Is Kentsfield really a power hunger monster?!let's see the power consumption !

Printable View

10-28-2006, 07:29 PM
coolaler

Is Kentsfield really a power hunger monster?!let's see the power consumption !

Is Kentsfield really a power hunger monster?!let's see the power consumption comparison between quad-core kentsfield and dual core core2 extreme!

platform
M/B ：ASUS P5B Deluxe
MEM ：TREEM DDRII 800 2GB Dual Channel
HDD ：Seagate 250g SATAII 16Mb Cache
VGA CARD ：MSI 7950GT
POWER: ACBel iPLUS 550W
OS ：WIN SEVER 2003
CPU COOLER：Tower 120
CPU ：Intel Core 2 Duo Conroe E6600 4mb cache
CPU ：Intel Core 2 Duo Kentsfield QX6600 8mb cache

ACBel iPLUS 550W

http://www.iamxtreme.net/coolaler/conroe/power/PSU.JPG

when dual core Vcore=1.35v,3.6G idle
From monitor reading,power consumption is 89W(stable)

http://www.iamxtreme.net/coolaler/conroe/power/C2D1.JPG

monitor software reading
5v 5A
3.3v 5A
12v2 4A

http://www.iamxtreme.net/coolaler/conroe/power/C2D1.gif

dual core 3.6G running Dual SP2004
the highest power consumption reading is 176W(150W~165W up and down)

http://www.iamxtreme.net/coolaler/conroe/power/C2D2.JPG

From monitor software reading
5v 5A increasing to 7A
3.3v 5A no changed
12v2 4A increasing to 10A

http://www.iamxtreme.net/coolaler/conroe/power/C2D2.gif
10-28-2006, 07:29 PM
coolaler

Now let's see quad-cores
kensfield Vcore=1.50v,3.6G idle
From monitorreadings,power consumption is 103W(stable),it's more than dual core around 14W

http://www.iamxtreme.net/coolaler/conroe/power/C2Q1.JPG

From monitoring software
5v 6A
3.3v 4A
12v2 6A

http://www.iamxtreme.net/coolaler/conroe/power/C2Q1.gif

Quad cores3.6Grunning the Dual SP2004,
The highest moment for power consumption is 257W(225W~241W jumping,even though it's only running 2 cores at the same time
; but it is still more than the dual core's 176W about 81W

http://www.iamxtreme.net/coolaler/conroe/power/C2Q3.JPG

from monitor reading software
5v 6A increased to 8A
3.3v 4A no chang
12v2 6A increased to 17A

http://www.iamxtreme.net/coolaler/conroe/power/C2Q3.gif

now, I let the quad core running Quad SP2004
the highest moment for power consumption increased to 298W,(most of the time for jumping around )

http://www.iamxtreme.net/coolaler/conroe/power/C2Q2.JPG

from monitor reading software
5v 6A increased to 9A
3.3v 4A no changed
12v2 6A increased to 19A

http://www.iamxtreme.net/coolaler/conroe/power/C2Q2.gif

Conclusion:At the same freq of 3.6GHZ
Quad core cpu power consumption is 1.16 times off dual core
for the highest moment of power consumption then will be the 1.69 times off dual core.
10-28-2006, 07:44 PM
NickS

Pretty neat result coolaler. Pop in some crossfire action and measure power consumption :)
10-28-2006, 08:05 PM
eva2000
nice coolaler... try also cpu load test that TAT has as it can provide a higher cpu /power load draw that stressprime can i.e. using Power-Mate power meter

System:
- Intel Core 2 Duo E6600 QPGE B1 Step 5 ES
- Scythe Infinity + 2x Thermaltake 120x25mm 81cfm push-pull
- Intel D975XBX Bad Axe rev304 1334 bios
- 128MB Gainward FX5200 PCI card
- 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 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.

@2400Mhz at 1.275v bios (1.259v)
Windows Idle watts = 122.5W / (0.563A)
Windows Orthos Load watts = 164.8W
Windows TAT Load watts = 173.5W / (0.777A)

@3600Mhz at 1.425v bios (1.394v)
Windows Idle watts = 146.2W / (0.673A)
Windows Orthos Load watts = 227.4W
Windows TAT Load watts = 248W / (1.094A)

As suspected TAT's internal cpu load test heats up the cpu more than dual stressprime Orthos load testing and it shows in the power consumption readings :)

I'm using 128MB Gainward FX5200 PCI card video card so naturally system power consumption is much lower. I will test again with higher end video cards later :)
10-28-2006, 08:08 PM
nn_step

almost double... thanks for the testing :up:
10-28-2006, 08:20 PM
zabomb4163

any way you could test an amd dual core setup?
10-28-2006, 08:31 PM
coolaler
Quote:
Originally Posted by eva2000

nice coolaler... try also cpu load test that TAT has as it can provide a higher cpu /power load draw that stressprime can i.e. using Power-Mate power meter

System:

Intel Core 2 Duo E6600 QPGE B1 Step 5 ES
Scythe Infinity + 2x Thermaltake 120x25mm 81cfm push-pull
Intel D975XBX Bad Axe rev304 1334 bios
128MB Gainward FX5200 PCI card
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 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.

