RE: IDE-RAID Drive Performance

["Nicklas Bondesson" <nicke _at_ xxxxxxxx>]

I don't know if you have read my latest post regarding this. anyway here it
is :)

It was the PCI latency value that was wrong. It was set to 0 in the BIOS so
I have changed it now to 64. I now get the following figures! :) BTW I'm
running two Western Digital WD800JB-00DUA3 (Special Edition 8 MB cache)
disks connected to a Promise TX2000 (PDC20271) card (RAID1 using ataraid
(pdcraid) drivers).

hdparm:

/dev/hda:
 Timing buffer-cache reads:   128 MB in  1.12 seconds =114.29 MB/sec
 Timing buffered disk reads:  64 MB in  1.53 seconds = 41.83 MB/sec

bonnie:

Version 1.02b       ------Sequential Output------ --Sequential Input-
--Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block--
--Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec
%CP
gollum         800M  4681  98 37827  76 17590  30  4888  95 39400  26 310.4
3
                    ------Sequential Create------ --------Random
Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read---
-Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec
%CP
                 16   310  99 +++++ +++ 21319 100   329  99 +++++ +++  1766
96
gollum,800M,4681,98,37827,76,17590,30,4888,95,39400,26,310.4,3,16,310,99,+++
++,+++,21319,100,329,99,+++++,+++,1766,96 

Thanks for the help!

/Nicke 

-----Original Message-----
From: Andre Tomt [mailto:lkml _at_ tomt.net] 
Sent: den 3 januari 2004 15:00
To: Nuno Alexandre
Cc: linux-kernel _at_ vger.kernel.org; Nicklas Bondesson
Subject: Re: IDE-RAID Drive Performance

On Tue, 2003-12-30 at 12:21, Nuno Alexandre wrote:
> /dev/hda:
>  Timing buffer-cache reads:   1320 MB in  2.00 seconds = 659.44 MB/sec
>  Timing buffered disk reads:  140 MB in  3.02 seconds =  46.40 MB/sec
> 
> Using:
> -d1 -u1 -m16 -c3 -W1 -A1 -k1 -X70 -a 8192

Wow, slow down for a minute. Most IDE chipset drivers does a excellent job
at autotuning the max *safe settings* for your drive/chipset combination.
Mucking around with hdparm parameters blindfolded will only cause you grief
in form of data loss and system instability sooner than later.

Usually when one gets into performance problems with IDE in Linux, the
chipset specific driver is not enabled, making the system fallback to the
generic driver - OR the drive and controller combination is considered
unsafe with faster settings.

Without any user intervention at all, my Seagate 7200 120G's does 55MB/s in
the infamious hdparm test, on a VIA KT266 based board, both in
2.6.1-rc1 and 2.4.23.



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