• turdas@suppo.fi
    link
    fedilink
    arrow-up
    9
    ·
    9 hours ago
    $ zramctl
    NAME       ALGORITHM DISKSIZE  DATA COMPR   TOTAL STREAMS MOUNTPOINT
    /dev/zram0 lzo-rle      62.6G  2.8G  972M 1011.4M         [SWAP]
    

    Already did

    • lambalicious@lemmy.sdf.org
      link
      fedilink
      English
      arrow-up
      1
      ·
      2 hours ago

      Based zramctl. Makes my 8GB RAM system run like I had 12 GB, which is quite significant in this new internet world where opening a second tab in a web browser costs almost 600 MB.

    • ominous ocelot@leminal.space
      link
      fedilink
      arrow-up
      2
      ·
      6 hours ago

      What’s the use case over RAM or disk swap? It’s compressed but faster than SSD? Hmm. That could help in distinct use cases…

      • turdas@suppo.fi
        link
        fedilink
        arrow-up
        4
        ·
        6 hours ago

        Yes, it’s basically faster than disk swap but uses some CPU cycles. The compression algorithms involved are very fast on modern CPUs so in some sense it’s “free RAM”.

        I set mine to almost 1:1 my physical RAM, because the way it works is that the zram disk size (62.6G there) is the amount of uncompressed data allowed on it, and the compression on real-world data is almost always at least 50% – so if the zram device fills up, it’ll be using something like 32G of physical memory. I’m yet to hit real-world usecases that would have tested these limits though, and the defaults are much more conservative.