Sign in with
Sign up | Sign in

Test Setup And A Side Note

AMD's Radeon HD 5000-Series: Measuring Power Efficiency
By

We're using the AMD-based platform primarily because it offers the flexibility to test an integrated graphics processor with both dual- and quad-core CPUs. The integrated graphics allows us to establish baseline power consumption without an add-on graphics card. The platform is also quite popular in the mainstream segment, so it's well-suited for testing mainstream graphics cards like the Radeon HD 5670 and HD 5770. It is system-limited with the Radeon HD 5870, as you will see.

We left the power-saving features on the Phenom II X4 955 BE and Athlon II X2 250 enabled in the BIOS, and we set Windows’ power policy to Balanced. To lower the base system power consumption even more, we altered Cool'n'Quiet with K10Stat, allowing our processors to run at even lower voltages. Using these settings, our test platform base power consumption hovered around 55 watts at idle and 80 watts during H.264 video playback. This is just right around the threshold where our platform's PSU efficiency drops off.

Our altered voltages mean that this isn’t a default setup, per se. So, these results are not directly comparable if you’re running stock voltage settings. You can see the differences in system power consumption between running with default voltages and our undervolted settings below.


IdleCrysisAdobe Photoshop CS 4Cinebench R11.5
Undervolted55
121
120
132
Default Voltage73
160
143
195

*Watts

We saw slightly higher power consumption figures with the add-on graphics cards at idle. Originally, we guessed that the results were just leakage variations (the differences are between 5 to 10 W). However, given further testing, it became clearer that this wasn’t the case. Running at default voltages gives us idle power consumption scores close or similar to each card’s specifications.

The reason for the higher power consumption is active power use by additional components. You can't really shut off all power to a  discrete graphics card, so there will always be some power drain there. Since it employs a lot more circuitry, the drain is larger than the simpler integrated graphics solution. There’s also PSU efficiency to consider. The power levels we’re seeing at idle are around 10% of the PSU’s maximum output.

For PowerDVD tests, the Athlon II X2 250 offers much lower power consumption, so it makes sense to use this processor instead, since it excels in an HTPC environment.

For reference purposes (and just for kicks), we also recorded GPU utilization data for most of the benchmarks used with GPU-Z on a Radeon HD 4670.

Display all 30 comments.
This thread is closed for comments
Top Comments
  • 18 Hide
    Lutfij , August 24, 2010 9:23 AM
    ^ nvidia would loose at this battle. period.
  • 14 Hide
    tony singh , August 24, 2010 6:32 AM
    Very innovative article tom keep it up!! Similar article consisting of various cpus would be really useful.
Other Comments
  • 14 Hide
    tony singh , August 24, 2010 6:32 AM
    Very innovative article tom keep it up!! Similar article consisting of various cpus would be really useful.
  • 3 Hide
    tacoslave , August 24, 2010 6:59 AM
    gtx 480 and 460 for reference?
  • 18 Hide
    Lutfij , August 24, 2010 9:23 AM
    ^ nvidia would loose at this battle. period.
  • 3 Hide
    spidey81 , August 24, 2010 11:20 AM
    I know the FPS/watt wouldn't be as good, but what if the 5670 was crossfired. Would it still be a better alternative, efficiency wise, than say a 5850?
  • 0 Hide
    nforce4max , August 24, 2010 11:50 AM
    Remember the R600 (2900xt) has a 80nm core while the 5870 has a 45nm core. Shrink the R600 and you will get the 3870 (55nm) that barely uses hardly any.
  • 7 Hide
    rhino13 , August 24, 2010 12:34 PM
    And now just for fun we should compare to Fermi.

    Oh, wait, this just in:
    Quote:
    There is a Fermi comparison chart that was avalible but you needed to have two screens to display the bar graph for Fermi's power consumption and temperature. So the decission was made to provide readers with the single screen only version.
  • 0 Hide
    aevm , August 24, 2010 12:51 PM
    I loved this part:

    Quote:
    A mere 20 watts separate the Radeon HD 3300, HD 5670, HD 5770, and HD 5870 1 GB. So, in certain cases, the Radeon HD 5870 1 GB can still save enough power to close in on its more mainstream derivatives. Again, this is the case because the cards use a fixed-function video engine to assist in decoding acceleration, which is the same from one board to the next. Thus, even a high-end card behaves like a lower-end product in such a workload. This is very important, as you will see later on.


    My next PC will be used mostly for movie DVDs and Diablo 3. Apparently if I get a 5870 1GB I get the best of both worlds - speed in Diablo and low power consumption when playing movies.

