Menu
▾
▴
mactel-linux-devel
|
From: Nicolas B. <ni...@bo...> - 2007-04-18 15:46:22
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, (switched to mactel-devel list, more appropriate I think) Sven Anders wrote: > Hello! > > Some data of my machine (for reference): > > # dmidecode | egrep '(Version|Release)' > Version: MBP22.88Z.00A5.B01.0611031551 > Release Date: 11/03/06 > Version: 1.0 > Version: PVT > Version: Mac-F42187C8 > Version: Intel(R) Core(TM)2 CPU > Version: Intel(R) Core(TM)2 CPU > > > This is what I found out: > ------------------------- Wow, impressive report, that's really unfortunate and frustrating it doesn't really helps us to improve power management on Linux... > The following hopes I have: > > 1) The auto USB suspend framework will solve the too-many interrupts problems. Have you tried this kernel option? CONFIG_USB_SUSPEND: USB selective suspend/resume and wakeup (EXPERIMENTAL) Looks interesting... You might be able to disable some devices and see if it improves power consumption... > 2) The dyn-ticks patches will allow the processor to sleep longer. At least in the current version, it doesn't help at all unfortunately. Best regards, Nicolas -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.3 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGJj0w01ajQnpJXgERAslGAJ9JrDniCCei9mYWg8LmiFCsRRQVCgCfWO/N tl7r5rUv6Nxxmnm4WWbhgYY= =btRZ -----END PGP SIGNATURE----- |
|
From: Kai W. <kai...@gl...> - 2007-04-20 15:35:42
|
* Sven Anders <an...@an...>: Great work Sven! Respect! BTW, did anyone had a look at the Apple PMU? Maybe that is a place to tune the power consumption... Kai -- * http://www.glorybox.de/ PGP 1024D/594D4132 B693 5073 013F 7F56 5DCC D9C2 E6B5 448C 594D 4132 |
|
From: Nicolas B. <ni...@bo...> - 2007-04-20 16:52:28
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Kai Weber wrote: > * Sven Anders <an...@an...>: > > Great work Sven! Respect! > > BTW, did anyone had a look at the Apple PMU? Maybe that is a place to > tune the power consumption... On the powerbook, maybe... PMU is a device on powerpc-based macs, SMC replaces it on intel-based macs. You can check this by running "kextstat" on OS X, and check if any kext related to PMU is loaded. I don't even think the PMU module has any x86 assembly section. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.3 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFGKO+t01ajQnpJXgERAkZMAJ9KEWl8wmtPogRMbZ3wMGS7p7mUGgCeJJG5 GekpcXho15uV9bY1B/LX3/4= =akGC -----END PGP SIGNATURE----- |
|
From: Sven A. <an...@an...> - 2007-04-24 06:10:32
Attachments:
signature.asc
anders.vcf
|
Hello!
I thought about the power consumption. Maybe we can track the problem down
some other way...
1. Is the output of '/proc/acpi/battery/BAT0/state' reliable? If it is,
we can use it for testing, otherwise we have to test how long the battery
really lasts.
2. As far as I remember, the MacBook is very similar to the MacBook Pro in
the hardware it's using. The main difference is the GPU.
How long does the battery of the MacBook under Linux last in comparison to
MacOS X?
If it's a small difference, so our problem is the power consumption of
the ATI GPU (assuming the Intel-GPU draws as less power under Linux as it
does under MacOSX). On the MacBook we have the advantage, that it's an
open-source driver.
If the power drain is as bad as on the MacBook Pro (and we checked the
Intel-GPU driver, that is's using all power-saving states available),
the problem lies somewhere else...
To check these, we need some output of the MacBook's /proc/acpi values or
somebody who helps us to prove these theory...
Any comments to these thoughts? Any errors in it?
Best regards
Sven
--
Sven Anders <an...@an...> () Ascii Ribbon Campaign
/\ Support plain text e-mail
ANDURAS service solutions AG
Innstraße 71 - 94036 Passau - Germany
Web: www.anduras.de - Tel: +49 (0)851-4 90 50-0 - Fax: +49 (0)851-4 90 50-55
Rechtsform: Aktiengesellschaft - Sitz: Passau - Amtsgericht Passau HRB 6032
Mitglieder des Vorstands: Sven Anders, Marcus Junker
Vorsitzender des Aufsichtsrats: Dipl. Kfm. Thomas Träger
|
|
From: Sheer El-S. <sh...@gm...> - 2007-04-24 08:06:40
|
Hi Sven, I asked about the MacBook issue in a previous post somewhere. I've also become convinced that the problem is the GPU (there was an article I mentioned on Phoronix where someone compared the temperature of the GPU of an ATI card in Linux and windows and found it much higher in the former). Regarding the MacBook, here's a webpage claiming a 3 hour battery life under linux: http://desrt.mcmaster.ca/macbook.xhtml http://bbs.archlinux.org/viewtopic.php?id=3D22960 The second link has a bit of discussion about battery life under Linux on the MB. I don't know what the OS X battery life is but I'm guessing its at least 4 hours (people there say 3-4 hrs). I agree that looking into this is the right approach. If the GPU is the problem there's very little we can do about it but wait for ATI so we might as well find out if this is the case. Hopefully somebody from the list can post some specific MB stats. It would also be good to see if its possible to get the MB battery life up to its OS X value using USB autosuspend and dynticks. cheers, Sheer On 4/24/07, Sven Anders <an...@an...> wrote: > Hello! > > I thought about the power consumption. Maybe we can track the problem dow= n > some other way... > > 1. Is the output of '/proc/acpi/battery/BAT0/state' reliable? If it is, > we can use it for testing, otherwise we have to test how long the batt= ery > really lasts. > > 2. As far as I remember, the MacBook is very similar to the MacBook Pro i= n > the hardware it's using. The main difference is the GPU. > How long does the battery of the MacBook under Linux last in compariso= n > to > MacOS X? > If it's a small difference, so our problem is the power consumption of > the ATI GPU (assuming the Intel-GPU draws as less power under Linux as= it > does under MacOSX). On the MacBook we have the advantage, that it's an > open-source driver. > If the power drain is as bad as on the MacBook Pro (and we checked the > Intel-GPU driver, that is's using all power-saving states available), > the problem lies somewhere else... > > To check these, we need some output of the MacBook's /proc/acpi values or > somebody who helps us to prove these theory... > > Any comments to these thoughts? Any errors in it? > > > Best regards > Sven > > -- > Sven Anders <an...@an...> () Ascii Ribbon Campaign > /\ Support plain text > e-mail > ANDURAS service solutions AG > Innstra=DFe 71 - 94036 Passau - Germany > Web: www.anduras.de - Tel: +49 (0)851-4 90 50-0 - Fax: +49 (0)851-4 90 > 50-55 > > Rechtsform: Aktiengesellschaft - Sitz: Passau - Amtsgericht Passau HRB 60= 32 > Mitglieder des Vorstands: Sven Anders, Marcus Junker > Vorsitzender des Aufsichtsrats: Dipl. Kfm. Thomas Tr=E4ger > |
