[Bug 717454] New: Power consumption of 3.0.4 kernel is 10-20W higher than 2.6.37 on new i7 based SandyBridge processors
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c0 Summary: Power consumption of 3.0.4 kernel is 10-20W higher than 2.6.37 on new i7 based SandyBridge processors Classification: openSUSE Product: openSUSE 11.4 Version: Final Platform: Other OS/Version: Other Status: NEW Severity: Major Priority: P5 - None Component: Kernel AssignedTo: kernel-maintainers@forge.provo.novell.com ReportedBy: awafaa@opensuse.org QAContact: qa@suse.de Found By: --- Blocker: --- Created an attachment (id=450402) --> (http://bugzilla.novell.com/attachment.cgi?id=450402) hwinfo for X220 User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/535.1 (KHTML, like Gecko) Chrome/13.0.782.220 Safari/535.1 Running on Intel® Core™ i7-2620M power usage is through the roof when using the 3.0.4 kernel. On average it is 10-20W higher than the 2.6.37 kernel. My 9 cell battery lasts ~4hrs instead of ~9hrs. hwinfo attached. Reproducible: Always Steps to Reproduce: 1. Boot up 2. Watch battery drain Expected Results: Battery should last ~9hrs+ using ~11W -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c Andrew Wafaa <awafaa@opensuse.org> changed: What |Removed |Added ---------------------------------------------------------------------------- AssignedTo|kernel-maintainers@forge.pr |gregkh@suse.com |ovo.novell.com | Severity|Major |Critical -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c1 --- Comment #1 from Richard Brown <rdb@ccb.ac.uk> 2011-09-14 21:23:02 UTC --- Created an attachment (id=450846) --> (http://bugzilla.novell.com/attachment.cgi?id=450846) hwinfo of another Thinkpad X220 with identical symptoms Attaching a hwinfo output from another Thinkpad X220 with an i7 CPU that exhibits identical power consumption issues. eg 11W in kernel 2.6 and over 17W in kernel 3.1 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c2 Richard Brown <rdb@ccb.ac.uk> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rdb@ccb.ac.uk --- Comment #2 from Richard Brown <rdb@ccb.ac.uk> 2011-09-14 21:23:36 UTC --- correction: 17W in kernel 3.0 - have not tested with 3.1 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c3 Greg Kroah-Hartman <gregkh@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO InfoProvider| |jlee@suse.com --- Comment #3 from Greg Kroah-Hartman <gregkh@suse.com> 2011-09-16 09:56:56 UTC --- Joey, can you dig around in your lab and see if you have a Sandybridge machine to see if you can duplicate this? If not, let me know and I'll push Intel to get me a sample box to track this down and fix it. Joey, also feel free to add anyone else to this bug that you think can help out. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c4 --- Comment #4 from Richard Brown <rdb@ccb.ac.uk> 2011-09-16 12:20:07 UTC --- Is there a good simple way of monitoring power consumption in desktops/servers? Powertop does not show consumption unless a battery is discharging but I am interested in seeing if this bug occurs in non-laptop hardware -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c5 --- Comment #5 from Greg Kroah-Hartman <gregkh@suse.com> 2011-09-16 14:21:12 UTC --- (In reply to comment #4)
Is there a good simple way of monitoring power consumption in desktops/servers? Powertop does not show consumption unless a battery is discharging but I am interested in seeing if this bug occurs in non-laptop hardware
I use a "Save-a-watt" device that plugs in between the power code and the power plug and it shows the power being used by the system. But it's pretty "crude". In order to do real power testing, you need a fancy scope and a bunch of wires on the battery of a laptop to really determine exactly how much power is being drawn by the system as powertop lies a lot when you are trying to tune things down to the sub-watt level. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c6 --- Comment #6 from Richard Brown <rdb@ccb.ac.uk> 2011-09-16 14:43:13 UTC --- Problem also present (possibly ~2W worse) in kernel 3.1.0-rc5-2-desktop from oS:Factory on the same X220 hardware in my previous hwinfo -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c7 --- Comment #7 from Richard Brown <rdb@ccb.ac.uk> 2011-09-16 17:57:02 UTC --- Created an attachment (id=451468) --> (http://bugzilla.novell.com/attachment.cgi?id=451468) hwinfo of Gulftown desktop I've bought a "Save-a-watt" like device and tested another one of my machines, a desktop with a Gulftown CPU (i7 980x) These tests are showing the problem also occurs on this hardware. After a straight boot, not logging in, kernel 2.6.38 draws 134W according to my meter Booting the identical install into kernel 3.0 from Tumbleweed draws 140W, even with the video card uninitialised (no nvidia prop. driver) hwinfo of this Gulftown machine attached. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c8 Joey Lee <jlee@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- InfoProvider|jlee@suse.com |rdb@ccb.ac.uk --- Comment #8 from Joey Lee <jlee@suse.com> 2011-09-17 07:34:42 UTC --- I will find a sandybridge machine to test it after back to Taipei office. That will be better use 2-channel power metter or even NetDAQ to measurement the Watt value. 