Announcement

Collapse
No announcement yet.

Intel Developer Finds 50 Watt Power Regression In Linux

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

  • Intel Developer Finds 50 Watt Power Regression In Linux

    Phoronix: Intel Developer Finds 50 Watt Power Regression In Linux

    An Intel Linux kernel developer has discovered that since the Linux 3.10 kernel one of his powerful Xeon-based systems is going through 50 Watts more energy while idling than on pre-3.10 kernels. This power regression could affect smaller systems too, but fortunately the issue has been bisected and an investigation is in process...

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

  • #2
    Is there any variable or loop or anything in the Linux kernel that can't be overridden by a kernel command line param?
    How about the variables from the stack, those too?

    Comment


    • #3
      Isn't that when intel_pstate was added?

      Comment


      • #4
        So maybe someone should buy a few energy meters for kernel regression testing. It wouldn't be hard to automate it.

        Comment


        • #5
          Originally posted by toyotabedzrock View Post
          So maybe someone should buy a few energy meters for kernel regression testing. It wouldn't be hard to automate it.
          It's already fully-automated by PTS. The only thing missing me from routinely monitoring it is only having one WattsUp meter (~$100 per unit) and not running the very high-end systems constantly (but only lower-end Core i3/i5s or Atoms) due to energy costs.
          Michael Larabel
          http://www.michaellarabel.com/

          Comment


          • #6
            Originally posted by Michael View Post
            It's already fully-automated by PTS. The only thing missing me from routinely monitoring it is only having one WattsUp meter (~$100 per unit) and not running the very high-end systems constantly (but only lower-end Core i3/i5s or Atoms) due to energy costs.
            Why don't you create some such amazon want list. Last i remember you can compile an i want list on amazon, like a santa christmas list. You can call it phoronix santa sack.

            And obviously post an article and link it.
            Hardware, tools, etc

            Comment


            • #7
              Originally posted by Michael View Post
              It's already fully-automated by PTS. The only thing missing me from routinely monitoring it is only having one WattsUp meter (~$100 per unit) and not running the very high-end systems constantly (but only lower-end Core i3/i5s or Atoms) due to energy costs.
              Hey, Michael, is there any way to do a per-component breakdown of watt usage? Was thinking maybe do a comparison between Linux and Windows 8 and figure out exactly which subsystems are more or less power hungry. Because everyone going "We need better power efficiency in the kernel!" is great and all, but it'd be a lot more effective if they could go "Okay, see, under Linux the CPU is using more energy even while idle compared to Windows." or "The memory is using more power while idle than Windows" give developers targeted areas to focus on, ya know?

              Comment


              • #8
                Originally posted by Ericg View Post
                Hey, Michael, is there any way to do a per-component breakdown of watt usage? Was thinking maybe do a comparison between Linux and Windows 8 and figure out exactly which subsystems are more or less power hungry. Because everyone going "We need better power efficiency in the kernel!" is great and all, but it'd be a lot more effective if they could go "Okay, see, under Linux the CPU is using more energy even while idle compared to Windows." or "The memory is using more power while idle than Windows" give developers targeted areas to focus on, ya know?
                I believe that's only becoming now possible with the very latest Intel (and maybe) AMD hardware. I believe it's mostly server-focused spec. I think I've only heard of the support/spec being implemented though on server hardware and not even Haswell desktop systems. There's also other unstandardized items like some AMD CPUs having a fam15h_power driver or whatever that is supposed to report like CPU power consumption individually, but have found that to be inaccurate garbage (at least as of a few kernels ago), etc.
                Michael Larabel
                http://www.michaellarabel.com/

                Comment


                • #9
                  Originally posted by Ericg View Post
                  Hey, Michael, is there any way to do a per-component breakdown of watt usage? Was thinking maybe do a comparison between Linux and Windows 8 and figure out exactly which subsystems are more or less power hungry. Because everyone going "We need better power efficiency in the kernel!" is great and all, but it'd be a lot more effective if they could go "Okay, see, under Linux the CPU is using more energy even while idle compared to Windows." or "The memory is using more power while idle than Windows" give developers targeted areas to focus on, ya know?
                  Didn't powertop do that?