@2400Mhz at 1.275v bios (1.259v)
Windows Idle watts = 122.5W / (0.563A)
Windows Orthos Load watts = 164.8W
Windows TAT Load watts = 173.5W / (0.777A)

@3600Mhz at 1.425v bios (1.394v)
Windows Idle watts = 146.2W / (0.673A)
Windows Orthos Load watts = 227.4W
Windows TAT Load watts = 248W / (1.094A)

As suspected TAT's internal cpu load test heats up the cpu more than dual stressprime Orthos load testing and it shows in the power consumption readings :)

I'm using 128MB Gainward FX5200 PCI card video card so naturally system power consumption is much lower. I will test again with higher end video cards later :)
hey!u r expert:banana:
10-28-2006, 09:04 PM
erwinz

Thanks Coolaler :woot:
10-28-2006, 09:24 PM
HKPolice

Errrrr.... The dual core was running @ 1.35v and the Quad was 1.5v. This makes the results invalid.
10-28-2006, 09:29 PM
eva2000

Quote:

Originally Posted by coolaler

hey!u r expert:banana:

thanks but I just have more money than sense just bought the Power-Mate power meter cost AUD$305 (US$230) :eek: and have been running around the house testing every appliance i can find with my new toy LOL

Would be interesting to see the QX6700 loads with TAT cpu test on 4 cores versus quad stressprime loads. If TAT increases power consumption by 10% over stressprime that would put +12v2 rail close to 21-22A maxing it out :D
10-28-2006, 09:43 PM
dinos22

Quote:

Originally Posted by eva2000

thanks but I just have more money than sense just bought the Power-Mate power meter cost AUD$305 (US$230) :eek: and have been running around the house testing every appliance i can find with my new toy LOL

Would be interesting to see the QX6700 loads with TAT cpu test on 4 cores versus quad stressprime loads. If TAT increases power consumption by 10% over stressprime that would put +12v2 rail close to 21-22A maxing it out :D

ROFL you jackass

watches George hook it up to microwave while making popcorn hahahah
10-29-2006, 12:15 AM
Tony

Quote:

Originally Posted by eva2000

thanks but I just have more money than sense just bought the Power-Mate power meter cost AUD$305 (US$230) :eek: and have been running around the house testing every appliance i can find with my new toy LOL

Would be interesting to see the QX6700 loads with TAT cpu test on 4 cores versus quad stressprime loads. If TAT increases power consumption by 10% over stressprime that would put +12v2 rail close to 21-22A maxing it out :D

hey George, I just have to get me one of those ;)
10-29-2006, 12:41 AM
bachus_anonym

Yeah, those meters are really useful and fun. I've had one one of those for about 2yrs now. Not as fancy as coolaler's and not as pretty as eva's, but it does what it's supposed to :)

That's my E6600 @ 4100MHz, 1.63v - during ORTHOS (almost 2hrs into LargeFFT as I'm playing with memory, as we speak). This power consumption includes: P5B DLX, 2GB DDR2, 1xHDD, 1xCDRW, 2x80mm fans, PCI videocard (7800GTX 256MB adds 30-40W, during ORTHOS since it's not in 3D mode).

Crappy pic, which reminds me I need to clean this thing a bit :D

http://www.clockmehigh.com/cooler/sc.../killawatt.jpg

Quote:

Originally Posted by HKPolice

Errrrr.... The dual core was running @ 1.35v and the Quad was 1.5v. This makes the results invalid.

Yeah, for the "apples-to-apples" comparison, coolaler would have to run his E6600 at 1.50v as well...
10-29-2006, 12:44 AM
Tony

So i just hit the buy button, for UK customers i just got this http://www.energyoptimizersdirect.co...log/index.html

30pounds delivered, and will work in australia also with a converter for both in and out ;)
10-29-2006, 03:13 AM
dinos22

Quote:

Originally Posted by Tony

So i just hit the buy button, for UK customers i just got this http://www.energyoptimizersdirect.co...log/index.html

30pounds delivered, and will work in australia also with a converter for both in and out ;)

lol hits buy now button
10-29-2006, 04:04 AM
The Asgard

Nice results. Better than I expected.

