Intel Core 2 Chipset Power Consumption Shootout
by Anand Lal Shimpi on October 12, 2006 12:53 PM EST- Posted in
- CPUs
The Test
We did our best to make this a chipset power consumption comparison, but in reality it is just as much of a motherboard power consumption shootout as a chipset one. All extra features -- things like Wireless LAN and IEEE-1394 support -- were disabled to level the playing field as much as possible between platforms. All memory timings and BIOS settings (where applicable) were set identically across all three systems, and of course we used the same physical CPU, power supply, memory and video card for each system to avoid any variation between components.
Our testing methodology was simple: we ran through our usual suite of CPU performance tests, varying only the motherboard, but also recorded power consumption for the duration of each benchmark (beginning when the benchmark starts measuring performance and concluding when it is done measuring performance). Using an Extech 380803 Power Analyzer we were able to log the instantaneous power consumption of our test systems every half a second, giving us reasonable accuracy, especially for the longer tests. For each benchmark that we ran, we plotted performance, total system power consumption and performance per watt.
We were able to run almost all of our tests while measuring power consumption with the exception of PC WorldBench 5, the reason being that some of the WorldBench tests were too short to get accurate power measurements with. We are working on a solution and hope to present WorldBench power consumption numbers in future articles.
| CPU: | Intel Core 2 Duo E6700 (2.66GHz/4MB) |
| Motherboard: | ASUS P5W DH Deluxe (975X) ASUS P5B Deluxe (P965) ASUS P5NSLI (nForce 570 SLI) |
| Chipset: | Intel 975X Intel P965 NVIDIA nForce 570 SLI |
| Chipset Drivers: | NVIDIA 8.22 Intel 8.1.1.1001 |
| Hard Disk: | Seagate 7200.9 300GB SATA |
| Memory: | Corsair XMS2 DDR2-800 4-4-4-12 (1GB x 2) |
| Video Card: | XFX GeForce 7900 GS |
| Video Drivers: | NVIDIA ForceWare 91.47 |
| Desktop Resolution: | 1280 x 1024 - 32-bit @ 60Hz |
| OS: | Windows XP Professional SP2 |
Facebook
Twitter
RSS
44 Comments
View All Comments
PotatoMAN - Friday, December 29, 2006 - link
I am currently trying to in work a PC into my car, and I need to conserve every watt I possibly can (trying to use a DC to DC power supply to avoid losing power from inverting and converting DC->AC->DC). I think there are some limitations right now with the amount of power you can draw from a DC current in a car, and I need to keep my wattage below a certain threshold. Thanks to this review I can base my computer purchasing decisions for my car with some information in hand. Thanks again, Anandtech.Stomper88 - Monday, October 16, 2006 - link
Please post idle wattage as well. My computers sit there at idle much longer than any other state.rkhpedersen - Friday, October 13, 2006 - link
Why didn't you include the idle power consumption. Most computers are running idle far longer than actually doing something. It seems that the most important peice of info is left out of this test.mino - Friday, October 13, 2006 - link
Second that.Please add that chart, these are 3 numbers so it shouldn't be a problem provided you want to add it.
MadBoris - Friday, October 13, 2006 - link
I was just going to close down the site after briefly looking at some of the graphs.But came here first to check and post comments.
This information is pretty useless to me, but atleast a corporation 'may' benefit. Although I don't think corporations buy computer parts with a few watts power draw as the governing factor, unless the IT group has really lost it's way.
As far as Anandtech articles go, testing power draw would be more beneficial practical if showing differences between Core 2 Duo and Athlon X2 platforms. Also between GPU's ATI vs NVIDIA, current gen vs. last gen, etc. now that could be fairly interesting and would likely show a difference that actually means something.
This one was a sleeper for me. I think the idea of testing power draw is sound, just that the focus of testing, in this case, wasn't worth the time. Maybe next time.
mino - Friday, October 13, 2006 - link
Well, maybe we have lost our way ..Chipset power started to really matter the moment NB power got above 10W period.
The problem is heated chipset means more heat in case, less life for board, bigger problem to cool it down => active cooling => lower long-term stability ...
Is it 3W or 6W ? I don't care.
Is it 6W or 15W? That 15W variant is off the table even before I look after the features.
BPB - Friday, October 13, 2006 - link
Too bad they didn't include ASUS' new http://www.zipzoomfly.com/jsp/ProductDetail.jsp?Pr...">P5N32-SLI Premium/WiFi-AP nForce 590 SLI Intel Edition board. That's a board I'm considering for a third system. Right now we have two P5W DH Deluxe boards.Askari77 - Friday, October 13, 2006 - link
I'd like to see the results from a Intel 945G motherboard test as I heard it will support the Core 2 processor line. Unfortunately though I have a Gateway proprietary 945G board (codename Big Lake) but still built by Intel. So I'm more interested in compatibility rather than power consumption. Currently have a Pentium D, though I like the sound of the Core 2's.peldor - Friday, October 13, 2006 - link
Really, charts are great when you've got a half dozen video cards in a roundup and the results are different in each test. But in this case the results are essentially the same in every test: high, middle, low. It would have been a lot more efficient to present these in a table. Red for high, blue for low. All the results fit on one page.Staples - Friday, October 13, 2006 - link
Anandtech has never done an article that pitted the 865, 875 and 570 head to head on a Core 2. Sure there have been old reviews with both the intel chipssets with P4s but those were long ago and the performance may not be the same with the Core 2.Glad to see some benchmarks finally. Maybe I am crazy but this is the only hardware site I visit and I have been disapointed that you were lacking head to head chipset benchmarks when the Core 2 has been out for several months already.