Announcement

Collapse
No announcement yet.

Does Chrome Burn Through More Power Than Firefox?

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Does Chrome Burn Through More Power Than Firefox?

    Phoronix: Does Chrome Burn Through More Power Than Firefox?

    With my recent work in tracking down Linux power regressions and looking at other areas of Linux power consumption, there's been a number of emails sent in by Phoronix readers concerning the power consumption of web-browsers. In particular, some users seem to think that Google's Chrome/Chromium web-browser causes the system to go through noticeably more power than Mozilla Firefox and other web-browsers. But how much is this really the case? Here's some benchmarks.

    http://www.phoronix.com/vr.php?view=16294

  • #2
    Thanks again for a nice test.
    I would appreciate similar tests for newer build of FF and Chrome to get a hint of what to expect from future releases

    Comment


    • #3
      Real Sites

      Approximately zero users are complaining about battery performance while running benchmarks. PTS needs to be updated to visit real sites and interact with them and measure the results of doing that. In previous testing I have done like that you also need to test with an ad blocker otherwise nothing is repeatable.

      Comment


      • #4
        How did you get Firefox?

        Was it a download from Mozilla.com, an Ubuntu repository, or did you compile it yourself?

        I would have liked to have seen FF6 results since it is coming out on Tuesday.

        I also think you may not have been testing fully realistically. What happens when you just open a few websites in tabs and let it idle with a bunch of Flash ads playing? That's probably a more interesting test most of the time than just looking at how much power it consumes while running a js benchmark, which you would hope would max out the CPU most of the time leading to pretty similar power results.

        Comment


        • #5
          id be interested to know about the power wasted by hidden tabs. as i type this i have several other tabs open in the background. some of these hidden tabs have active javascript checking for new messages or scrolling headlines, animation (gifs, etc). from looking at the CPU monitor it appears that some of this is still active even though no one is looking at it.

          firefox 5 set limits on javascript timeouts in background tabs which probably helps, but i am sure there is still some waste. i think gifs and flash still get rendered in a hidden tab.

          Comment


          • #6
            This is a test, that might be biased, but uses real word loads: https://blogs.msdn.com/b/ie/archive/...xplorer-9.aspx

            Comment


            • #7
              ... and opera?

              srsly, I would be highly interrested in how they compete ... esp. as they use the same rendering core in opera mobile ...

              Comment


              • #8
                This is from Microsoft's tests on the Windows platform.



                Comment


                • #9
                  Dev builds

                  I'm a little confused on why you would be running a developmental kernel, developmental video drivers, and not developmental builds of Firefox and Chromium. I'm biased as a Firefox guy and there are noticeable performance gains upstream. They're just started in their rapid release system, but the Azure Direct2D backend for canvas is in version 7, Electrolysis is being implemented in the Nightlies as well as considerable memory reductions.

                  I don't feel like the test is an accurate measure of today's performance when the stable releases are months old (not to mention that Firefox 4 is obsolete). If anything, the test should have compared the latest stable AND dev versions of each respective browser.

                  Comment


                  • #10
                    Originally posted by abral View Post
                    This is a test, that might be biased, but uses real word loads: https://blogs.msdn.com/b/ie/archive/...xplorer-9.aspx
                    The MS tests are very interesting, but I'm not sure they mean much when it comes to Linux testing.

                    For starters, Firefox/IE9 are able to use the GPU much more on Win7 than Chrome is, which they claim is the reason for their lower power usage. That's not really the case on Linux since OpenGL layers aren't enabled and I think both should be using XRender (I know FF is, and assume that Chrome probably is). They are also using completely different toolkits, etc. on Linux which could cause big differences.

                    Comment


                    • #11
                      Originally posted by smitty3268 View Post
                      I also think you may not have been testing fully realistically. What happens when you just open a few websites in tabs and let it idle with a bunch of Flash ads playing?
                      Surely if you care about power consumption then blocking Flash ads will be one of the first things you do?

                      When I still ran Windows on my old laptop I used to be regularly annoyed when I noticed my battery had dropped 50% in the last half hour and discovered some Flash crap running in a window I hadn't looked at in that time.

                      Comment


                      • #12
                        To me, on older hardware, Chromium has been the obvious choice for a while now. It really does squeeze a lot of performance out of a little bit of hardware. The luxury of having parts of the browser written in assembly language.

                        My Desktops run firefox though.
                        Last edited by Sidicas; 08-15-2011, 11:51 AM.

                        Comment


                        • #13
                          Originally posted by movieman View Post
                          Surely if you care about power consumption then blocking Flash ads will be one of the first things you do?

                          When I still ran Windows on my old laptop I used to be regularly annoyed when I noticed my battery had dropped 50% in the last half hour and discovered some Flash crap running in a window I hadn't looked at in that time.
                          Well, OK. I can think of a few useful tests.

                          1. Playing some Flash video on Youtube.
                          2. Loading a few common websites in tabs and idling, with Adblock active. I'm guessing Michael hates Adblock enough to never do this, though.
                          3. Do the same, while not blocking Flash/ads.

                          I can sort of see how a 4th test run while doing some JS test might be an interesting additional test, but it would be by FAR the least useful IMHO. Unfortunately, that's the only 1 michael tested here.

                          Comment


                          • #14
                            In my opinion, I think a proper Browser test should be to take a few random pages from each of the top 100 websites on the Internet (ads, flash content, animated gifs, javascript, flash videos (play from start to end), etc. etc..).. Create a local download / copy of all that and see how the browsers perform rendering through all that, pushing the content into 3 or 4 open tabs...

                            The existing benchmarks seem to be very specialized making them pretty irrelevant for real-world use.

                            As far as CPU usage on flash content, I'd think that entirely depends on what version of the flash plugin you have and what features are getting pushed off to the GPU..

                            On some of the javascript benchmarks (like the V8 javascript test) Chromium wins by a landslide, I've seen some results show Firefox getting slaughtered nearly 10:1 (with newer versions of Firefox, it now appears to only be about 2:1).. On other Javascript benchmarks, Firefox is faster.. So it only makes sense to pull REAL content right from the web to benchmark with.. No more synthetics!
                            Last edited by Sidicas; 08-15-2011, 01:44 PM.

                            Comment


                            • #15
                              Note to Michael - Yes, I would like to see further testing with Firefox 6, Firefox 7 beta and Chrome/Chromium 14 please.

                              Originally posted by smitty3268 View Post
                              The MS tests are very interesting, but I'm not sure they mean much when it comes to Linux testing.

                              For starters, Firefox/IE9 are able to use the GPU much more on Win7 than Chrome is, which they claim is the reason for their lower power usage. That's not really the case on Linux since OpenGL layers aren't enabled and I think both should be using XRender (I know FF is, and assume that Chrome probably is). They are also using completely different toolkits, etc. on Linux which could cause big differences.
                              More hardware-accelerated OpenGL stuff has been enabled on a wider array of hardware in Firefox 6, which is one reason why I would like to see testing with Firefox 6. I agree about the potential difference between Linux and Win7 due to different toolkits.

                              ssam wrote - "I think [animated] gifs and flash still get rendered in a hidden tab."

                              In Flash Player 10.1 SWFs (flash videos) on hidden tabs and/or SWFs scrolled out of view are throttled, resulting in a dramatic reduction in CPU usage. See http://www.kaourantin.net/2010/03/timing-it-right.html
                              Unfortunately this does not work on Linux (nor on Opera 10.1 on Windows) - see the second part of http://www.kaourantin.net/2010/04/pr...-continue.html

                              Comment

                              Working...
                              X