Thanks for taking the time to perform the test :)
10-29-2006, 04:19 AM
lawrywild

Quote:

Originally Posted by Tony

So i just hit the buy button, for UK customers i just got this http://www.energyoptimizersdirect.co...log/index.html

30pounds delivered, and will work in australia also with a converter for both in and out ;)

20quid, oo that's pretty good :)
10-29-2006, 04:22 AM
DTU_XaVier

Great article Coolaler... Goes to show that the nice, powerful (no pun intended) PSU will go a long way without needing upgrade, whereas if you buy inferior (lower-wattage) PSUs, you'll be changing them in no time :)
Again, great read, now show us som 3dmark 06 action ;)

Best Regards :toast:
10-29-2006, 04:54 AM
rozzyroz

Quote:

Originally Posted by nn_step

almost double... thanks for the testing :up:

actually, coolaler said that it was 16% more at idle and 69% more under load.

Quote:

Originally Posted by coolaler

Conclusion:At the same freq of 3.6GHZ
Quad core cpu power consumption is 1.16 times off dual core
for the highest moment of power consumption then will be the 1.69 times off dual core.
10-29-2006, 07:08 AM
Hooker

Ok, question from the uneducated.

Should I want to build a QX6700 with a G80-GTX/r600-XTX with 4GB PC2-6400 & 3 HDD's (2 of which are Raptors)

Would I need a 600 of 700 Watt PSU? considering the G80/r600 use about 450W on load - Seasonic M12 series is my PSU of choice ATM
10-31-2006, 02:44 PM
thunderstruck!

Quote:

Originally Posted by Hooker

Ok, question from the uneducated.

Should I want to build a QX6700 with a G80-GTX/r600-XTX with 4GB PC2-6400 & 3 HDD's (2 of which are Raptors)

Would I need a 600 of 700 Watt PSU? considering the G80/r600 use about 450W on load - Seasonic M12 series is my PSU of choice ATM

You will be taxing your whole PSU at stock. If you want to overclock or even upgrade down the road, that PSU will be worthless.
11-01-2006, 03:06 AM
mach82

Quote:

Originally Posted by Hooker

Ok, question from the uneducated.

Should I want to build a QX6700 with a G80-GTX/r600-XTX with 4GB PC2-6400 & 3 HDD's (2 of which are Raptors)

Would I need a 600 of 700 Watt PSU? considering the G80/r600 use about 450W on load - Seasonic M12 series is my PSU of choice ATM

E6400 3GHz+8800GTX - a 470W 30A 12V PSU works absoutely fine (1HDD, 1DVD-RW, 2GB RAM). So it's not so bad when it comes to energy consumption by the newest G80.
11-01-2006, 03:13 AM
K404

Quote:

Originally Posted by Hooker

considering the G80/r600 use about 450W on load

That for SLI?
11-01-2006, 03:53 AM
irev210

I will post kentfield w/ 1.7vcore under load shortly

you will be amused. lol
11-01-2006, 05:31 AM
uOpt

BTW, here is the consumption of my 4x 875 Opteron for comparision:

Code:

zippy 700W OCZ 700W OCZ/fanctrl Zip/fanctrl soft off: 7 6 6 7 idle: 289 282 308 8x r@h: 398 416 8x mprime: 487 455 453 482
11-01-2006, 06:25 AM
IluvIntel

Peaking at 69% increse over dualcore with full load.... not sure if these Kentsfields are going to be worth it for my use.
11-01-2006, 06:39 AM
dinges

Quote:

Originally Posted by Hooker

Ok, question from the uneducated.

Should I want to build a QX6700 with a G80-GTX/r600-XTX with 4GB PC2-6400 & 3 HDD's (2 of which are Raptors)

Would I need a 600 of 700 Watt PSU? considering the G80/r600 use about 450W on load - Seasonic M12 series is my PSU of choice ATM

:eek: een mede utrechter op xs :toast:

if you are spending that much on a new rig, i'd go with one of the 1k models (pp&c, ocz, others..?). with that you'll be having more then enough power for this and future rig's (for example, you can add an extra vid. card later for xfire/sli without having to worry about you're psu).

OT: Sick power consumption on the kentfield, close to 300w for a cpu..:shock: thanks for sharing this coolaler.
although i have to agree with hkpolice, both cpu's at the same vcore would make a better compare.
11-01-2006, 05:42 PM
Frank M

Quote:

Originally Posted by dinges

if you are spending that much on a new rig, i'd go with one of the 1k models (pp&c, ocz, others..?).

OT: Sick power consumption on the kentfield, close to 300w for a cpu..:shock: thanks for sharing this coolaler.

