lkml.org 
[lkml]   [1999]   [Oct]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    /
    SubjectRe: Speeding up swap
    From
    Date
    kris@koehntopp.de (Kristian Koehntopp) writes:

    > In netuse.lists.linux-kernel you write:
    > >Benchmarking the harddisk area which is being used for swap, he
    > >can get sustained write rates of 26 MB/sec under Windows NT
    > >4.0SP5 as well as under Linux 2.2.12.
    >
    > Because it was asked on the list: The disk is a 10000 rpm IBM
    > Ultrastar 18 and the swap is in the fastest zone of this disk.
    > This was the fastest disk they could get on short notice. The
    > disk is monstrously noisy when seeking. :-)
    >
    > Question: How would the kernel benefit from additional
    > information about user space memory when swapping, had Linux a
    > madvise() call? Is it useful for the kernel to know that a
    > certain block of memory is always addressed sequentially or that
    > a certain block of memory should be swapped en-bloc?

    It could optimize throwing away pages that won't be reused soon again
    (the optimal caching policy is very different when you can assume
    sequential accesses, instead of random accesses)

    BTW, 2.3 has some extensions for 64KB mmap cluster IO. Unfortunately
    its general MM balance seems to be still a bit unoptimal (i can easily
    produce swap storms at 64MB machines - seems try_to_free_pages is too
    aggressive). But maybe it helps.


    -Andi
    --
    This is like TV. I don't like TV.

    -
    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.rutgers.edu
    Please read the FAQ at http://www.tux.org/lkml/

    \
     
     \ /
      Last update: 2005-03-22 13:54    [W:3.103 / U:0.008 seconds]
    ©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site