6 Watt is a huge power leak, we should check the machine on idle status, first. Richard and Andrew, could you please you powertop to check the CPU go to C-state? When the machine is idle, the CPU state most of the time in C4 state. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c9 Richard Brown <rdb@ccb.ac.uk> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|rdb@ccb.ac.uk | --- Comment #9 from Richard Brown <rdb@ccb.ac.uk> 2011-09-17 09:51:18 UTC --- On my X220 (Sandybridge CPU) with 2.6 (Stock oS 11.4 kernel) running powertop in runlevel 1, I see power consumption as previously reported (11W) with the CPU spending almost all of its time in C4 state (99-101% according to powertop) Changing to kernel 3.0 (Tumbleweed version) running powertop in rl 1, power consumption is as previously reported (17W). powertop reports that the CPU is spending just as much of its time in C4 state (99-101%) Looks like the power drain is coming from something other than the CPU, or powertop is lying. Will be doing these same tests on my Gulftown hardware later today when I'm back at home -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c10 --- Comment #10 from Richard Brown <rdb@ccb.ac.uk> 2011-09-17 19:58:27 UTC --- Situation is identical on my Gulftown (i7 980x) workstation CPU when idle is spending most of its time in C4 state (99-101% according to powertop) but power consumption is still 6-10W higher with kernel 3.0 compared to 2.6 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c11 --- Comment #11 from Richard Brown <rdb@ccb.ac.uk> 2011-09-17 20:14:31 UTC --- As CPU doesn't appear to be related to the issue I've been investigating other potential causes. Phoronix have lots of articles on recent Linux kernel power issues (eg: http://www.phoronix.com/scan.php?page=article&item=linux_2638_aspm&num=2 http://www.phoronix.com/scan.php?page=news_item&px=OTYwNA ) Based on the claims of the article, I have attempted booting up kernel 3.0 with pcie_aspm=force This has seen a small (~1W) improvement in 3.0 power consumption on both of my machines, which might be a good start but clearly not the root cause. HTH -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c12 --- Comment #12 from Andrew Wafaa <awafaa@opensuse.org> 2011-09-25 08:09:59 UTC --- Adding the following kernel parameter to the 3.0.4 kernel brings usage down from 22W to 11W: i915.i915_enable_rc6=1 This also seems to fix fan over utilisation/heat which may have been the cause of the additional power drain. For reference see https://bugzilla.redhat.com/show_bug.cgi?id=727579 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c13 --- Comment #13 from Joey Lee <jlee@suse.com> 2011-09-26 10:49:05 UTC --- I got a SandyBridge notebook in Taipei, now. Will also apply Richard and Andrew's suggestion then verify the power status by powertop. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c14 --- Comment #14 from Joey Lee <jlee@suse.com> 2011-09-27 08:10:57 UTC --- Just tested on a IronLake machine (found I was wrong after testing), but still put the powertop result here for compare with SandyBridge: Acer TravelMate 8572 openSUSE 11.4 2.6.37.1-1.2-desktop i586 kernel battery discharge brightness lowest level wifi off Desktop 3D effect enabled powertop 11.3 after 10 minues: 9.5 W long term: 11.6W Acer TravelMate 8572 openSUSE 11.4 3.0.4-43.3-desktop i586 kernel (openSUSE:Tumbleweed) battery discharge brightness lowest level wifi off Desktop 3D effect enabled powertop 11.3 after 10 minutes: 8.4 W long term: 12.2W Didn't have huge difference between 2.6.37 and 3.0.4 kernel, I am testing a SandyBridge machine, will attach result. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c15 --- Comment #15 from Joey Lee <jlee@suse.com> 2011-09-28 03:01:44 UTC --- I cann't reproduce issue by using a HP machine: HP openSUSE 11.4 2.6.37.1-1.2-desktop i586 kernel battery discharge brightness level one wifi off Desktop 3D effect enabled powertop 11.3 boot after 10 minues: 7.5 W long term: 10.8W HP openSUSE 11.4 3.0.4-43.3-desktop i586 kernel (openSUSE:Tumbleweed) battery discharge brightness lowest level wifi off Desktop 3D effect enabled powertop 11.3 boot after 10 minutes: 8.5 W long term: 10.3W AC Plug-in power monitor(removed battery): 12 W Didn't have big difference between 2.6.37.1-1 and 3.0.4 kernel. The HP machine's VGA is: 8086:0116 -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c16 --- Comment #16 from Joey Lee <jlee@suse.com> 2011-09-28 15:18:16 UTC --- I cann't reproduce the huge power leak on the sandybridge machine with 8086:0116 VGA. But, still add i915_enable_rc6=1 and pcie_aspm=force to testing: HP openSUSE 11.4 3.0.4-43.3-desktop i586 kernel (openSUSE:Tumbleweed) battery discharge brightness lowest level wifi off Desktop 3D effect enabled enabled i915.i915_enable_rc6=1 powertop 11.3 boot after 10 minutes: 6.6 W long term: 8.4W AC Plug-in power monitor(removed battery): 10 - 11 W HP openSUSE 11.4 3.0.4-43.3-desktop i586 kernel (openSUSE:Tumbleweed) battery discharge brightness lowest level wifi off Desktop 3D effect enabled enabled i915.i915_enable_rc6=1 pcie_aspm=force powertop 11.3 boot after 10 minutes: 6.7 W long term: 8.3 W AC Plug-in power monitor(removed battery): 9 - 10 W Compare with didn't add parameters: i915_enable_rc6=1 save 1.x W pcie_aspm=force save 0.x W On my machine, the above parameters useful for battery, but didn't have huge improve like Thinkpad X220 with i7. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c17 Jon Nelson <jnelson-suse@jamponi.net> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |jnelson-suse@jamponi.net --- Comment #17 from Jon Nelson <jnelson-suse@jamponi.net> 2011-10-12 19:09:33 UTC --- This may help. I have a ThinkPad T520i. I placed the following in /etc/modprobe.d/99-local.conf and rebuilt the initrd (between '====' lines): ============= blacklist firewire_ohci blacklist mei options i915 powersave=1 i915_enable_rc6=1 ============= (I don't use firewire). I've also sometimes used i915_enable_fbc=1 but I'm not sure that helps or not. Furthermore, in /etc/modprobe.d/50-sound.conf I have: ============= options snd-hda-intel enable_msi=1 model=thinkpad power_save=30 power_save_controller=1 ============= and that helps quite a bit, too. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c18 Greg Kroah-Hartman <gregkh@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO InfoProvider| |awafaa@opensuse.org --- Comment #18 from Greg Kroah-Hartman <gregkh@suse.com> 2011-10-23 08:50:52 UTC --- Andrew and others who can duplicate this, please try the above i915_enable_rc6=1 option for the i915 driver, that's what the intel driver authors just suggested to me. Can you let me know if that works or not? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c19 Bruno Friedmann <bruno@ioda-net.ch> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |bruno@ioda-net.ch --- Comment #19 from Bruno Friedmann <bruno@ioda-net.ch> 2011-10-23 09:19:58 UTC --- Just a guess here Did someone check if the content of /etc/modprobe.d/50-thinkpad_acpi.conf is correct for the x220 model ? which come with acpid rpm package # # IBM/Lenovo ThinkPad ACPI driver options # # A detailed description of the parameters for the ThinkPad ACPI driver can be found # in /usr/src/linux/Documentation/laptops/thinkpad-acpi.txt which is part of the package # 'kernel-source'. # # If you encounter problems with the hotkey mask please file a bug on # http://bugzilla.novell.com/ # options thinkpad_acpi experimental=1 hotkey=0xffffff -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c20 Richard Brown <RBrownCCB@opensuse.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|awafaa@opensuse.org | --- Comment #20 from Richard Brown <RBrownCCB@opensuse.org> 2011-10-23 12:37:27 UTC --- Hi Greg, I've been doing some experiments with those i915 module params - all of the below numbers are from Factory with kernel 3.1.0-rc9-1 X220 powertop recorded idle power consumption with full screen brightness, wifi enabled and no i915 parameters 18-20W General use (mouse movements, browser opening, etc) sees consumption spike upto 35W+ X220 powertop recorded idle power consumption with full screen brightness, wifi enabled and i915.i915_enable_rc6=1 12.8-13.1W General use (mouse movements, browser opening, etc) sees consumption spike upto 20W for brief periods X220 powertop recorded idle power consumption with full screen brightness, wifi enabled and i915.i915_enable_rc6=1 i915.powersave=1 i915.i915_enable_fbc=1 i915.lvds_downclock=1 12.5-13.0W General use (mouse movements, browser opening, etc) sees consumption spike upto 16W for brief periods So, I would say that the suggestion that i915_enable_rc6 is the key change that dramatically improves battery life looks like, with some other parameters also providing a small nudge with no side effects in my testing, yet. Removing the NEEDINFO as hopefully this answers GregKH's inforequest, still going to check out Bruno's idea though -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c21 Greg Kroah-Hartman <gregkh@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO InfoProvider| |awafaa@opensuse.org --- Comment #21 from Greg Kroah-Hartman <gregkh@suse.com> 2011-10-23 12:58:22 UTC --- Thanks, yes, that answers my question. So, with those options enabled, we are back to the old power consumption levels, right? So there's nothing else we can do here. We can't enable those options "by default", otherwise the kernel would have done so, as some hardware can't handle them, while others, like yours, needs them due to various integration issues (BIOS/motherboard/etc.) I still would like to have Andrew try this out on his platform, for verification that this does resolve the issue he sees. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c22 --- Comment #22 from Andrew Wafaa <awafaa@opensuse.org> 2011-10-23 13:22:03 UTC --- (In reply to comment #21)
Thanks, yes, that answers my question.
