Questions to energy consumption test in German Linux Magazin 3/2012

[Paul Menzel]

Dear Lesswatts folks,


issue 3/2012 of the German Linux Magazin looked at the energy
consumption of different GNU/Linux distributions.

A Fujitsu Lifebook e751 vPro with an Intel Core i7-2620M and Intel
graphics is used.

As I read the result OpenSUSE 12.1(?) with `kernel-default` and GNOME
has the lowest value of 11,4 Watts shown in PowerTOP.

Using their benchmark to simulate user input the distributions compare
as follows.

Ubuntu (Unity) 3:04 hours
Mint (GNOME) 3:14 hours
openSUSE (KDE) 3:23 hours
Mint (LXDE) 3:25 hours
Windows 7 Professional 4:05 hours

As far as I could see they enable our power saving features like RC6 for
the Intel graphics hardware.

Additionally they also write that although tuning a lot of setting and
PowerTOP also showing that less Watts are consumed that this did not
have a big effect on the runtime. Is there something wrong with this
setup?

Additionally are there obvious explanations why Windows still runs that
much longer?


Thanks,

Paul


[1] http://www.linux-magazin.de/Heft-Abo/Ausgaben/2012/03/Rat-und-Tat

[Kok, Auke-jan H]

On Fri, Feb 3, 2012 at 2:08 AM, Paul Menzel

Post by Paul Menzel
As I read the result OpenSUSE 12.1(?) with `kernel-default` and GNOME
has the lowest value of 11,4 Watts shown in PowerTOP.
Using their benchmark to simulate user input the distributions compare
as follows.
Ubuntu (Unity) 3:04 hours
Mint (GNOME) 3:14 hours
openSUSE (KDE) 3:23 hours
Mint (LXDE) 3:25 hours
Windows 7 Professional 4:05 hours
As far as I could see they enable our power saving features like RC6 for
the Intel graphics hardware.
Additionally they also write that although tuning a lot of setting and
PowerTOP also showing that less Watts are consumed that this did not
have a big effect on the runtime. Is there something wrong with this
setup?

yes - powertop is a diagnostics tool, and only limited as to what it
can see. If hardware drivers conceal power savings features, or omit
to implement them, then powertop is often (obviously) in the dark as
to what is really going on in the hardware.

Post by Paul Menzel
Additionally are there obvious explanations why Windows still runs that
much longer?

Missing specs, undocumented firmware, vendor modifications to
hardware, firmware, BIOS, ACPI, etc...

Also, obviously (?) the vendor shipped the unit with a non-Linux OS,
so a good question would be to actually ask the vendor if they have a
version optimized for Linux.

Auke