|
From: Sheer El-S. <sh...@gm...> - 2007-04-24 20:46:02
|
Hi Davide, Thanks for the info. I'm hope you don't mind that I'm CCing the list but I think others will appreciate this info. How did you measure the power drain in OS X (i.e. 1200 mA)? Cheers, Sheer On 4/24/07, Davide Bertola <da...@da...> wrote: > At fist macbooks do not have ATI card. It's Intel (macbook non-pro). > > I use a macbook and i get ~3 hours on linux and ~4.5 hours on OSX. > It's a macbook core duo (not core 2 duo) and the battery is new. > > Linux measures capacity in Watts (W), osx in Amperes * 10^(-3) (mA). > Under linux I see a maximum capacity of ~52000 mW, and using OSX it's ~52= 00 > mA. > Since mW -> mA is a simple conversion because we only have to /10 > (should be 11 'cause battery works at 11 volts average but let's say > it's 10) we can compare the idle power consumpition. > > On my system osx idles at ~1200 mA, linux does 16 W -> 1600 mA. > It's also hotter when using linux. Why ? > I guess: > - Lack of C4 state support. > - No way to turn off bluetooth (always looking for devices). > - maybe GPU Power Management support is not so efficent. > - don't know > > Anyway, I think that after a battery calibration, it's ok to compare > Watts and Amperes outputs of Linux and OSX. > > On 4/24/07, Sheer El-Showk <sh...@gm...> wrote: > > Hi Sven, > > > > I asked about the MacBook issue in a previous post somewhere. I've > > also become convinced that the problem is the GPU (there was an > > article I mentioned on Phoronix where someone compared the temperature > > of the GPU of an ATI card in Linux and windows and found it much > > higher in the former). Regarding the MacBook, here's a webpage > > claiming a 3 hour battery life under linux: > > > > http://desrt.mcmaster.ca/macbook.xhtml > > http://bbs.archlinux.org/viewtopic.php?id=3D22960 > > > > The second link has a bit of discussion about battery life under Linux > > on the MB. I don't know what the OS X battery life is but I'm > > guessing its at least 4 hours (people there say 3-4 hrs). > > > > I agree that looking into this is the right approach. If the GPU is > > the problem there's very little we can do about it but wait for ATI so > > we might as well find out if this is the case. Hopefully somebody > > from the list can post some specific MB stats. It would also be good > > to see if its possible to get the MB battery life up to its OS X value > > using USB autosuspend and dynticks. > > > > cheers, > > Sheer > > > > On 4/24/07, Sven Anders <an...@an...> wrote: > > > Hello! > > > > > > I thought about the power consumption. Maybe we can track the problem > down > > > some other way... > > > > > > 1. Is the output of '/proc/acpi/battery/BAT0/state' reliable? If it i= s, > > > we can use it for testing, otherwise we have to test how long the > battery > > > really lasts. > > > > > > 2. As far as I remember, the MacBook is very similar to the MacBook P= ro > in > > > the hardware it's using. The main difference is the GPU. > > > How long does the battery of the MacBook under Linux last in > comparison > > > to > > > MacOS X? > > > If it's a small difference, so our problem is the power consumptio= n > of > > > the ATI GPU (assuming the Intel-GPU draws as less power under Linu= x > as it > > > does under MacOSX). On the MacBook we have the advantage, that it'= s > an > > > open-source driver. > > > If the power drain is as bad as on the MacBook Pro (and we checked > the > > > Intel-GPU driver, that is's using all power-saving states > available), > > > the problem lies somewhere else... > > > > > > To check these, we need some output of the MacBook's /proc/acpi value= s > or > > > somebody who helps us to prove these theory... > > > > > > Any comments to these thoughts? Any errors in it? > > > > > > > > > Best regards > > > Sven > > > > > > -- > > > Sven Anders <an...@an...> () Ascii Ribbon > Campaign > > > /\ Support plain tex= t > > > e-mail > > > ANDURAS service solutions AG > > > Innstra=DFe 71 - 94036 Passau - Germany > > > Web: www.anduras.de - Tel: +49 (0)851-4 90 50-0 - Fax: +49 (0)851-4 = 90 > > > 50-55 > > > > > > Rechtsform: Aktiengesellschaft - Sitz: Passau - Amtsgericht Passau HR= B > 6032 > > > Mitglieder des Vorstands: Sven Anders, Marcus Junker > > > Vorsitzender des Aufsichtsrats: Dipl. Kfm. Thomas Tr=E4ger > > > > > > > -----------------------------------------------------------------------= -- > > This SF.net email is sponsored by DB2 Express > > Download DB2 Express C - the FREE version of DB2 express and take > > control of your XML. No limits. Just data. Click to get it now. > > http://sourceforge.net/powerbar/db2/ > > _______________________________________________ > > Mactel-linux-devel mailing list > > Mac...@li... > > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > > > |
|
From: Sven A. <an...@an...> - 2007-04-24 21:05:58
Attachments:
signature.asc
anders.vcf
|
Sheer El-Showk schrieb:
>> On my system osx idles at ~1200 mA, linux does 16 W -> 1600 mA.
My MacBook Pro idles at 32W almost constantly.
>> It's also hotter when using linux. Why ?
>> I guess:
>> - Lack of C4 state support.
I think Linux is using C4(E) state too, but it does not stay as long
as MacOSX in it.
>> - No way to turn off bluetooth (always looking for devices).
Just try to unload the module... Maybe this is enough.
>> - maybe GPU Power Management support is not so efficent.