So, with those options enabled, we are back to the old power consumption levels, right?
So there's nothing else we can do here. We can't enable those options "by default", otherwise the kernel would have done so, as some hardware can't handle them, while others, like yours, needs them due to various integration issues (BIOS/motherboard/etc.)
I still would like to have Andrew try this out on his platform, for verification that this does resolve the issue he sees.
I'll verify this shortly, hopefully by tomorrow. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c23 Andrew Wafaa <awafaa@opensuse.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|awafaa@opensuse.org | --- Comment #23 from Andrew Wafaa <awafaa@opensuse.org> 2011-10-24 08:51:56 UTC --- I've managed to run the same tests as Richard and my results are as follows based on my system which is running 11.4+GNOME3+Kernel 3.0.7: X220 powertop recorded idle power consumption with full screen brightness, wifi enabled and no i915 parameters ~21.8W General use (mouse movements, browser opening, etc) sees consumption spike upto ~40.1W X220 powertop recorded idle power consumption with full screen brightness, wifi enabled and i915.i915_enable_rc6=1 ~14.9W General use (mouse movements, browser opening, etc) sees consumption spike upto ~19.7W X220 powertop recorded idle power consumption with full screen brightness, wifi enabled and i915.i915_enable_rc6=1 i915.powersave=1 i915.i915_enable_fbc=1 i915.lvds_downclock=1 ~13.8W General use (mouse movements, browser opening, etc) sees consumption spike upto ~18.8W -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c24 Greg Kroah-Hartman <gregkh@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |NEEDINFO InfoProvider| |awafaa@opensuse.org --- Comment #24 from Greg Kroah-Hartman <gregkh@suse.com> 2011-10-24 09:54:51 UTC --- So, is the 13.8W the same as the older kernel version? I don't think you ever posted the powertop numbers for the .37 kernel, right? And if so, I'm guessing we can close this bug now? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c25 Andrew Wafaa <awafaa@opensuse.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEEDINFO |NEW InfoProvider|awafaa@opensuse.org | --- Comment #25 from Andrew Wafaa <awafaa@opensuse.org> 2011-10-24 10:22:35 UTC --- Whoops - that was a rather key omission on my part, sorry. Booting into the 2.6.37 kernel I have the following: X220 powertop recorded idle power consumption with full screen brightness, wifi enabled and no i915 parameters ~13.2W General use (mouse movements, browser opening, etc) sees consumption spike upto ~18.0W These are with no kernel parameter passed at all and as you can see they are still better than with an optimised 3.0.7 kernel :-( Out of interest, would there be much difference in power consumption between -desktop and -default? -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c26 --- Comment #26 from Greg Kroah-Hartman <gregkh@suse.com> 2011-10-25 01:29:23 UTC --- Thanks for the numbers. So, overall, with the module options, we are back at the original power usage, which is a good thing to know. Now, how can we enable these options "by default" without having to set them by hand is going to be the trick. I'll track down the upstream kernel developers and bug them about this issue today. Unfortunately, I'm worried that the response is going to be, "we can't safely always enable them, due to BIOS bugs, which is why I think they are not set this way to start with. Then we run the problem of trying to tell vendors that their BIOSes are broken, which, while always true, and it makes kernel developers feel smug, isn't always the most helpful way to handle this.. And no, there should not be any power differences between the -default and -desktop kernels that I know of. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=717454 https://bugzilla.novell.com/show_bug.cgi?id=717454#c27 Greg Kroah-Hartman <gregkh@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #27 from Greg Kroah-Hartman <gregkh@suse.com> 2011-12-05 20:05:53 UTC --- There are plans in the future to hopefully integrate some automatic detection here, but it's not in the 12.1 timeframe as that's already shipping sorry. I'll close this for now, as the base problem is solved. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com