E6600 Power consumption tests - Power-Mate meter

[eva2000]

I recently purchased an Aussie made Power-Mate (http://www.power-mate.com.au/) power meter for testing a few power supplies I have been playing with and decided to do some system tests to see how much power is drawn at the mains socket :)

The Power-Mate is a device that is inserted between the main’s power outlet and an appliance’s power plug. The device can accurately measure and calculate power, volts, amps and energy being used by an electrical appliance. The Power-Mate can also measure the greenhouse gas emission created by an appliance. The highlight of the device is its ability to determine the actual running cost of an electrical appliance.

http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/thumbnails/thumb_powermate_001.jpg (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_001.html) http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/thumbnails/thumb_powermate_002.jpg (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_002.html) http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/thumbnails/thumb_powermate_003.jpg (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_003.html)

http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/thumbnails/thumb_powermate_004.jpg (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_004.html) http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/thumbnails/thumb_powermate_007.jpg (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_007.html) http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/thumbnails/thumb_powermate_009.jpg (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_009.html)

System:

Intel Core 2 Duo E6600 QPGE B1 Step 5 ES
Scythe Infinity + 2x Thermaltake 120x25mm 81cfm push-pull (http://i4memory.com/showthread.php?t=3218)
Intel D975XBX Bad Axe rev304 1334 bios
2x1GB Gskill PC6400PHU2 HZ Micron D9GMH 0629 batch
2x 320GB Seagate 7200.10 SATAII hdds
Pioneer DVD-RW Burner
1.44MB FDD
600W OCZ EvoStream modular (http://i4memory.com/showthread.php?t=3140) power supply


I tested 2 system clocks @2400Mhz at 1.275v bios (1.259v) and @3600Mhz at 1.425v bios (1.394v) with dual stressprime Orthos and TAT cpu load tests with the only variance being the video card used.

I tested first with 128MB Gainward FX5200 PCI video, then with 2x 256MB BFG 7800GTX SLI with modded bios @473(525)/1350 which is my 24/7 clocks.


128MB Gainward FX5200 PCI

@2400Mhz at 1.275v bios (1.259v)
Windows Idle watts = 122.5W (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_012.html) / (0.563A (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_017.html))
Windows Orthos Load watts = 164.8W (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_015.html)
Windows TAT Load watts = 173.5W (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_016.html) / (0.777A (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_019.html))

@3600Mhz at 1.425v bios (1.394v)
Windows Idle watts = 146.2W (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_020.html) / (0.673A (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_024.html))
Windows Orthos Load watts = 227.4W (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_021.html)
Windows TAT Load watts = 248W (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_022.html) / (1.094A (http://i4memory.com/reviewimages/powersupply/powermate/e6600_fx5200_evostream/1/photos/powermate_023.html))

As suspected TAT's internal cpu load test (http://i4memory.com/showthread.php?t=3082) heats up the cpu more than dual stressprime Orthos load testing and it shows in the power consumption readings :)


2x256MB BFG 7800GTX SLI @473(525)/1350

Ran 3dmark06 and PCmark05 and recorded the peak power draw values listed below.

@2400Mhz at 1.275v bios (1.259v)
Windows Idle watts = 187.7W
Windows Orthos Load watts = 231.9W
Windows TAT Load watts = 240.8W

PCmark05 = 359.8W peak

3dmark06 = 344.1W peak
Return to Proxycon = 337.9W
Firefly Forest = 344.1W
CPU1 - Red Valley = 253.5W
CPU2 - Red Valley = 266.3W
1 - Canyon Flight (SM 3.0) = 334.1W
2 - Deep Freeze (SM 3.0) = 335.7W



@3600Mhz at 1.425v bios (1.394v)
Windows Idle watts = 218.9W
Windows Orthos Load watts = 303.7W
Windows TAT Load watts = 324.4W

PCmark05 = 401.3W peak

3dmark06 = 394.8W peak
Return to Proxycon = 386.5W
Firefly Forest = 394.8W
CPU1 - Red Valley = 290.6W
CPU2 - Red Valley = 289.6W
1 - Canyon Flight (SM 3.0) = 384.9W
2 - Deep Freeze (SM 3.0) = 384.9W

For the stressprime 'Orthos' and TAT load tests there wouldn't be as much a power draw difference between 7800GTX SLI vs FX5200 PCI since there isn't any video card load in those tests. You can see the difference in power draw when video card load comes into play with 7800GTX SLI configuration.

[eva2000]

Now to compare all the data in a nice graph :)

http://i4memory.com/reviewimages/powersupply/powermate/E6600_FX5200_7800GTXSLI_powerdraw.png

[Elfear]

Thanks for the testing eva. Any chance you could do the same tests on your Crossfire rig?

[eva2000]

Well my X1900 CF master card is on water cooling and the loop might not reach where i setup my rig right now, but I do plan on tests with CF.