Maybe...
>> Anyway, I think that after a battery calibration, it's ok to compare
>> Watts and Amperes outputs of Linux and OSX.
Regards
Sven
--
Sven Anders <an...@an...> () Ascii Ribbon Campaign
/\ Support plain text e-mail
ANDURAS service solutions AG
Innstraße 71 - 94036 Passau - Germany
Web: www.anduras.de - Tel: +49 (0)851-4 90 50-0 - Fax: +49 (0)851-4 90 50-55
Rechtsform: Aktiengesellschaft - Sitz: Passau - Amtsgericht Passau HRB 6032
Mitglieder des Vorstands: Sven Anders, Marcus Junker
Vorsitzender des Aufsichtsrats: Dipl. Kfm. Thomas Träger
|
|
From: Thomas M. <tho...@we...> - 2007-04-24 22:10:43
|
Sven Anders schrieb: > Sheer El-Showk schrieb: > > >>> On my system osx idles at ~1200 mA, linux does 16 W -> 1600 mA. >>> > > My MacBook Pro idles at 32W almost constantly. > cat /proc/acpi/battery/BAT0/state present: yes capacity state: ok charging state: discharging present rate: 26189 mW -> without wlan and without appletouch, using an external mouse... |
|
From: Michel <mi...@tu...> - 2007-04-24 08:32:13
|
On Tue, 2007-04-24 at 08:10 +0200, Sven Anders wrote: > > 2. As far as I remember, the MacBook is very similar to the MacBook Pro in > the hardware it's using. The main difference is the GPU. > How long does the battery of the MacBook under Linux last in comparison to > MacOS X? > If it's a small difference, so our problem is the power consumption of > the ATI GPU (assuming the Intel-GPU draws as less power under Linux as it > does under MacOSX). On the MacBook we have the advantage, that it's an > open-source driver. > If the power drain is as bad as on the MacBook Pro (and we checked the > Intel-GPU driver, that is's using all power-saving states available), > the problem lies somewhere else... AFAICT the xf86-video-intel driver only has rudimentary power management code, mostly powering down unused output components. However, the integrated graphics core probably has significantly lower maximum and average power consumption than discrete parts to begin with. Hope this helps, -- Earthling Michel Dänzer | http://tungstengraphics.com Libre software enthusiast | Debian, X and DRI developer |
|
From: Davide B. <da...@da...> - 2007-04-25 05:22:23
|
you can see the power drain using System Profiiler.app under section 'power= ' On 4/24/07, Sheer El-Showk <sh...@gm...> wrote: > Hi Davide, > > Thanks for the info. I'm hope you don't mind that I'm CCing the list > but I think others will appreciate this info. > > How did you measure the power drain in OS X (i.e. 1200 mA)? > > Cheers, > Sheer > > > On 4/24/07, Davide Bertola <da...@da...> wrote: > > At fist macbooks do not have ATI card. It's Intel (macbook non-pro). > > > > I use a macbook and i get ~3 hours on linux and ~4.5 hours on OSX. > > It's a macbook core duo (not core 2 duo) and the battery is new. > > > > Linux measures capacity in Watts (W), osx in Amperes * 10^(-3) (mA). > > Under linux I see a maximum capacity of ~52000 mW, and using OSX it's ~= 5200 > > mA. > > Since mW -> mA is a simple conversion because we only have to /10 > > (should be 11 'cause battery works at 11 volts average but let's say > > it's 10) we can compare the idle power consumpition. > > > > On my system osx idles at ~1200 mA, linux does 16 W -> 1600 mA. > > It's also hotter when using linux. Why ? > > I guess: > > - Lack of C4 state support. > > - No way to turn off bluetooth (always looking for devices). > > - maybe GPU Power Management support is not so efficent. > > - don't know > > > > Anyway, I think that after a battery calibration, it's ok to compare > > Watts and Amperes outputs of Linux and OSX. > > > > On 4/24/07, Sheer El-Showk <sh...@gm...> wrote: > > > Hi Sven, > > > > > > I asked about the MacBook issue in a previous post somewhere. I've > > > also become convinced that the problem is the GPU (there was an > > > article I mentioned on Phoronix where someone compared the temperatur= e > > > of the GPU of an ATI card in Linux and windows and found it much > > > higher in the former). Regarding the MacBook, here's a webpage > > > claiming a 3 hour battery life under linux: > > > > > > http://desrt.mcmaster.ca/macbook.xhtml > > > http://bbs.archlinux.org/viewtopic.php?id=3D22960 > > > > > > The second link has a bit of discussion about battery life under Linu= x > > > on the MB. I don't know what the OS X battery life is but I'm > > > guessing its at least 4 hours (people there say 3-4 hrs). > > > > > > I agree that looking into this is the right approach. If the GPU is > > > the problem there's very little we can do about it but wait for ATI s= o > > > we might as well find out if this is the case. Hopefully somebody > > > from the list can post some specific MB stats. It would also be good > > > to see if its possible to get the MB battery life up to its OS X valu= e > > > using USB autosuspend and dynticks. > > > > > > cheers, > > > Sheer > > > > > > On 4/24/07, Sven Anders <an...@an...> wrote: > > > > Hello! > > > > > > > > I thought about the power consumption. Maybe we can track the probl= em > > down > > > > some other way... > > > > > > > > 1. Is the output of '/proc/acpi/battery/BAT0/state' reliable? If it= is, > > > > we can use it for testing, otherwise we have to test how long th= e > > battery > > > > really lasts. > > > > > > > > 2. As far as I remember, the MacBook is very similar to the MacBook= Pro > > in > > > > the hardware it's using. The main difference is the GPU. > > > > How long does the battery of the MacBook under Linux last in > > comparison > > > > to > > > > MacOS X? > > > > If it's a small difference, so our problem is the power consumpt= ion > > of > > > > the ATI GPU (assuming the Intel-GPU draws as less power under Li= nux > > as it > > > > does under MacOSX). On the MacBook we have the advantage, that i= t's > > an > > > > open-source driver. > > > > If the power drain is as bad as on the MacBook Pro (and we check= ed > > the > > > > Intel-GPU driver, that is's using all power-saving states > > available), > > > > the problem lies somewhere else... > > > > > > > > To check these, we need some output of the MacBook's /proc/acpi val= ues > > or > > > > somebody who helps us to prove these theory... > > > > > > > > Any comments to these thoughts? Any errors in it? > > > > > > > > > > > > Best regards > > > > Sven > > > > > > > > -- > > > > Sven Anders <an...