    How about nVidia cards, would I get the same behavior with a GTX 480 for example?
  • -2 Hide
    Onus , August 24, 2010 12:57 PM
    For those not needing the absolute maximum eye candy at high resolutions in their games, the HD5670 looks like a very nice choice for a do-it-all card that won't break the budget.
    Next questions: First, where does the HD5750 fall in this? Second, if you do the same kinds of manual tweaking for power saving that you did in your Cool-n-Quiet analysis, how will that change the results? And finally, if you run a F@H client, what does that do to "idle" scores, when the GPU is actually quite busy processing a work unit?
  • 0 Hide
    eodeo , August 24, 2010 1:34 PM
    Very interesting article indeed.

    I'd love to see nvidia cards and beefier CPUs used as well. Normal non green hdds too. Just how big of a difference in speed/power do they make?

    Thank you for sharing.
  • 3 Hide
    arnawa_widagda , August 24, 2010 3:42 PM
    Hi guys,

    Thanks for reading the article.

    Quote:
    Next questions: First, where does the HD5750 fall in this? Second, if you do the same kinds of manual tweaking for power saving that you did in your Cool-n-Quiet analysis, how will that change the results? And finally, if you run a F@H client, what does that do to "idle" scores, when the GPU is actually quite busy processing a work unit?


    Have no 5750 sample yet, but they should relatively be close to 5770. For this article, we simply chose the best bin for each series (Redwood, Juniper and Cypress).

    The second question, what will happen when you tweak the chip? Glad you ask!! I can't say much yet, but you'll be surprised what the 5870 1 GB can do.

    As for NVIDIA cards, I'm hoping to have the chance to test GF100 and derivatives very soon.

    Take care.

  • 0 Hide
    JohnnyLucky , August 24, 2010 4:03 PM
    Thanks for including mainstream applications.

    Interesting comments about Furmark.

  • 0 Hide
    xbsoft , August 24, 2010 6:24 PM
    >> TEST SYSTEM: AMD Phenom II X4 955 Black Edition (3.2 GHz, >>>>>> 8
  • -2 Hide
    mattmock , August 24, 2010 6:30 PM
    Quote:
    It just means you typically will never encounter such an extreme usage scenario.

    I have to disagree, there are several ways a user can fully load their graphics card in normal use. I have found that my GPU utilization and fan speed go to %100 when I play the dice mini-game in The Witcher. The game only has to render a small game board and the frame rate goes into the 200-300 range. Some thing similar occurs when I hit the pause key in stalker.
  • 0 Hide
    Fokissed , August 24, 2010 11:03 PM
    xbsoft(3.2 GHz, >>>>>> 8

    bad penis joke?
  • 1 Hide
    MartenKL , August 24, 2010 11:13 PM
    FPS/watt uses average FPS during the test but max wattage? I am very disappointed by this flawed logic from toms hardware. Spending an entire page describing why everyone else uses flawed testing for benchmarking power efficiency and then doing this simple error is just embarrassing.
  • 1 Hide
    MartenKL , August 24, 2010 11:30 PM
    I forgot to say I am very interested in this kind of benchmarks and I am glad Toms Hardware is writing a big piece on it, sorry for the harsh words. For me total Wh per completed task for the entire system is the most interesting number. To me that is the only way to measure efficiency. add to that idle power draw and every user can calculate their own usage (by adding tasks and idle hours). Sorry and thanks yet again for an article with a very important topic. My interest is noise and mechanical wear rather than power cost and environment.
  • 0 Hide
    tubers , August 25, 2010 2:50 AM
    Fermi comparison please :) 
  • 0 Hide
    mayne92 , August 25, 2010 2:45 PM
    What a great article by Tom's (Arnawa)! Probably one of the best articles I have read in a long time! Enjoyed the article because was very detailed and you explained everything so well and I LOVE my tech reviews! A Fermi comparison would have been nice but I know that you said that you don't have them to play with so it's said as a request. Hats off to you Arnawa...for a great read...
  • 0 Hide
    eddieroolz , August 25, 2010 11:07 PM
    Really paints in perspective the power of GPUs compared to CPUs. I really wish that one day we'll be able to use the GPU for central processing.
  • 0 Hide
    EDIGX2 , August 26, 2010 6:00 PM
    Hello everyone
    Well i think this article inspired form the movie that AMD has release lately.
    That movie called as i think "Mis understanding"here's the link: http://www.youtube.com/watch?v=2QkyfGJgcwQ
    As we all know AMD is innovative in power consumption as well It's Graphics I read such this review in Anandtech.com ...Just WOW....Loads Of Noise and power flowed for Fermi VGAs . In this review we see the smooth performance for 5670 and 5770.
    and another thing that we should give a hint on is You know releasing Fermi after six month of releasing 5000 series...I think it's good in performance but not after 6 Months!!! but awful in power consuming and noise and heat!!
    Take care guys
Display more comments