mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Badari Pulavarty <pbadari@us.ibm.com>
To: linux-kernel@vger.kernel.org (lkml)
Subject: 2.5.47-mm3 IO rate question
Date: Fri, 15 Nov 2002 10:37:48 -0800 (PST)	[thread overview]
Message-ID: <200211151837.gAFIbm610839@eng2.beaverton.ibm.com> (raw)

Hi,

I am using 2.5.47-mm3 and qlogic driver qla2x00src-v6.03.00b8.
I have 2 qlogic controllers connected to 2 trays (10 disks/tray).

I do 256K IO on raw devices using "dd".

dd of=/dev/null bs=256k count=2000 if=/dev/raw/raw1 &
dd of=/dev/null bs=256k count=2000 if=/dev/raw/raw2 &
dd of=/dev/null bs=256k count=2000 if=/dev/raw/raw3 &
...
dd of=/dev/null bs=256k count=2000 if=/dev/raw/raw20 &

IO throughput changes just by adjusting SG_SEGMENTS (from 32 to 64).
I was wondering why such a significant difference. (20MB/sec).

I used to get 194-198MB/sec with 2.5.37 (with DIO code sending
(fixed) 64K requests - qlogic max_sector=512, SG_SEGMETS=32).

Any thoughts ?

Thanks,
Badari

max_sectors = 512
SG_SEGMENTS = 64
[root]# vmstat 5
   procs                      memory      swap          io     system      cpu
 r  b  w   swpd   free   buff  cache   si   so    bi    bo   in    cs us sy id
 0 20  0      0 3642680   8832  42836    0    0 194765    15 1762  2183  0 100  0
 0 20  0      0 3642680   8836  42836    0    0 194611     3 1761  2122  0 100  0
 0 20  0      0 3642680   8840  42836    0    0 194509     3 1774  2150 10 81  8
 0 20  0      0 3642680   8844  42836    0    0 194816     3 1762  2132  2 98  0
 0 20  0      0 3642680   8848  42836    0    0 194765     3 1764  2177  2 98  0


max_sectors = 512
SG_SEGMENTS = 32
[root]# vmstat 5
   procs                      memory      swap          io     system      cpu
 r  b  w   swpd   free   buff  cache   si   so    bi    bo   in    cs us sy id
 0 20  0      0 3652472   7752  34464    0    0 170755    15 2354  3477  2 98  0
 0 20  0      0 3652472   7756  34464    0    0 172771     3 2363  3512  2 98  0
 0 20  0      0 3652472   7760  34464    0    0 174031     3 2372  3576  0 100  0
 0 20  0      0 3652408   7764  34464    0    0 173074     7 2384  3538 10 87  3
 0 20  0      0 3652408   7768  34464    0    0 172670     3 2364  3502  2 98  0


             reply	other threads:[~2002-11-15 18:32 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-11-15 18:37 Badari Pulavarty [this message]
2002-11-17  1:00 ` Andrew Morton

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200211151837.gAFIbm610839@eng2.beaverton.ibm.com \
    --to=pbadari@us.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®