Yeah, but those others are nowhere near as silent ;)

As for #2: that was not just the cpu, but the entire pc.

Also, I don't think the next generation flagships will use
that much power. 450Ws under load just for a graphics
card sound like an awful lot to me... not to mention, how
to cool that beast? :eek:
11-01-2006, 08:23 PM
G H Z

Quote:

Originally Posted by HKPolice

Errrrr.... The dual core was running @ 1.35v and the Quad was 1.5v. This makes the results invalid.

He's showing us power consumption at identical Mhz. Even if they are not valid in the strictest sense they are pretty useful either way you look at it. It's possible to infer approximate numbers anyway and this also shows us that Kentsfield needs more juice than C2D for stability at similar Mhz.
11-01-2006, 08:53 PM
Revv23

Good test thanks coolaler, imo for 2x the amount of cores at a higher voltage 60-70% more power is not so much.
11-01-2006, 10:13 PM
OC_Windforce

wow~! good power :woot:
11-02-2006, 06:42 PM
computerman0416

Am I reading the temperatures right, 87 degrees celcius on the CPU. Thats roasting!! Whats the thermal specifications for the Kentsfields???
11-02-2006, 07:04 PM
Revv23

Quote:

Originally Posted by Revv23

Good test thanks coolaler, imo for 2x the amount of cores at a higher voltage 60-70% more power is not so much.

another thing to consider is that it will be rare for us to use all 4 cores.
11-02-2006, 07:05 PM
uOpt

Quote:

Originally Posted by computerman0416

Am I reading the temperatures right, 87 degrees celcius on the CPU. Thats roasting!! Whats the thermal specifications for the Kentsfields???

As with most Core2 temperature reports - it's bogus.
11-02-2006, 07:44 PM
775911

Whats the real temp then because personally I think one day heatsinks will come with little copper frying pans on top :toast:
11-02-2006, 08:20 PM
Revv23

i dunno, its hot though! :p:

IMO as long as its stable it doesnt matter
11-02-2006, 08:24 PM
nn_step

Quote:

Originally Posted by uOpt

BTW, here is the consumption of my 4x 875 Opteron for comparision:

Code:

zippy 700W OCZ 700W OCZ/fanctrl Zip/fanctrl soft off: 7 6 6 7 idle: 289 282 308 8x r@h: 398 416 8x mprime: 487 455 453 482

That kind of kills the performance per watt now doesn't it
11-02-2006, 08:30 PM
Ychocky

Very informative, thanks for the testing :)
11-03-2006, 07:21 AM
zerolise

Hey Coolaler, is it possible to keep both chips' Vcore the same in your set-up? That might be a better way to compare.
11-03-2006, 07:26 AM
uOpt

Quote:

Originally Posted by nn_step

That kind of kills the performance per watt now doesn't it

Why?

It think it's pretty low. Keep in mind it is 8 cores. So it mprimes at 56 watts per core using the OCZ unit. And that is 56 watts out of the A/C. Since the OCZ has 80% efficiency pretty much everywhere that means each core takes 45 watts, mpriming, and that is with 1 GB of RAM per core.
11-03-2006, 12:41 PM
JoeBar

Thanks for testing! :clap:
11-06-2006, 06:10 PM
coolaler

kentsfield 4200mhz(phase change)+ 8800GTX single(650/2100)
when ran 3D06,peak value 308 watts(all platform)

http://www.iamxtreme.net/coolaler/8800GTX/308W.JPG

3D06

http://www.iamxtreme.net/coolaler/88...ENTS4200_3.gif
11-06-2006, 06:19 PM
eva2000

nice coolaler.. what's the evap idle and load temps for QX6700 @4.2Ghz ? vcore ?
11-06-2006, 06:23 PM
shimmishim

first of all, nice results coolaler! :)

AT released an article a couple of days ago that does a performance per watt analysis with the new quad core vs. the current dual core

http://www.anandtech.com/cpuchipsets...oc.aspx?i=2866
11-06-2006, 07:21 PM
kraggy

Where can I get one of those cool watt meters that coolaler has? Looks like I'm gona need one for my Kentsfield setup too. Can you buy them here in the states?
11-06-2006, 07:48 PM
Revv23

nice 14k in 06 single card!!!

Thanks coolaler!
01-16-2007, 05:49 PM
Fixt00l

Thank you so much for the wonderful testing...It confirms that Kentsfield is way too early to appear for the masses,as it consumes a lot of power,runs H O T,which means OC cant be as good as the E6400:fact: ,and,of course,the price is...:stick:

All times are GMT -8. The time now is 09:38 AM.

XtremeSystems