Bug 388581 - Option to show combined CPU load history
Summary: Option to show combined CPU load history
Status: RESOLVED UNMAINTAINED
Alias: None
Product: ksysguard
Classification: Unmaintained
Component: ksysguard (show other bugs)
Version: 5.11.3
Platform: Other Linux
: NOR wishlist
Target Milestone: ---
Assignee: KSysGuard Developers
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-01-05 16:51 UTC by Gregor Mi
Modified: 2024-09-23 21:01 UTC (History)
2 users (show)

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:


Attachments
Resulting graph with a simple endless loop program (98.74 KB, image/png)
2018-01-05 16:51 UTC, Gregor Mi
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Gregor Mi 2018-01-05 16:51:46 UTC
Created attachment 109696 [details]
Resulting graph with a simple endless loop program

I have 4 CPUs and in KSysGuard > System Load > CPU History the CPUs are shown separately.

In some situations this causes "wild" graphs: E.g. run this python script:

while True:
    print("Infinite Loop")

and the graph looks like in the attachment.

I wonder if it makes sense to have an option to show a graph with the combined CPU load. This would mean for 4 CPUs that each CPU would contribute with max. 25 % to the 100 % Y axis.
Comment 1 Becker Béla 2018-01-19 08:49:17 UTC
Have you tried stacking the beams on top of each other? It is similar to what you ask for.
Comment 2 Gregor Mi 2018-01-20 12:56:27 UTC
I didn't know the option before. It is similar, yes, but is missing one thing I find useful: with my infinite loop example it still shows a 100% load instead a much lower percentage since not all CPUs are loaded 100 %.
Comment 3 Becker Béla 2018-01-20 18:38:02 UTC
Usage will go above 100% with more CPU's.
Most applications will not normalize the usage to the number of cores.
I think this is OK.
Comment 4 Gregor Mi 2018-01-20 19:46:36 UTC
Thanks for pointing this out. I have 4 cores, therefore I went to the graph properties > Scales and set the graph range to 0 - 400 %.

I used then this command

$ dd if=/dev/zero of=/dev/null

(see https://stackoverflow.com/questions/2925606/how-to-create-a-cpu-spike-with-a-bash-command) to generate 100 % load for one, two, three and four cores to verify. It works fine.

Two usability things:
1. When I had the scale at 100 % (automatic) there it was not intuitive to see that the graph was "overflowing" when the processor was loaded more than 100 %.
2. A "Show combined CPU usage" option could either normalize to 100 % or automatically adjust the scale to the fitting percentage.
Comment 5 Christoph Cullmann 2024-09-23 21:01:27 UTC
ksysguard is no longer maintained, in Plasma 6 there is the Plasma system monitor for this task.

If your wish is still valid for the Plasma 6 replacement, please re-open and we can move this bug to the new product, thanks!