@an...> () Ascii Ribbon > > Campaign > > > > /\ Support plain t= ext > > > > e-mail > > > > ANDURAS service solutions AG > > > > Innstra=DFe 71 - 94036 Passau - Germany > > > > Web: www.anduras.de - Tel: +49 (0)851-4 90 50-0 - Fax: +49 (0)851-= 4 90 > > > > 50-55 > > > > > > > > Rechtsform: Aktiengesellschaft - Sitz: Passau - Amtsgericht Passau = HRB > > 6032 > > > > Mitglieder des Vorstands: Sven Anders, Marcus Junker > > > > Vorsitzender des Aufsichtsrats: Dipl. Kfm. Thomas Tr=E4ger > > > > > > > > > > ---------------------------------------------------------------------= ---- > > > This SF.net email is sponsored by DB2 Express > > > Download DB2 Express C - the FREE version of DB2 express and take > > > control of your XML. No limits. Just data. Click to get it now. > > > http://sourceforge.net/powerbar/db2/ > > > _______________________________________________ > > > Mactel-linux-devel mailing list > > > Mac...@li... > > > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > > > > > > |
|
From: Davide B. <da...@da...> - 2007-04-25 05:22:38
|
On 4/24/07, Sven Anders <an...@an...> wrote: > Sheer El-Showk schrieb: > > >> On my system osx idles at ~1200 mA, linux does 16 W -> 1600 mA. > > My MacBook Pro idles at 32W almost constantly. 32W means you get almost 1,5 hours. That's ridicolus. Maybe your scaling governor is not working, or battery is not calibrated.. Then you have a macbook pro so you may have that problem with ati drivers. > > >> It's also hotter when using linux. Why ? > >> I guess: > >> - Lack of C4 state support. > > I think Linux is using C4(E) state too, but it does not stay as long > as MacOSX in it. > You think, but it's not true. linux sees onlin C1 C2 and C3 state. On some macbooks pro only C1 and C2. You can verify that looking somewhere in /proc or /sys I don't remember now. > >> - No way to turn off bluetooth (always looking for devices). > > Just try to unload the module... Maybe this is enough. > It's not. If I unload the module for bluetooth and the userspace bluetooth daemon.. the bluetooth act in some sort of emulation. My bluetooth mouse keeps working but it's seen as an USB mouse. It's cool, but drains power. Another thing I've seen is that linux is much faster pairing devices when they get in range, so i guess that bluetooth under linux polls for devices more often. (draining more power) > >> - maybe GPU Power Management support is not so efficent. > > Maybe... > > >> Anyway, I think that after a battery calibration, it's ok to compare > >> Watts and Amperes outputs of Linux and OSX. > > Regards > Sven > > -- > Sven Anders <an...@an...> () Ascii Ribbon Campaign > /\ Support plain text e-= mail > ANDURAS service solutions AG > Innstra=DFe 71 - 94036 Passau - Germany > Web: www.anduras.de - Tel: +49 (0)851-4 90 50-0 - Fax: +49 (0)851-4 90 5= 0-55 > > Rechtsform: Aktiengesellschaft - Sitz: Passau - Amtsgericht Passau HRB 60= 32 > Mitglieder des Vorstands: Sven Anders, Marcus Junker > Vorsitzender des Aufsichtsrats: Dipl. Kfm. Thomas Tr=E4ger > > |
|
From: Soeren S. <mac...@nn...> - 2007-04-25 07:12:02
|
On Tue, 2007-04-24 at 08:10 +0200, Sven Anders wrote: > Hello! > > I thought about the power consumption. Maybe we can track the problem down > some other way... > > 1. Is the output of '/proc/acpi/battery/BAT0/state' reliable? If it is, > we can use it for testing, otherwise we have to test how long the battery > really lasts. well at least the predicted time matched very well the time when the mbp went out of batteries and as that time is based on powerconsumption estimates in /proc/*/BAT0/state I would tend to say it is OK. > 2. As far as I remember, the MacBook is very similar to the MacBook Pro in > the hardware it's using. The main difference is the GPU. > How long does the battery of the MacBook under Linux last in comparison to > MacOS X? > If it's a small difference, so our problem is the power consumption of > the ATI GPU (assuming the Intel-GPU draws as less power under Linux as it > does under MacOSX). On the MacBook we have the advantage, that it's an > open-source driver. > If the power drain is as bad as on the MacBook Pro (and we checked the > Intel-GPU driver, that is's using all power-saving states available), > the problem lies somewhere else... > > To check these, we need some output of the MacBook's /proc/acpi values or > somebody who helps us to prove these theory... well I am using a mbp c1d and it eats about 27W with everything on + full brightness. as you know powerplay is supported on this book, aticonfig --set-powerstate=1 (low power mode) gives me 1-2W less boiling down to 25-26W or max. 10 extra minutes (*), if you further set display-brightness to the lowest level it goes down to 21-22W and backlight off still has 20W. In OSX I get 23.8W (=2h 18min) with everything on and 18.5W (=2h 58min) with everything off+low brightness so we are looking for reasons (I don't expect it to be a single one) that explain a gap of ~4W ... which seems a lot considering that this is essentially what the display at full brightness eats up but still may not be as big as we expected. Soeren (*): this assumes that your battery can be charged to 55Wh mine after half a year of intensive usage is already down to 35Wh |
|
From: Till S. <str...@sl...> - 2007-04-25 09:52:01
|
