No announcement yet.

Ubuntu 10.04 Is More Power Hungry Than Windows 7

  • Filter
  • Time
  • Show
Clear All
new posts

  • Ubuntu 10.04 Is More Power Hungry Than Windows 7

    Phoronix: Ubuntu 10.04 Is More Power Hungry Than Windows 7

    Yesterday we published our first benchmarks of Windows 7 vs. Ubuntu 10.04 that provided an initial look at the OpenGL graphics performance between these two operating systems on six different systems. Today we are continuing to compare the two operating systems as we look at the power consumption of Ubuntu and Windows on a netbook and notebook.

  • #2
    Ubuntu is pretty bad power-wise out of the box. On UNR 9.10 laptop mode was enabled out of the box (I think), but all the useful power saving tricks were turned off. On my Dell Mini 10v I got another 30-45 minutes using powertop and tweaking stuff in laptop mode. Most of the tips given by powertop are settings in the laptop mode configuration directory.



    • #3
      I think what's interesting here is that in case of Linux a C2D+dedicated video+4GB+15" consume less that Atom330+onboard video+2GB+12" !!!


      • #4
        Very interesting subject for me, but at the same time very heart-breaking content for me.

        Thx for the (bad) news Michael (jaja).

        Can you please try with Debian? Stable or testing you'd choose.


        • #5
          Excuse me. Do you have tested Ubuntu Desktop or Ubuntu Netbook Remix over Netbooks?
          Thank you.


          • #6
            Thank you Michael.
            On my Lenovo T400 I have noticed that Ubuntu 10.04 by default runs at max LCD screen brightness, but Windows Vista at 75%. It is very easy to adjust, but the default setting of Ubuntu Desktop are not good for notebook at all. Did you notice the same difference on your test laptops?


            • #7
              1 I don't know if the difference between 6 hours or 8 hours on a Acer Aspire One would even matter?

              2 You did not take into account the fact that on battery, Windows can be hellish slow, certainly on Atoms. This might be because it stays to long in a low-power state when the user asks for responsiveness. I would prefer Ubuntu over that.

              In the previous articles, the performance was tested and the battery power, but not a combination of both.


              • #8
                Nothing productive to add, but I think my dad has that exact same AC power meter.


                • #9
                  Thanks for the results.

                  One thing i have been wishing for with ubuntu is the they give up the drive for boot time and now focus on power.

                  If desktop linux was the most power efficient operating system then it would have a massive selling point. I feel that people don't really care about software freedom or about database benchmarks. But tell them that you could cut 30 from the power bill each year and they will start to pay attention.

                  My personal experience on an acer timeline 3810tg is on stock Ubuntu (i.e open ati drivers) comes in at 20 watts. With the ati propitiatory driver it comes in at around 10-11 watts. With a few of tweaks it comes in at 9 watts.

                  Could i plead with phoronix to do more power benchamrks, id like to see what distros use the least power, the effect of different kernels on power. i.e the Ubuntu-prempt kernel vs the generic vs zen.


                  • #10
                    @Phoronix: Did you check, that Ubuntu really uses it's powersave function? I ask, because I've two Notebooks, that were affected by a bug in the Kernel (or Gnome-Power-Manager in constellation with the Lucid Kernel), that refuses the backlight-dimming to work. It worked great with Karmic on both Notebooks, but didn't work with Lucid. You can't even change the brightness with the Fn+F[7|8] keys.


                    • #11
                      Most likely the backlight is set to 100%, thats what needs much more power. Brighness is not controlled via backlight by default. I would disable compiz too.


                      • #12
                        Do note that Ubuntu boots into performance and reverts to ondemand after 60 seconds or so. This is done to improve boot times and does impact power consumption.

                        (I disabled this feature today in favor of using the "conservative" governor everywhere. My mobile Core 2 doesn't really need the extra performance to boot).


                        • #13
                          a bit to generous headline for the article

                          you test a netbook and use also one of the few ion-versions. Thats not very representive because there is no good open nvidia driver, and the closedsource drivers all consume to much energie. Use a Netbook with intel grafics or use a small notebook with integrated ati hardware, thats representative.

                          That Nvidia is the worst grafics solution for linux do we all now. Call it with nividia grafics + ubuntu linux is more power hungry than it is fair. Or alternativly test at least one of each other grafic-chip vendors.


                          • #14
                            Originally posted by BlackStar View Post
                            I disabled this feature today in favor of using the "conservative" governor everywhere. My mobile Core 2 doesn't really need the extra performance to boot).
                            You should be aware that you are actually crippling[1] the ability of your processor to save power by forcing a lower clockrate. Unless you are worried about your thermal envelope, there is absolutely no reason to deviate from the ondemand governor.

                            Modern processors save the most power when they have long periods of inactivity as they can then spend longer in deeper C-states. By slowing the processor down, you are requiring that it take longer to complete its work and therefore spend more time dissipating more power.



                            • #15
                              Sorry, but is there nobody who finds the results of the test a little bit strange?

                              To refresh your memories, the raw numbers from the test :

                              Netbook 25W, 39W, 20W, 33W (W7,Ubuntu,W7,Ubuntu)
                              T61     28W, 32W, 24W, 25W (W7,Ubuntu,W7,Ubuntu)
                              Diff.   -3W,  7W, -4W,  8W
                              Come on, a far more powerful T61-Notebook consumes less energy or in worst case 4W more than a netbook? A Core2Duo and a Nvidia Quadro card consume less energy than a Atom 330 and Geforce 9400 aka ION? How possible is this? The numbers absolutely make no sense.

                              The only explanation is that the power-meter, used in the test, does not work correctly with the netbook battery-charger.

                              Please Michael, check your numbers before writing an such an article.

                              P.S. :

                              I've just checked my own notebook (Dell Latitude D505, P-M 1.7Ghz, Linux-PHC-patched kernel) with a power-meter (No-name brand, bought at ALDI South, Germany) :

                              Low Brightness, 600Mhz (Idle)      -> ~15W
                              Low Brightness, 1700Mhz (Idle)     -> ~16W
                              Low Brightness, 1700Mhz (CPUburn)  -> ~25W
                              High Brightness,  600Mhz (Idle)    -> ~21W
                              High Brightness, 1700Mhz (Idle)    -> ~22W
                              High Brightness, 1700Mhz (CPUburn) -> ~30W