lkml.org 
[lkml]   [2000]   [Mar]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: kernel performance & profiles under heavy disk I/O
On 20 Mar 2000, Dimitris Michailidis wrote:

>faster than 2.3.99-pre2. Notice the increase in the number of IRQs from the
>disk controller (~250%) and the increase in scheduling activity (~400%), a
>good deal of which is due to __get_request_wait.

Since 2.3.99-pre1 you can tune the elevator with an userspace tools
avaiable in the latest util-linux (see Changes for the URL). With elvtune
you can tune the elevator at runtime to give you a behaviour close to the
one of 2.3.40 if you really want to do so.

The current values gives good performance and good inteactiveness, if you
don't care about interactiveness you are free to change that.

laser:~ # elvtune /dev/hda

/dev/hda elevator ID 0xc02370e4
read_latency: 128
write_latency: 8192
max_bomb_segments: 4

laser:~ # elvtune -r 100000 -w 100000 -b 128 /dev/hda

/dev/hda elevator ID 0xc02370e4
read_latency: 100000
write_latency: 100000
max_bomb_segments: 128

Something like the above should give you bad interactiveness (not infinite
as 2.2.x at least) but more aggressive ordering and coalescing of the
requests.

There's also a manpage that explains some of the details of the parameter.

Thank you for the feedback.

Andrea


-
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 13:57    [W:4.047 / U:0.032 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site