[davexl]

Thanks for the great tests as always Eva....

one thing I am interested in is if it is worth running Rightmark to lower power consumption during idle - as explored here. (http://www.xtremesystems.org/forums/showthread.php?p=1814283#post1814283)

Were your tests using any sort of power saving at idle?

If you have the time I would love to see what actual wattage difference this would make in a barebones overclocked system...

[GWillakers]

Thanks EVA,

I mirror DaveXL's interest in the power consumption at idle.

I noticed a while back that the E6600 runs much hotter in W2K pro than XP Pro, due mainly because W2K lacks speedstep drivers.

When you measure at idle, is the E6600 running at 2.4GHZ or has it downshifted to 1.58GHz?

Right click "My computer" ... Properties.... General Tab..
3rd line under Computer:

[Lestat]

now you fully understand why i have said these cpu's when overclocked heat up to hotter than my 955 XE @ 4.5ghz

i find in just mind numbing how much wattage these little chips create/use when they aer overclocked.

you also understand wh, when i see people say that they are running 1.5~1.6v on AIR cooling, how i find it laughingly bogus.

the insane amount of heat dissipation ability(s) thats needed for that amount of wattage/heat is not even remotely possible with the current air coolers on the market unless they are sitting right under an a/c unit.


Thank you eva for all your work, i have been wanting to do this for a long time now, infact ever since the day i got my conroe.
thank you for doing what i should have done months ago.
thank you from all of us!!

[eva2000]

thanks guys glad it's useful info for you

AFAIK, yes at idle E6600 is down from 2400mhz to 1600mhz - probably explains the very low idle system power consumption figures.

[davexl]

you also understand wh, when i see people say that they are running 1.5~1.6v on AIR cooling, how i find it laughingly bogus.


Wow, strong words.

Rather than calling us liars, why not try it yourself, it is not like it is a secret.

Always a good move to know what you are talking about before calling people bogus..

My load temps (at 20C ambient) with quiet fans running 10K Orthos are 72C Core Temp - which is 10C under thermal throttling. (3750MHz, 1.5625v in BIOS) This means real world apps like 3d renders get to about 60C Core Temp - perfectly acceptable for 24/7.




the insane amount of heat dissipation ability(s) thats needed for that amount of wattage/heat is not even remotely possible with the current air coolers on the market unless they are sitting right under an a/c unit.


You haven't tried it have you? Or else you would not say something so silly.

a) Buy a modern cooler eg Ultra 120
b) Install it with good paste like AS5, and a quiet fan.
c) wipe egg from face

[thunderstruck!]

I would be careful davexl, that is a dangerously high temp. The insides of your case must be steaming. Anyway, eva, please test x1900 cf soon!!!:)

[fhpchris]

you also understand wh, when i see people say that they are running 1.5~1.6v on AIR cooling, how i find it laughingly bogus.

I could easily run 1.55v 24-7 on a stock typhoon, how is that bogus?

1.4xv is low letstat, you are not even close to unleashing the potential of that cpu.


Week 28 @ Typhoon + denki -- Ambient 72f
Idle temp after run @ 1.65v = 53c Asus probe, 45c Core temp
http://www.sspmustang.com/OT/E6600/3700-tg512-32m-run2.JPG

Temps during run
59c @ Asus probe, 54c Core Temp
http://www.sspmustang.com/OT/E6600/temp.JPG

On Water:

Water setup
------------
Swiftech MCR-320
Iwaki MD-20R
MP-05 SP LE
Maze4
dual 5.25 bay res
3*Sanyo Denki 120x38mm

Changes to comp
----------------
Trying out my Teamgroup 4gb 667 3-4-4-8
X1950XTX from X1900 Crossfire Edition on air
Ambients @ 68f (appartment complex now includes power in rent ;) )
6400 Ballistix @ 1:1 cas 5-5-5 because they suck :(
downloading CS:S off steam during 32m and prime ;)
SAME CPU, SAME BOARD, SAME BIOS VER 0004
Boot @ 400*9, clockgen to 412*9

Dual 32m load temp
39c Asus probe, 44/41c Core temp
http://www.sspmustang.com/ot/E6600/t1.jpg

Dual 32m idle temp after run
34c Asus probe, 36c Core temp
http://www.sspmustang.com/ot/E6600/t2.jpg

30 minute SP2004 Load temps!
53c Asus probe, 56c Core temp
http://www.sspmustang.com/ot/E6600/t3.jpg

If I leave fans on high, I have seen Asus probe report 30c @ CPU and 23c @ NB

ATI Tool reports 30c idle core and 33c ambient on card when Asus probe reads 35c @ CPU.

With a bad mount and higher ambients, I have seen 72c sp2004 load temps, so these temps ARE correct ;)