/I can confirm your report (mbp, c1d here) - I also found similar results (~17-18W) under osx. On linux I get down to 20-21W (low brighness, disk at rest, CPU freq at 1GHz, ati powerstate=1) but I found that unloading the 'uvcvideo' driver (isight) seems to save almost another ~1W or so - I'm at 19-20W right now. -- Till PS I still have the wireless up; shutting it down probably saves also some power. / Soeren Sonnenburg wrote: > On Tue, 2007-04-24 at 08:10 +0200, Sven Anders wrote: > >> Hello! >> >> I thought about the power consumption. Maybe we can track the problem down >> some other way... >> >> 1. Is the output of '/proc/acpi/battery/BAT0/state' reliable? If it is, >> we can use it for testing, otherwise we have to test how long the battery >> really lasts. >> > > well at least the predicted time matched very well the time when the mbp > went out of batteries and as that time is based on powerconsumption > estimates in /proc/*/BAT0/state I would tend to say it is OK. > > >> 2. As far as I remember, the MacBook is very similar to the MacBook Pro in >> the hardware it's using. The main difference is the GPU. >> How long does the battery of the MacBook under Linux last in comparison to >> MacOS X? >> If it's a small difference, so our problem is the power consumption of >> the ATI GPU (assuming the Intel-GPU draws as less power under Linux as it >> does under MacOSX). On the MacBook we have the advantage, that it's an >> open-source driver. >> If the power drain is as bad as on the MacBook Pro (and we checked the >> Intel-GPU driver, that is's using all power-saving states available), >> the problem lies somewhere else... >> >> To check these, we need some output of the MacBook's /proc/acpi values or >> somebody who helps us to prove these theory... >> > > well I am using a mbp c1d and it eats about 27W with everything on + > full brightness. as you know powerplay is supported on this book, > aticonfig --set-powerstate=1 (low power mode) gives me 1-2W less boiling > down to 25-26W or max. 10 extra minutes (*), if you further set > display-brightness to the lowest level it goes down to 21-22W and > backlight off still has 20W. > > In OSX I get 23.8W (=2h 18min) with everything on and 18.5W (=2h 58min) > with everything off+low brightness > > so we are looking for reasons (I don't expect it to be a single one) > that explain a gap of ~4W ... which seems a lot considering that this is > essentially what the display at full brightness eats up but still may > not be as big as we expected. > > Soeren > > (*): this assumes that your battery can be charged to 55Wh mine after > half a year of intensive usage is already down to 35Wh > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Mactel-linux-devel mailing list > Mac...@li... > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > |
|
From: Davide B. <da...@da...> - 2007-04-25 17:27:48
|
http://www.flickr.com/photos/dade/472608054/ see description :) On 4/25/07, Till Straumann <str...@sl...> wrote: > > /I can confirm your report (mbp, c1d here) - I also > found similar results (~17-18W) under osx. On linux > I get down to 20-21W (low brighness, disk at rest, CPU freq > at 1GHz, ati powerstate=1) but I found that unloading > the 'uvcvideo' driver (isight) seems to save almost > another ~1W or so - I'm at 19-20W right now. > > -- Till > > PS I still have the wireless up; shutting it down probably > saves also some power. > / > Soeren Sonnenburg wrote: > > On Tue, 2007-04-24 at 08:10 +0200, Sven Anders wrote: > > > >> Hello! > >> > >> I thought about the power consumption. Maybe we can track the problem > down > >> some other way... > >> > >> 1. Is the output of '/proc/acpi/battery/BAT0/state' reliable? If it is, > >> we can use it for testing, otherwise we have to test how long the > battery > >> really lasts. > >> > > > > well at least the predicted time matched very well the time when the mbp > > went out of batteries and as that time is based on powerconsumption > > estimates in /proc/*/BAT0/state I would tend to say it is OK. > > > > > >> 2. As far as I remember, the MacBook is very similar to the MacBook Pro > in > >> the hardware it's using. The main difference is the GPU. > >> How long does the battery of the MacBook under Linux last in > comparison to > >> MacOS X? > >> If it's a small difference, so our problem is the power consumption > of > >> the ATI GPU (assuming the Intel-GPU draws as less power under Linux > as it > >> does under MacOSX). On the MacBook we have the advantage, that it's > an > >> open-source driver. > >> If the power drain is as bad as on the MacBook Pro (and we checked > the > >> Intel-GPU driver, that is's using all power-saving states > available), > >> the problem lies somewhere else... > >> > >> To check these, we need some output of the MacBook's /proc/acpi values > or > >> somebody who helps us to prove these theory... > >> > > > > well I am using a mbp c1d and it eats about 27W with everything on + > > full brightness. as you know powerplay is supported on this book, > > aticonfig --set-powerstate=1 (low power mode) gives me 1-2W less boiling > > down to 25-26W or max. 10 extra minutes (*), if you further set > > display-brightness to the lowest level it goes down to 21-22W and > > backlight off still has 20W. > > > > In OSX I get 23.8W (=2h 18min) with everything on and 18.5W (=2h 58min) > > with everything off+low brightness > > > > so we are looking for reasons (I don't expect it to be a single one) > > that explain a gap of ~4W ... which seems a lot considering that this is > > essentially what the display at full brightness eats up but still may > > not be as big as we expected. > > > > Soeren > > > > (*): this assumes that your battery can be charged to 55Wh mine after > > half a year of intensive usage is already down to 35Wh > > > > > ------------------------------------------------------------------------- > > This SF.net email is sponsored by DB2 Express > > Download DB2 Express C - the FREE version of DB2 express and take > > control of your XML. No limits. Just data. Click to get it now. > > http://sourceforge.net/powerbar/db2/ > > _______________________________________________ > > Mactel-linux-devel mailing list > > Mac...@li... > > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > > > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Mactel-linux-devel mailing list > Mac...@li... > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > |
|
From: Sheer El-S. <sh...@gm...> - 2007-04-25 20:47:58
|
