Announcement

Collapse
No announcement yet.

What To Do If Still Seeing Poor Linux Battery Life

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

  • #16
    But I do have to say that all these Bavaria references kinda get annoying over time.
    Annoying? Every time i read such a reference i barely can hold my stomach content ;-)

    Comment


    • #17
      I'm the owner of a new dell E6420 laptop (sandybridge cpu) and tested the pcie_aspm=force boot option impact.

      Results are here: http://openbenchmarking.org/result/1...MAXR-E6420PO59

      Without this option i have the following lines in dmesg:
      Code:
      [    0.780418] ACPI FADT declares the system doesn't support PCIe ASPM, so disable it
      [    3.430392] Unable to assume _OSC PCIe control. Disabling ASPM
      The first test doesn't seem to be valid (i may have touched the touchpad, don't know why it's so different from the other results). At first i thought i was impacted by the bug but repeating each test 3 times did not showed significant differences.

      Phoronix Test Suite v3.0.1
      System Information

      Hardware:
      Processor: Intel Core i5-2520M @ 2.50GHz (4 Cores), Motherboard: Dell 0K0DNP, Chipset: Intel 2nd Generation Core Family, Memory: 7680MB, Disk: 250GB Western Digital WDC WD2500BEVT-7, Graphics: Intel 2nd Generation Core Family IGP 256MB, Audio: IDT 92HD90BXX

      Software:
      OS: Ubuntu 11.04, Kernel: 2.6.39-3-generic (x86_64), Desktop: Unity 3.8.16, Display Server: X Server 1.10.1, Display Driver: intel 2.14.0, OpenGL: 2.1 Mesa 7.10.2, Compiler: GCC 4.5.2, File-System: ext4, Screen Resolution: 1600x900

      PS: is there a way to remove, rename and reorganize benchmarks on openbenchmarking.org ?

      Comment


      • #18
        Originally posted by danwood76 View Post
        Use powertop (properly) and see how many wakeups from what apps you get.
        Thanks, powertop is very useful. My kernel was missing CONFIG_USB_SUSPEND. Also this helped a bit:

        Code:
        Suggestion: Enable SATA ALPM link power management via:
          echo min_power > /sys/class/scsi_host/host0/link_power_management_policy
        Unfortunately Linux still uses 15% - 20% more power than Win 7.

        Linux AMD: 34W
        Wakeups-from-idle per second: 6.3 interval: 30.0s

        Comment


        • #19
          Still having issues despite grub.conf fix

          MSI Wind U100
          Intel Atom CPU N270 @ 1.60GHz
          Intel 945GME

          Comment


          • #20
            Hardware:
            Code:
            Processor: Intel Core 2 Duo T6500 @ 2.10GHz (2 Cores), Motherboard: Dell 0D176M, Chipset: Intel Mobile 4 MCH + ICH9M, Memory: 3584MB, Disk: 320GB Seagate ST9320320AS + 160GB Hitachi HTS64161, Graphics: Intel Mobile 4 IGP 256MV, Audio: IDT 92HD73C1X5
            Software:
            Code:
            OS: Ubuntu 11.04, Kernel: 2.6.38-8-generic (x86_64), Desktop: KDE 4.6.4, Display Server: X Server 1.10.1, Display Driver: intel 2.14.0, Compiler: GCC 4.5.2, File-System: ext4, Screen Resolution:1366x768
            and also

            Code:
            OS: Ubuntu 11.04, Kernel: 3.0.0-rc5-customdell1555 (x86_64), Desktop: KDE 4.6.4, Display Server: X Server 1.10.1, Display Driver: intel 2.14.0, Compiler: GCC 4.5.2, File-System: ext4, Screen Resolution:1366x768

            Comment


            • #21
              Dell Latitude E6410

              On my Dell Latitude E6410, forcing ASPM on results in ~5-10% power savings.

              Running pts/battery-power-usage-1.0.0:
              Battery Power Usage (Avg): 13748.75 (1-default) vs 12903.14 (2-pcie_aspm=force).

              http://openbenchmarking.org/result/1...IV-UBUNTUNAT12

              Phoronix Test Suite v3.0.1
              System Information

              Hardware:
              Processor: Intel Core i3 M 380 @ 2.53GHz (4 Cores), Motherboard: Dell 0667CC, Chipset: Intel Core, Memory: 7680MB, Disk: 250GB Western Digital WDC WD2500BEKT-7, Graphics: Intel Core IGP 256MB, Audio: IDT 92HD81B1C5

              Software:
              OS: Ubuntu 11.04, Kernel: 2.6.38-8-generic (x86_64), Desktop: Unity 3.8.16, Display Server: X Server 1.10.1, Display Driver: intel 2.14.0, Compiler: GCC 4.5.2, File-System: ext4, Screen Resolution: 1280x800

              Comment


              • #22
                I have a desktop, so the excess power consumption isn't too much of a problem -- but it's measurable. Core i7 2600K, Asus P8P67 Deluxe rev3. My CyberPower UPS measures power consumption in 8-watt increments, and ASPM reduces it by one [8-watt jump]. I'll dig up a Kill-a-watt and test it better later....

