How much power does your computer use? If you have a kill-a-watt post the numbers

[hans007]

So i just bought a kill-a-watt and I wanted to see how low power a computer I could build.

so i used the following components.

intel dg31pr board
2 x 1gb ddr2-800 wintec ram
samsung 750GB spinpoint f1
liteon dvdrw sata
380 watt earthwatts 80
and an e7200 at stock
I have a card in there with a silicon image adapter that gives it DVI i'm assuming it uses some miniscule amount of energy
There is a 120mm case fan antec tri cool @ low setting, and the stock intel heatsink as well.



anyhow i hooked the kill-a-watt to that.

Idle in windows : 47 watts

orthos @ load: 72 watts


all in all i'm pretty sure this means the e7200 is quite possibly the lowest power intel cpu out there. hell it might be lower than a celeron 420 (i might try this as i have one in my other machine).

The samsung spinpoint f1 lowerd it 1 watt from a WD5000aaks also! @ idle.




so according to xbit labs

http://www.xbitlabs.com/articl...ad-q9300_13.html#sect0

an ati 790fx , with a phenom 9600 , 8800GTX and a 150GB raptor uses 205 watts or so at idle!!!

i calculated out the power difference though. if i ran them at idle 24 / 7 for 30 days, its like... 3 gallons of gasoline worth of power.


and yes i'm really bored today.

 

[JackMDS]

Windows XP Pro on a VIA C-7 (1.5GHz) based computer used as s file server/NAS.

36 Watts at full load.

6 Watts a sleep (Network traffic would wakes it up)

 

[Fullmetal Chocobo]

E6400 on P5N32-E SLI Plus with 7 hard drives: 329 watts max load. Full specs & power usage here.

 

[0roo0roo]

well the one i checked a while back was an e2200+ecs intel 945+4hd+1dvd+2gb ddr2+pci ide+soundcard=~77watt idle

 

[Modelworks]

Just keep in mind that the killawatt is not accurate .
It should be used just to get a rough idea.
The smaller the wattage used, the worse the accuracy.
It also is not accurate with inductive loads.

It should not be used to compare usage with numbers that are close, like comparing 60 and 62 watts.

 

[betasub]

Should be fairly useful for measuring draws like Fullmetal Chocobo's rig. Or my R600 set-up, for that matter

 

[DarkRogue]

E8400 @ 3.6, 8800GT OC'd, 2x2GB DDR2-800, GA-EP35-DS3P, SB Audigy, Samsung DVDRW, 5 hard drives (1 Raptor 150GB, 1 Raptor 74GB, 3 WD 500GB's) and a 380w Seasonic S12II
~120w idle, ~200w load

Seems a bit higher than I'd like but I suppose having 5 HD's contributes to that (and I've a few more waiting to be added via external enclosures... I can't seem to toss them away.) Although, it is actually better than my old Athlon XP 2600+ system which drew ~155w at idle and 200w at load.

 

[JackMDS]

Originally posted by: Modelworks
Just keep in mind that the killawatt is not accurate .
It should be used just to get a rough idea.
The smaller the wattage used, the worse the accuracy.
It also is not accurate with inductive loads.

It should not be used to compare usage with numbers that are close, like comparing 60 and 62 watts.

This is absolutely true.

However almost all of the so call measures that we do are actually more of a cultural phenomenon rather than be Technologica sound.

Almost None of the benchmarks/measures done take into consideration Standard Deviation, Standard Estimate of error, and the band of Rejecting the Zero assumption according to pre-defined percentage.

It all comes from the dire pathetic believe that a difference in few frames of Video makes any significance impact in real game performance.

However if according to Kill-A-Watt my C-7 computer does 36Watt with SD=1.5, I would not get excited for a second if someone claims that his computer takes only 30Watt. But No matter what is much better than my IP-35 computer that does 150 with an SD=10.

 

[DarkRogue]

Man, my head hurts after reading that post.. I just remembered how horrified I was when I had to do standard deviations back in High School lol.
I agree with what you're saying though.

Edit:
Yay 400th post :Q
Possibly the most posts I've ever accrued in any one forum..

 

[JackMDS]

Originally posted by: DarkRogue
Yay 400th post :Q
Possibly the most posts I've ever accrued in any one forum..

:thumbsup:

 

[Modelworks]

Originally posted by: DarkRogue
Man, my head hurts after reading that post.. I just remembered how horrified I was when I had to do standard deviations back in High School lol.
I agree with what you're saying though.

Edit:
Yay 400th post :Q
Possibly the most posts I've ever accrued in any one forum..

Congrats !

I learned about the killawatt accuracy when a friend brought over a new toy.
A fluke meter for measuring power, cost $1800
It showed how far off the killawatt really was.

But then for the cost of the killawatt, you really can't complain.

 

[DarkRogue]

lol Thank you guys.

Anyway, by how much is the Killawatt off by? A lot or is it reasonably accurate (within a couple watts?)

 

[Modelworks]

Originally posted by: DarkRogue
lol Thank you guys.

Anyway, by how much is the Killawatt off by? A lot or is it reasonably accurate (within a couple watts?)

These are the results from the post when I compared the two.

Comparing it with a fluke 1735 I found a few differences.
The killawatt is terrible at inductive loads, so don't use it for measuring those.
Inductive loads would be things typically with motors, compressors.
So not good for measuring a refirgertator or washing machine.
It measures them , but its results are not .2% accurate.
Refrigerator with compressor running 921 watts, fluke 841.8 watts

Its not accurate at measurements of small wattages.
Things like the power usage of a dvd player in standby mode are not accurate.
Killawatt said dvd player was using about 5 watts, fluke 2.64 watts.

The sampling rate is very low compared to meters like the fluke, so it can miss quick spikes or surges in usage . When I used it to measure the power usage of a 51" hdtv at turn on, it constantly gave different readings. Range from 410 watts to 504 watts. Fluke 448 to 453 watts.

PC, using a 750watt Ultra lsp model.
Core2, running at 3.3ghz, 7900gtx , pc running gelato (nvidia software that uses the gpu as a fpu, while also using the host cpu), maxes both.
Killawatt 316watts, fluke 329watts.

All that said its still a good device to use.
Its not as good as some of the professional tools like the fluke 1735, but then the fluke meter is over $2000 while the killawatt is less than 20.00

 

[DarkRogue]

Oh wow, thanks for that information - it's good to know.

I'm wondering if the Killawatt is "accurate enough" or not now. Certainly with the fridge it was off by nearly 100w which is huge, but on a PC it was off by about 15 watts which seems ok.. Still, I guess it's good for a very "general" estimate of how much power you're using.

 

[Zepper]

The OP's Earthwatts 80 Plus certified PSU is contributing a lot to that low power number. An old, inefficient PSU could be using at least 20% more. My old sock A system would pull about 150W including the speakers, printer and monitor all at idle - might go as high as 200 under load. I haven't tested the new system. And don't forget to multiply the outlet watts by the efficiency of your PSU (assuming you are only measuring the wall draw of the system box) to approximate what the system is actually using. The difference is wasted as heat.

.bh.