct values. I set it up for $0.11 a kWh, also a big higher estimate.

My father insisted that my computer was the reason the power bill was over $200 one month. It's nice to be able to tell him that it only costs about $15 a month to use. (: