• 34 Posts
  • 3.04K Comments
Joined 6 years ago
cake
Cake day: May 31st, 2020

help-circle



  • It’s just really oversimplifying memory usage. OS designers had that same thought decades ago already, so they introduced disk caching. If data gets loaded from disk, then it won’t be erased from memory as soon as it isn’t needed anymore. It’s only erased, if something else requests memory and this happens to be the piece of “free” memory that the kernel thinks is the most expendable.

    For example, this is what the situation on my system looks like:

    free -h
                   total        used        free      shared  buff/cache   available
    Mem:            25Gi       9,8Gi       6,0Gi       586Mi       9,3Gi        15Gi
    

    Out of my 32 GiB physical RAM, 25 GiB happens to be usable by my applications, of which:

    • 9.8 GiB is actually reserved (used),
    • 9.2 GiB is currently in use for disk caching and buffers (buff/cache), and
    • only 6.1 GiB is actually unused (free).

    If you run cat /proc/meminfo, you can get an even more fine-grained listing.

    I’m sure, I could get the number for actually unused memory even lower, if I had started more applications since booting my laptop. Or as the Wikipedia article I linked above puts it:

    Usually, all physical memory not directly allocated to applications is used by the operating system for the page[/disk] cache.

    So, if you launch a memory-heavy application, it will generally cause memory used for disk caching to be cleared, which will slow the rest of your system down somewhat.

    Having said all that, I am on KDE myself. I do not believe, it’s worth optimizing for the speed of the system, if you’re sacrificing features that would speed up your usage of it. Hell, it ultimately comes down to how happy you are with your computer, so if it makes you happy, then even gaudy eye-candy can be the right investment.
    I just do not like these “unused RAM is wasted RAM” calls, because it is absolutely possible to implement few features while using lots of memory, and that does slow your system down unnecessarily.









  • To be honest, what I’m most mad about isn’t the typoes, it’s that someone generated this image and figured, yeah alright, that will clear things up.

    On some level you want to believe that even if someone does not come up with a proper concept for a visualization, that they still check what the AI shat out, so that it’s at the very least not conceptually wrong and not confusing.

    This image isn’t just shitty, it’s actively worse than having no visualization. They could’ve generated that, chuckled, and not used it. Just how do you blunder your perception check so badly that you decide to include it anyways?




  • Ephera@lemmy.mltoLinux@lemmy.mlKDE Plasma 6.6 released
    link
    fedilink
    English
    arrow-up
    1
    ·
    3 days ago

    Yeah, I’ve done that occasionally, too, but it adds a load of friction for moving windows between screens, in particular also when un-/replugging the screen, so it’s still painful enough that I don’t bother with a second screen.

    I guess, it also plays a role that I do use lots of workspaces, so it’s 1) extra painful and 2) I don’t have as big of a need for a second screen, since I can just switch out what first screen displays very quickly.



  • Ephera@lemmy.mltoLinux@lemmy.mlKDE Plasma 6.6 released
    link
    fedilink
    English
    arrow-up
    6
    ·
    3 days ago

    Oh boy, feature freeze for Ubuntu 26.04 is on Thursday. Hopefully, they still include this update.

    My work laptop unfortunately comes with Kubuntu LTS and I desperately want the virtual-desktops-only-on-the-primary-screen feature on there. Currently, I’m the guy that actively disables all but one screen, because my workflow does not work at all with the secondary screen switching in sync with the primary screen.


  • Ephera@lemmy.mltoLinux@lemmy.mlKDE Plasma 6.6 released
    link
    fedilink
    English
    arrow-up
    1
    ·
    3 days ago

    I still wouldn’t assume it to actually go further than that. It’s a limitation of the EWMH standard, which is used for controlling the placement of windows.

    I don’t have in-depth knowledge of the standard, but I assume, it can only represent 1 desktop as the active desktop and stuff like that.
    Maybe you could try to be clever by e.g. always reporting the active desktop of the active screen and stuff like that, but yeah, no idea if you can do that for all aspects of the standard, and whether applications will still behave as expected.