On 4/25/07, Till Straumann <str...@sl...> wrote: > /I can confirm your report (mbp, c1d here) - I also > found similar results (~17-18W) under osx. On linux > I get down to 20-21W (low brighness, disk at rest, CPU freq > at 1GHz, ati powerstate=1) but I found that unloading > the 'uvcvideo' driver (isight) seems to save almost > another ~1W or so - I'm at 19-20W right now. This is considerably better than an mbp c2d. If you don't set powerstate to the lowest state what do you get? Do your fans run at a higher speed than in OS X? On OS X now I'm at 18 W with wireless and a dim screen. This is around 3 hrs (though it goes up to 3.5+ hrs if I let make sure nothing is running in the background). If I could get 19-20 W in linux with the fans quiet that would be spectacular. What do you use to set the disk to rest - laptop mode? thanks, Sheer > > -- Till > > PS I still have the wireless up; shutting it down probably > saves also some power. > / > Soeren Sonnenburg wrote: > > On Tue, 2007-04-24 at 08:10 +0200, Sven Anders wrote: > > > >> Hello! > >> > >> I thought about the power consumption. Maybe we can track the problem down > >> some other way... > >> > >> 1. Is the output of '/proc/acpi/battery/BAT0/state' reliable? If it is, > >> we can use it for testing, otherwise we have to test how long the battery > >> really lasts. > >> > > > > well at least the predicted time matched very well the time when the mbp > > went out of batteries and as that time is based on powerconsumption > > estimates in /proc/*/BAT0/state I would tend to say it is OK. > > > > > >> 2. As far as I remember, the MacBook is very similar to the MacBook Pro in > >> the hardware it's using. The main difference is the GPU. > >> How long does the battery of the MacBook under Linux last in comparison to > >> MacOS X? > >> If it's a small difference, so our problem is the power consumption of > >> the ATI GPU (assuming the Intel-GPU draws as less power under Linux as it > >> does under MacOSX). On the MacBook we have the advantage, that it's an > >> open-source driver. > >> If the power drain is as bad as on the MacBook Pro (and we checked the > >> Intel-GPU driver, that is's using all power-saving states available), > >> the problem lies somewhere else... > >> > >> To check these, we need some output of the MacBook's /proc/acpi values or > >> somebody who helps us to prove these theory... > >> > > > > well I am using a mbp c1d and it eats about 27W with everything on + > > full brightness. as you know powerplay is supported on this book, > > aticonfig --set-powerstate=1 (low power mode) gives me 1-2W less boiling > > down to 25-26W or max. 10 extra minutes (*), if you further set > > display-brightness to the lowest level it goes down to 21-22W and > > backlight off still has 20W. > > > > In OSX I get 23.8W (=2h 18min) with everything on and 18.5W (=2h 58min) > > with everything off+low brightness > > > > so we are looking for reasons (I don't expect it to be a single one) > > that explain a gap of ~4W ... which seems a lot considering that this is > > essentially what the display at full brightness eats up but still may > > not be as big as we expected. > > > > Soeren > > > > (*): this assumes that your battery can be charged to 55Wh mine after > > half a year of intensive usage is already down to 35Wh > > > > ------------------------------------------------------------------------- > > This SF.net email is sponsored by DB2 Express > > Download DB2 Express C - the FREE version of DB2 express and take > > control of your XML. No limits. Just data. Click to get it now. > > http://sourceforge.net/powerbar/db2/ > > _______________________________________________ > > Mactel-linux-devel mailing list > > Mac...@li... > > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > > > > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Mactel-linux-devel mailing list > Mac...@li... > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > |
|
From: Soeren S. <mac...@nn...> - 2007-04-26 08:57:57
|
On Wed, 2007-04-25 at 22:47 +0200, Sheer El-Showk wrote: > On 4/25/07, Till Straumann <str...@sl...> wrote: > > /I can confirm your report (mbp, c1d here) - I also > > found similar results (~17-18W) under osx. On linux > > I get down to 20-21W (low brighness, disk at rest, CPU freq > > at 1GHz, ati powerstate=1) but I found that unloading > > the 'uvcvideo' driver (isight) seems to save almost > > another ~1W or so - I'm at 19-20W right now. > > This is considerably better than an mbp c2d. If you don't set > powerstate to the lowest state what do you get? around 30W > Do your fans run at a > higher speed than in OS X? no. except I force them to (which I do (~2500RPM) as this mbp gets too hot and this is independent of the OS you use) Soeren |
|
From: Till S. <str...@sl...> - 2007-04-26 16:23:28
|
Soeren Sonnenburg wrote: > On Wed, 2007-04-25 at 22:47 +0200, Sheer El-Showk wrote: > >> On 4/25/07, Till Straumann <str...@sl...> wrote: >> >>> /I can confirm your report (mbp, c1d here) - I also >>> found similar results (~17-18W) under osx. On linux >>> I get down to 20-21W (low brighness, disk at rest, CPU freq >>> at 1GHz, ati powerstate=1) but I found that unloading >>> the 'uvcvideo' driver (isight) seems to save almost >>> another ~1W or so - I'm at 19-20W right now. >>> >> This is considerably better than an mbp c2d. If you don't set >> powerstate to the lowest state what do you get? >> > > around 30W > > I can't confirm that the ATI powerstate makes 10W of a difference. In my case savings are only somewhere between 0-2W (mbp c1d). I don't do anything to the fans and unless the CPU gets busy (ondemand governor) the temperature doesn't seem to be excessive. T. >> Do your fans run at a >> higher speed than in OS X? >> > > no. except I force them to (which I do (~2500RPM) as this mbp gets too > hot and this is independent of the OS you use) > > Soeren > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Mactel-linux-devel mailing list > Mac...@li... > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > |
|
From: Soeren S. <mac...@nn...> - 2007-04-26 16:30:24
|
On Thu, 2007-04-26 at 09:21 -0700, Till Straumann wrote: > Soeren Sonnenburg wrote: > > On Wed, 2007-04-25 at 22:47 +0200, Sheer El-Showk wrote: > > > >> On 4/25/07, Till Straumann <str...@sl...> wrote: > >> > >>> /I can confirm your report (mbp, c1d here) - I also > >>> found similar results (~17-18W) under osx. On linux > >>> I get down to 20-21W (low brighness, disk at rest, CPU freq > >>> at 1GHz, ati powerstate=1) but I found that unloading > >>> the 'uvcvideo' driver (isight) seems to save almost > >>> another ~1W or so - I'm at 19-20W right now. > >>> > >> This is considerably better than an mbp c2d. If you don't set > >> powerstate to the lowest state what do you get? > >> > > > > around 30W > > > > > I can't confirm that the ATI powerstate makes 10W > of a difference. In my case savings are only somewhere > between 0-2W (mbp c1d). Ahh, sorry, I misread the email. I meant when I put the c1d to 2.1GHz I get around 30W. regarding ati, I observe exactly the same thing ~1-2W savings with powersave. > I don't do anything to the fans and unless the CPU gets > busy (ondemand governor) the temperature doesn't > seem to be excessive. I am also using the ondemand governor, but in addition force the minimum fan speed a little higher (it is already 24 C in Berlin nowadays...) Soeren |
