• ssd - i/o scheduler

    I read wiki how improve system and ssd disc and it`s many tutorial how change i/o scheduler but my question is which is the best for ssd? Once people recommend noop (arch wiki) another time deadline (debian wiki) on manjaro they prefer bfq so sorry but I have no experience in this field and I am completely lost different opinions are not clear for me.

    Maybe someone will share his experience?

  • @joekamprad

    no. These topics are about TRIM and I want to know what scheduler is the best for ssd:

    use this command to check what you use

    dmesg | grep scheduler
  • Scheduler for SSDs The Deadline Scheduler is well-suited for interacting with SSDs because it does not rely on the excessive merging of requests. In many places, NOOP is praised as the best scheduler for SSDs because it simply passes the requests to the device. For performance reasons this is quite conclusive, but the fairness of the scheduler times under applications can suffer.
    Source: Thomas-Krenn

  • Do you already try to change the scheduler?
    i try it and adding elevator=noop to kernel parameters does not change it.
    But temporarly it works:

    [[email protected] killajoe]# echo noop > /sys/block/sdb/queue/scheduler
    [[email protected] killajoe]# exit
    ╭─[email protected] ~  
    ╰─➤  cat /sys/block/sdb/queue/scheduler
    [noop] deadline cfq 
  • systemd-tmpfiles solution works:
    Also the best solution if you get different HD in system

  • yes i try change scheduler

    GRUB_CMDLINE_LINUX_DEFAULT="quiet elevator=noop"

    and after that

    sudo grub-mkconfig -o /boot/grub/grub.cfg

    and for me works so i dont need temporary solution. but thx for help

  • systemd-tmpfiles solution is permanent too but per device:
    create: /etc/tmpfiles.d/10_ioscheduler.conf with this inside
    w /sys/block/sdX/queue/scheduler - - - - noop

ssd30 scheduler1 Posts 8Views 1270
Log in to reply
Bloom Email Optin Plugin

Looks like your connection to Antergos Community Forum was lost, please wait while we try to reconnect.