                  Comment


                  • #10
                    Originally posted by mrugiero View Post
                    Didn't powertop do that?
                    Only by switching things on and off to see how the overall power consumption changed, I think. Still handy though.

                    Comment


                    • #11
                      Originally posted by Michael View Post
                      I believe that's only becoming now possible with the very latest Intel (and maybe) AMD hardware. I believe it's mostly server-focused spec.
                      How are you going with getting PTS used/sponsored by companies like Intel for doing automated testing to find these regressions? Has there been much take up? It's sounding increasingly useful for finding these kinds of issues and I can understand not wanting to run high power drain systems continuously and out of your own pocket.

                      Comment


                      • #12
                        Originally posted by Kamikaze View Post
                        How are you going with getting PTS used/sponsored by companies like Intel for doing automated testing to find these regressions? Has there been much take up? It's sounding increasingly useful for finding these kinds of issues and I can understand not wanting to run high power drain systems continuously and out of your own pocket.
                        PTS is already used by multiple groups within Intel, NVIDIA, AMD, and countless other unnamed companies. I don't go to them seeking funds though for running a public test farm here but assisting them in their custom engineering and deployment needs of which they generally can't make results public.
                        Michael Larabel
                        http://www.michaellarabel.com/

                        Comment


                        • #13
                          Originally posted by Michael View Post
                          It's already fully-automated by PTS. The only thing missing me from routinely monitoring it is only having one WattsUp meter (~$100 per unit) and not running the very high-end systems constantly (but only lower-end Core i3/i5s or Atoms) due to energy costs.
                          The other thing here is the indication was that this was a forty core monster machine ( monster for this year) so we are talking a watt and a fraction for each core. That really isn't a lot so I wonder if I'm interpreting things right here.

                          The other thing is if you are going to benchmark power usage you really need to test a few Apple Mavericks based systems. Comparing or judging Linux against Windows is a bit like packing the panel of judges in your favor.

                          As for systems I'm not convinced a lot of people are interested in the high end systems. I'd rather see good benchmarking on portables and small form factor machines myself. That is machines that are very likely purchased for power savings in the first place.

                          Comment


                          • #14
                            Originally posted by wizard69 View Post
                            The other thing here is the indication was that this was a forty core monster machine ( monster for this year) so we are talking a watt and a fraction for each core. That really isn't a lot so I wonder if I'm interpreting things right here.

                            The other thing is if you are going to benchmark power usage you really need to test a few Apple Mavericks based systems. Comparing or judging Linux against Windows is a bit like packing the panel of judges in your favor.

                            As for systems I'm not convinced a lot of people are interested in the high end systems. I'd rather see good benchmarking on portables and small form factor machines myself. That is machines that are very likely purchased for power savings in the first place.
                            In the lkml thread, the user with a Q6600 quad core reported a 25W difference. That's 6.25W per core.

                            Comment


                            • #15
                              Originally posted by wizard69 View Post
                              The other thing here is the indication was that this was a forty core monster machine ( monster for this year) so we are talking a watt and a fraction for each core. That really isn't a lot so I wonder if I'm interpreting things right here.

                              The other thing is if you are going to benchmark power usage you really need to test a few Apple Mavericks based systems. Comparing or judging Linux against Windows is a bit like packing the panel of judges in your favor.

                              As for systems I'm not convinced a lot of people are interested in the high end systems. I'd rather see good benchmarking on portables and small form factor machines myself. That is machines that are very likely purchased for power savings in the first place.
                              I disagree completely. I think that the people who purchase and operate such "monster machines" are immensely more concerned about the power consumptions of these machines than the owners of portables and small form factor machines are of theirs.

                              Comment

                              Working...
                              X