|
From: Davide B. <da...@da...> - 2007-04-25 07:12:46
|
Ok, I trust him. I was wrong On 4/25/07, Thomas Meyer <th...@m3...> wrote: > > Davide Bertola schrieb: > > On 4/24/07, Sven Anders <an...@an...> wrote: > > > >> Sheer El-Showk schrieb: > >> > >> > >>>> On my system osx idles at ~1200 mA, linux does 16 W -> 1600 mA. > >>>> > >> My MacBook Pro idles at 32W almost constantly. > >> > > > > 32W means you get almost 1,5 hours. That's ridicolus. > > Maybe your scaling governor is not working, or battery is not > calibrated.. > > Then you have a macbook pro so you may have that problem with ati > drivers. > > > > > >>>> It's also hotter when using linux. Why ? > >>>> I guess: > >>>> - Lack of C4 state support. > >>>> > >> I think Linux is using C4(E) state too, but it does not stay as long > >> as MacOSX in it. > >> > >> > > You think, but it's not true. linux sees onlin C1 C2 and C3 state. On > > some macbooks pro only C1 and C2. You can verify that looking > > somewhere in /proc or /sys I don't remember now. > > > > > > Better trust him. Linux does use C4 state. See also > > http://www.mail-archive.com/mac...@li.../msg00918.html > |
|
From: Sheer El-S. <sh...@gm...> - 2007-04-26 16:57:47
|
Hi Soeren and Till, Thanks for the info. I'm a bit confused with your remark Soeren - you're using the ondemand governer but you also say you put the CPU to 2.1 Ghz. Do you get the powersaving when you set the CPU to run at 1 Ghz manually or when you put the ondemand governer. I'm not using Linux now but last time I checked I think I still got over 30 watts when I had the ondemand governer on and the brightness down to near minimum. Other C2D users seem to confirm this (e.g. Sven). In any case the cores seem to be running at 1k most of the time but this doesn't seem to help. I didn't turn off the backlight (I'm not even sure how to do that) nor disable the isight and I was probably using wireless. I think the C2D has a higher min fan speed of 2k rather but even then it often shoots up to 3-5k once you start doing anything. I'll try to see what I can do to reduce my power usage in Linux -- if the ATI card is only 0-2W then there shouldn't be too much of a difference between the C2D and the CD. Any recommendations on what I might be missing or what the issue might be? cheers, Sheer On 4/26/07, Soeren Sonnenburg <mac...@nn...> wrote: > On Thu, 2007-04-26 at 09:21 -0700, Till Straumann wrote: > > Soeren Sonnenburg wrote: > > > On Wed, 2007-04-25 at 22:47 +0200, Sheer El-Showk wrote: > > > > > >> On 4/25/07, Till Straumann <str...@sl...> wrote: > > >> > > >>> /I can confirm your report (mbp, c1d here) - I also > > >>> found similar results (~17-18W) under osx. On linux > > >>> I get down to 20-21W (low brighness, disk at rest, CPU freq > > >>> at 1GHz, ati powerstate=1) but I found that unloading > > >>> the 'uvcvideo' driver (isight) seems to save almost > > >>> another ~1W or so - I'm at 19-20W right now. > > >>> > > >> This is considerably better than an mbp c2d. If you don't set > > >> powerstate to the lowest state what do you get? > > >> > > > > > > around 30W > > > > > > > > I can't confirm that the ATI powerstate makes 10W > > of a difference. In my case savings are only somewhere > > between 0-2W (mbp c1d). > > Ahh, sorry, I misread the email. I meant when I put the c1d to 2.1GHz I > get around 30W. regarding ati, I observe exactly the same thing ~1-2W > savings with powersave. > > > I don't do anything to the fans and unless the CPU gets > > busy (ondemand governor) the temperature doesn't > > seem to be excessive. > > I am also using the ondemand governor, but in addition force the minimum > fan speed a little higher (it is already 24 C in Berlin nowadays...) > > Soeren > |
|
From: Sven A. <an...@an...> - 2007-04-26 21:36:44
|
Hello!
Here are my last result:
Normally I'm running a mactel-patched Linux 2.6.19 kernel.
I'm using the laptop-mode (I attached my config) and dim the brightness down
to 1/3. I'm using the userspace governor with the powernowd (Ubuntu default).
Via laptop-mode I automatically enable "sched_mc_power_savings" and try to set
the wlan (ndiswrapper) power saving mode (by executing
"iwconfig wlan0 power on power max period 3").
I'm getting about 2 hours of battery time and do not experiencing loud fans.
I tried the linux-2.6.21-rc6-mm1 kernel with dynticks, usb-autosuspend, cpufreq
and cpuidle enabled, but it does not help. If I remove as many modules I could,
the lowest power consumption was about 26-27W. I got about 2 hours, too...
I'm running the latest ATI driver (8.36.5) and if I run their new tool
'Catalyst Control Center' it reports a memory clock of 450MHz and a
core clock of 423MHz. Because I own an C2D, I cannot lower these values.
The german computer-magazine C't state a saving of about only 2W, too, if
you set the values via the aticonfig tool.
I'm not sure, if we can trust the aticonfig tool, because maybe there are
some other power-settings of the ATI card, which this tool does not use.
On the other hand, I took a look on MacOSX's power consumption and it was
(when I disable bluetooth and WLAN) about 12-14W. These are still 12-14W lower
than my best...
Maybe I should take another look on any of OSX's PowerManagement extensions. I
only analyzed the "AppleIntelCPUPowerManagement" extension. Are these some
others?
Maybe some of the other devices (the 945GM or ICH7 family chips) have special
power-management settings, which we should enable... Some volunteers to read
the specs? ;-)
The 945GM chipset has a graphic-part. Is it disabled by hardware or has it
to be disabled via software?
