[lkml]   [1998]   [Mar]   [11]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    SubjectRe: 2.1.89 broken?
    On Wed, 11 Mar 1998, Rik van Riel wrote:

    >On Tue, 10 Mar 1998, Scott Lampert wrote:
    >> Far be it from me to criticize ideas made by those who know
    >> much more than me, but personally I don't see how transfering stuff to swap
    >> immediately instead of grabbing what ram we can from the disk
    >> cache first could be considered faster or preferable. As far as
    >> I can see, it would be much better performance-wise to delay swapping as
    >> long as there's ram to be had from the disk cache. Isn't this the main
    >> reason the disk cache is dynamic in size?

    I think so:

    total used free shared buffers cached
    Mem: 31188 30228 960 3604 704 23512
    -/+ buffers/cache: 6012 25176
    Swap: 52412 7084 45328

    In my system there is more swap than used memory while disk cache is more
    than 3 times the used memory. I don' t need such cache (maybe it was
    created by a recursive grep that if I relaunch from the beginning it will
    not use the cache because it needs the top of the cache while the one
    present is relative to the end of the recursive grep). Due this cache
    priority against process memory, every time I return in a sleeping task it
    take a _while_ to wakeup from the swap. I think we should swap only
    process that are sleeping for more time than in 2.1.90-pre1, and free the
    cache instead. What I am tring tell is that I like to have I/O access
    during a recursive grep instead of when I return to emacs. Also the
    buffer memory is very low but I don' t know what does it mean.

    andrea@arca:~/devel/kernel/linux$ grep pippo `find -name \*.[ch]`
    andrea@arca:~/devel/kernel/linux$ free
    total used free shared buffers cached
    Mem: 31188 30548 640 604 520 26808
    -/+ buffers/cache: 3220 27968
    Swap: 52412 10628 41784

    I think that removing the 10Mbyte swapped from the cache will make
    performance better.

    Andrea[s] Arcangeli

    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to

     \ /
      Last update: 2005-03-22 13:41    [W:0.021 / U:189.304 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site