Bug 921 - Workrave generates unreasonable number of wakeups
Status:
NEW
Component:
Core
Version:
1.9.1
Hardware:
All Linux
Importance:
P5 normal
Target Milestone:
---
Assignee:
Rob Caelers
URL:
Depends on:
Blocks:
Reported:
Sep 1 2010 17:18:25 UTC
by:
Craig Younkins
Modified:
Jun 8 2020 14:53:04 UTC
CC List:
Filipus Klutiero
WhoWhenWhatRemovedAdded
Filipus KlutieroJun 8 2020 14:53:04 UTCccFilipus Klutiero
Description
Craig Younkins  Sep 1 2010 17:18:25 UTC
I have workrave set up with just one timer (rest break), no status window, no applet, and no network config. With this basic setup, powertop is reporting that workrave issues ~11 wakeups per second. 

I feel this is unreasonable for what workrave is doing. Please try to increase the time between wakeups by sleeping for longer periods of time. I think most people would agree that updating the timer once per second is enough.

Reducing the wakeups per second from workrave will improve battery life in mobile devices. Thanks!
Comment 1
Rob Caelers  Sep 5 2010 21:22:32 UTC
I tried to reproduce this but I get other results:

Top causes for wakeups:
...
  0.1% (  1.3)   workrave

Workrave only updates its GUI once every second. So 11 wakeups per second is quite a lot.

Does you X server have the record extension enabled? You can check this by running:

$xdpyinfo -queryExtensions | grep -i record
    RECORD  (opcode: 136, base error: 160)


With record, Workrave may use some more CPU. 

What Linux distribution are you using? I would like to try powertop with the same distribution.
Comment 2
Craig Younkins  Sep 5 2010 22:11:33 UTC
"Does you X server have the record extension enabled? You can check this by
running:"

Yes, apparently it does.

"What Linux distribution are you using?"

64-bit Arch.

$ pacman -Q xorg-server
xorg-server 1.8.1.902-1
Comment 3
Filipus Klutiero  Jun 8 2020 14:53:04 UTC
I am also experiencing a somewhat excessive CPU usage from Workrave on Windows 10 if Workrave's task is considered.
I do not know how to tell the average consumption, but I would estimate it to 0.1%. I can install software to provide a more accurate measure if needed.