Regards
Sven
--
Sven Anders <an...@an...> () Ascii Ribbon Campaign
/\ Support plain text e-mail
ANDURAS service solutions AG
Innstraße 71 - 94036 Passau - Germany
Web: www.anduras.de - Tel: +49 (0)851-4 90 50-0 - Fax: +49 (0)851-4 90 50-55
Rechtsform: Aktiengesellschaft - Sitz: Passau - Amtsgericht Passau HRB 6032
Mitglieder des Vorstands: Sven Anders, Marcus Junker
Vorsitzender des Aufsichtsrats: Dipl. Kfm. Thomas Träger
|
|
From: Tino K. <tin...@ti...> - 2007-04-27 09:09:12
|
On Thu, Apr 26, 2007 at 23:36:26 +0200, Sven Anders wrote: [...] > On the other hand, I took a look on MacOSX's power consumption and it was > (when I disable bluetooth and WLAN) about 12-14W. These are still 12-14W lower > than my best... The numbers in MacOS X are common for a modern notebook. > Maybe I should take another look on any of OSX's PowerManagement extensions. I > only analyzed the "AppleIntelCPUPowerManagement" extension. Are these some > others? > > Maybe some of the other devices (the 945GM or ICH7 family chips) have special > power-management settings, which we should enable... Some volunteers to read > the specs? ;-) > The 945GM chipset has a graphic-part. Is it disabled by hardware or has it > to be disabled via software? Even if it is not disabled in Linux, it shouldn't make a difference of more then 5 Watt IMHO. Are there any any power consumption numbers from non-pro MacBooks? Maybe with the same environment, with lid closed to exclude display power consumtpion etc. Regards, Tino |
|
From: Michael G. <mig...@gm...> - 2007-04-27 09:25:10
|
What about the Windows on MB(P) consumption? I think its even better than the linux one (don't have times, not using it so often). But perhaps it would be good to have a look at the windows site too, not only linux and osx. Perhaps someone finds a solution there. 2.6.21 brings no enhancement under my linux MBP Regards, Michael |
|
From: Sheer El-S. <sh...@gm...> - 2007-04-27 10:15:56
|
Well I think one of the best comparisons is just a 1st gen MBP vs a 2nd gen MBP. Yesterday I tried running Feisty on my MBP C2D with screen dimmed to its lowest, ndiswrapper (wifi) unloaded, using only the vesa driver and I couldn't get it below 28W. When I shutdown X it would go down to 27 W but that's as low as it got. By comparison 1st gens seem to get down to 19-20 W which I think is reasonable. Besides ATI's powerplay (which is reported to only make 1-2 W difference) why are these numbers so different? Could it be the open-source wireless driver enabling power management settings on the wireless? Could it be the higher default fan speed (2k vs 1k I think)? I also think the ATI number are naive because the MBP 1st gens had (afaik) underclocked ATI cards while in the C2D machines not only do they not support power states but they are set by default to a higher speed. If this difference is really consistent then its worth exploring a bit more. Can any other MBP CD owner confirm that they can get 19-20W on a usable configuration (i.e. without disabling everything). What are C2D owners minimum energy consumption rates? I know windows gets reduced battery life when compared to OS X because they mention this on the FAQ on the apple site (I think I found it by googling for applesmc on the apple site) - I think Apple claims this is an applesmc driver issue on windows that they won't do anything about. One useful piece of info in comparing Linux to Windows on this machine is to figure out how much of the power difference is inherant in Linux (i.e. lack of usb suspend, etc...) and how much of it is OS X specific tweaking for these machines (applesmc driver, other hidden power management features). cheers, Sheer On 4/27/07, Michael Gangolf <mig...@gm...> wrote: > What about the Windows on MB(P) consumption? I think its even better than the > linux one (don't have times, not using it so often). > But perhaps it would be good to have a look at the windows site too, not only > linux and osx. Perhaps someone finds a solution there. > > 2.6.21 brings no enhancement under my linux MBP > > Regards, > Michael > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Mactel-linux-devel mailing list > Mac...@li... > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > |
|
From: Sheer El-S. <sh...@gm...> - 2007-04-19 07:55:12
|
Another check which someone suggested to me a while ago but I've been too lazy to setup is to ssh into your MBP remotely and remove as many modules as possible, including usb. I read reports on the kernel mailing list that some machines don't go into C4 if the usb module is loaded but unfortunately on the MBPs this is needed for the keyboard so to test this you have to log in remotely. Also, there's a post on phoronix.com http://www.phoronix.com/scan.php?page=article&item=683&num=2 where the author claims he's checked and the ATI radeon x1650 pro (not exaclty our model) runs much hotter on Linux than on windows (and presumably os x) so a large part of the problem might be the ATI drivers. I'm not sure if powerplay works on this card but the problem may be that the driver is so ineffecient that even using powerplay won't buy us too much energy saving. Do macbooks take as bad a bit on battery life as MBPs? I've heard three hours from some MB owners on the list which seems to still be pretty bad compared to what they should be able to get but it would be interesting to get more concrete data on this. cheers, Sheer On 4/18/07, Nicolas Boichat <ni...@bo...> wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Hi, > > (switched to mactel-devel list, more appropriate I think) > > Sven Anders wrote: > > Hello! > > > > Some data of my machine (for reference): > > > > # dmidecode | egrep '(Version|Release)' > > Version: MBP22.88Z.00A5.B01.0611031551 > > Release Date: 11/03/06 > > Version: 1.0 > > Version: PVT > > Version: Mac-F42187C8 > > Version: Intel(R) Core(TM)2 CPU > > Version: Intel(R) Core(TM)2 CPU > > > > > > This is what I found out: > > ------------------------- > > Wow, impressive report, that's really unfortunate and frustrating it > doesn't really helps us to improve power management on Linux... > > > The following hopes I have: > > > > 1) The auto USB suspend framework will solve the too-many interrupts problems. > > Have you tried this kernel option? > CONFIG_USB_SUSPEND: USB selective suspend/resume and wakeup (EXPERIMENTAL) > > Looks interesting... You might be able to disable some devices and see > if it improves power consumption... > > > 2) The dyn-ticks patches will allow the processor to sleep longer. > > At least in the current version, it doesn't help at all unfortunately. > > Best regards, > > Nicolas > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v2.0.3 (GNU/Linux) > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org > > iD8DBQFGJj0w01ajQnpJXgERAslGAJ9JrDniCCei9mYWg8LmiFCsRRQVCgCfWO/N > tl7r5rUv6Nxxmnm4WWbhgYY= > =btRZ > -----END PGP SIGNATURE----- > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Mactel-linux-devel mailing list > Mac...@li... > https://lists.sourceforge.net/lists/listinfo/mactel-linux-devel > |
Thanks for helping keep SourceForge clean.
X