                Comment


                • #23
                  HP Pavilion dv6730ed
                  Intel Core 2 Duo
                  NVidia 8400M GS running open-source driver

                  There's something not OK in the hardware power management; there's no OS or program able to measure the power consumption.

                  BIOS updates can do wonders on HP laptops. Even on Windows there used to be a power and fan control related bug in the BIOS of many HP's with an Intel processor.

                  Now that I've upgraded the BIOS, the laptop doesn't feel as hot as it used to be and the battery life is 100% OK. Actually, Windows Vista consumes a lot of memory.

                  My mother's Acer Aspire 5530G, AMD Turion X2 64, ATI Radeon HD 3650, overheats after 30 minutes using Linux Mint 11. I haven't yet tested upgrading the BIOS.
                  Please people, try upgrading your BIOS and report here if it solved your problems.
                  Last edited by AlbertP; 07-02-2011, 06:21 AM.

                  Comment


                  • #24
                    My data

                    First, I do agree with "danwood76", Michael Larabel doesn't make things in the Community Spirit but for such a cause I am ready to generate him some revenue. I even disabled AdBlock for this site...
                    Anyway, I think he should post what he found to the Community instead of saying that I found something but I would not tell you for a time cause I care more to make money than help faster and more the community...

                    Btw, my data:

                    System Information

                    Hardware:
                    Processor: Intel Pentium Dual T3200 @ 2.00GHz (2 Cores), Motherboard: FUJITSU SIEMENS ESPRIMO Mobile V5505 v20, Chipset: Intel Mobile PM965/GM965/GL960 + ICH8M, Memory: 3072MB, Disk: 250GB Western Digital WDC WD2500BEVS-2, Graphics: Intel Mobile GM965/GL960 IGP 256MB, Audio: Conexant CX20549 (Venice)

                    Software:
                    OS: Ubuntu 10.10, Kernel: 2.6.35-30-generic (i686), Desktop: GNOME 2.32.0, Display Server: X Server 1.9.0, Display Driver: intel 2.12.0, Compiler: GCC 4.4.5, File-System: ext4, Screen Resolution: 1280x800

                    Later the "battery-power-usage" benchmark, cause I need to close all the programs...

                    Comment


                    • #25
                      My test results

                      On a dell xps 15, nvidia graphics off

                      http://openbenchmarking.org/result/1...GR-MIRKOROSS05

                      Comment


                      • #26
                        Originally posted by AlbertP View Post
                        HP Pavilion dv6730ed
                        Intel Core 2 Duo
                        NVidia 8400M GS running open-source driver

                        There's something not OK in the hardware power management; there's no OS or program able to measure the power consumption.
                        My HP has that problem also:
                        cat: /sys/class/power_supply/BAT0/current_now: No such device

                        Originally posted by AlbertP View Post
                        BIOS updates can do wonders on HP laptops.
                        A good idea. Unfortunately I already have the latest BIOS/UEFI.
                        I have upgraded it 2-3 times in the past.

                        Comment


                        • #27
                          At me Linux and Windows at least recognise the battery and show the %, and I can see the tension (volt) but not the power consumption. My other HP (nc6120 with a BIOS from 2005) does not suffer much from the regression.

                          Comment


                          • #28
                            Unstable when activating pcie_aspm=force

                            On my computer (HP Compaq 6730b) enable aspm into the kernel makes cold down the computer but makes it unstable when awakes from the suspend state (sometimes it crash just after the awakening - sometimes not).

                            System Information

                            Hardware:
                            Processor: Intel Core 2 Duo P8400 @ 2.26GHz (2 Cores), Motherboard: HP 30DD, Chipset: Intel Mobile 4 MCH + ICH9M, Memory: 2 x 2048 MB DDR2-800MHz, Disk: 160GB FUJITSU MHZ2160B, Graphics: Intel Mobile 4 IGP, Audio: Analog Devices AD1984A

                            Software:
                            OS: Arch, Kernel: 2.6.39-ARCH (x86_64), Desktop: GNOME Shell 3.0.2, Display Server: X Server 1.10.2, Display Driver: intel 2.15.0, OpenGL: 2.1 Mesa 7.10.3, Compiler: GCC 4.6.0, File-System: ext4, Screen Resolution: 1280x800

                            Comment


                            • #29
                              4.1 and macbook battery problem also

                              forums muylinux.com in a Spanish page on linux I found the case of a guy who quotes the following: I have installed Ubuntu 11.04 on a macbook 4.1 and I have that battery problem,
                              it seems that these computers do not have what which is a BIOS (I'm not sure a lot of computer science is not it ...), let any developer writing to compare, look for or report, it is an important detail, with which to discover the problem, which Kernel is safer

                              Comment


                              • #30
                                MacBooks indeed do not have an IBM-compatible BIOS like other computers, but the ACPI tables (the FADT is one of them) are still there so it's still affected by the power bug if Mac have used ASPM the Windows way.

                                Comment

                                Working...
                                X