Power consumption on Linux 3.10

Ritesh Raj Sarraf's picture

AddThis: 

The power consumption on the Linux kernel 3.10 is pretty bad.

On kernel 3.10, with the follwing config, the PowerTop results are:

 
#
# Timers subsystem
#
CONFIG_TICK_ONESHOT=y
CONFIG_NO_HZ_COMMON=y
# CONFIG_HZ_PERIODIC is not set
CONFIG_NO_HZ_IDLE=y
# CONFIG_NO_HZ_FULL is not set
CONFIG_NO_HZ=y
CONFIG_HIGH_RES_TIMERS=y
 
PowerTOP v2.0     Overview   Idle stats   Frequency stats   Device stats   Tunables                                  
 
The battery reports a discharge rate of 28.0 W
The estimated remaining time is 23 minutes
 
Summary: 1785.5 wakeups/second,  0.0 GPU ops/second, 0.0 VFS ops/sec and 22.1% CPU use
 
Power est.      Usage       Events/s    Category       Description
  16.3 W     2915 rpm                   Device         Laptop fan
  5.11 W    100.0%                      Device         USB device: WALTON Primo-X1 Primo-X1
  1.70 W     33.3%                      Device         Display backlight
  849 mW     33.3%                      Device         Display backlight
  425 mW     86.0 ms/s     330.7        Process        /usr/bin/konsole
  316 mW     63.9 ms/s      66.1        Process        /usr/bin/plasma-desktop
  142 mW     28.6 ms/s     396.8        Process        /usr/bin/X :0 -auth /var/run/lightdm/root/:0 -nolisten tcp vt7
 64.1 mW     13.0 ms/s     198.4        Process        kwin -session 101261418fe3000136103713100000053880000_13746081
 53.6 mW     10.8 ms/s      0.00        Process        powertop
 35.9 mW      7.3 ms/s      66.1        Process        /usr/lib/chromium/chromium --type=plugin --plugin-path=/usr/li
 24.3 mW      4.9 ms/s     396.8        Interrupt      PS/2 Touchpad / Keyboard / Mouse
 6.92 mW      1.4 ms/s      0.00        Interrupt      [48] i915
 5.94 mW      1.2 ms/s      66.1        Interrupt      [9] RCU(softirq)
 3.98 mW      0.8 ms/s      0.00        kWork          flush_to_ldisc
 3.78 mW      0.8 ms/s      66.1        Process        [ksoftirqd/2]
 3.33 mW    673.3 us/s      66.1        Process        [rcu_sched]
 1.80 mW    363.1 us/s      66.1        Interrupt      [1] timer(softirq)
 1.79 mW    363.0 us/s      0.00        Process        [ksoftirqd/4]
 

Where as on the 3.9 kernel:

 
 
The battery reports a discharge rate of 13.2 W
The estimated remaining time is 43 minutes
 
Summary: 611.5 wakeups/second,  0.0 GPU ops/second, 0.0 VFS ops/sec and 14.2% CPU use
 
Power est.      Usage       Events/s    Category       Description
  14.0 W     2722 rpm                   Device         Laptop fan
  1.72 W     33.3%                      Device         Display backlight
  862 mW     33.3%                      Device         Display backlight
  255 mW     65.7 ms/s      58.0        Process        /usr/bin/plasma-desktop
 91.9 mW     23.7 ms/s      27.5        Process        /usr/lib/chromium/chromium --type=renderer --lang=en-US --forc
 60.1 mW     15.5 ms/s      96.1        Process        /usr/lib/chromium/chromium --type=plugin --plugin-path=/usr/li
 25.0 mW      6.4 ms/s      25.1        Process        kwin -session 101261418fe3000136103713100000053880000_13746094
 21.5 mW      5.6 ms/s      34.2        Process        /usr/bin/X :0 -auth /var/run/lightdm/root/:0 -nolisten tcp vt7
 13.1 mW      3.4 ms/s       5.6        Process        /usr/bin/konsole
 9.82 mW      2.5 ms/s      53.7        Process        [irq/48-iwlwifi]
 9.11 mW      2.4 ms/s       2.2        Process        /usr/bin/knemo
 8.62 mW      2.2 ms/s      12.3        Process        /usr/lib/chromium/chromium --password-store=detect
 8.32 mW      2.1 ms/s      45.5        Interrupt      [48] iwlwifi
 6.96 mW      1.8 ms/s      35.3        Interrupt      [7] sched(softirq)
 5.13 mW      1.3 ms/s      57.1        Interrupt      [47] i915
 4.24 mW      1.1 ms/s       0.4        Process        powertop
 3.38 mW      0.9 ms/s       1.5        Timer          tcp_keepalive_timer
 2.85 mW      0.7 ms/s      11.9        Process        /usr/sbin/mysqld --basedir=/usr --datadir=/var/lib/mysql --plu

Categories: 

Keywords: 

Comments

USB device

Did you detach the USB device at the second try? It looks like your test conditions are not equal.

Ritesh Raj Sarraf's picture

Nope. The setup was as it is.

Nope. The setup was as it is. Just a reboot, nothing else. The USB controller not being suspended in 3.10 could be one reason. But I have more suspicion on the Timer changes. Anyways, I will wait till 3.11.

Add new comment