Sata speed - slowed by half. SATA_SIL

drbenway · n00b Joined: 26 Jan 2006 Posts: 56

Hi.

I have a SIL3112 Sata (Nforce2) motherboard.

I have noticed that the speed of the sata drive in now 1/2 of what it was.

Here is what hdparm shows now:-
-------------------------------------
/dev/sda:
Timing cached reads: 844 MB in 2.00 seconds = 421.08 MB/sec
Timing buffered disk reads: 136 MB in 3.02 seconds = 44.99 MB/sec

Heres what hdparm used to show
-------------------------------------

/dev/sda:
Timing cached reads: 1944 MB in 2.00 seconds = 971.71 MB/sec
Timing buffered disk reads: 170 MB in 3.01 seconds = 56.48 MB/sec

..........

Also from looking through /var/log/messages I notice that i now get these lines :-
-------------------------------------------------------

Mar 10 17:14:51 localhost SCSI device sda: 320173056 512-byte hdwr sectors (163929 MB)
Mar 10 17:14:51 localhost sda: Write Protect is off
Mar 10 17:14:51 localhost sda: Mode Sense: 00 3a 00 00
Mar 10 17:14:51 localhost SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA
Mar 10 17:14:51 localhost sda: sda1 sda2 sda3 < sda5 sda6 sda7 >
Mar 10 17:14:51 localhost sd 1:0:0:0: Attached scsi disk sda

Whilst i used to get this:-
-------------------------------------------------
localhost SCSI device sda: 320173056 512-byte hdwr sectors (163929 MB)
Feb 5 22:54:55 localhost sda: Write Protect is off
Feb 5 22:54:55 localhost sda: Mode Sense: 00 3a 00 00
Feb 5 22:54:55 localhost SCSI device sda: drive cache: write back

The only changes i have done since I used to get proper speeds is a udev update and a kernel update (2.6.18 --> 2.6.20)

Any ideas anyone ?
_________________
Studies show that 9 out of 10 paedophiles use Windows ...
------------------------------------------------------------------------------------------------

NeddySeagoon · Posted: Sat Mar 10, 2007 5:45 pm Post subject:

drbenway,

It is a change in hdparam. Old versions give one answer, new versions the other answer.
Maybe they are both wrong ?
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

drbenway · n00b Joined: 26 Jan 2006 Posts: 56

Hi.

Thank you for your response.

Are you saying that the hdparm test speed is not important ?

I have noticed the same speed on ubuntu feisty and suse 10.3 alpha 1, however ubuntu edgy give the faster older speed.

What is the correct way of benchmarking my sata drive ?
_________________
Studies show that 9 out of 10 paedophiles use Windows ...
------------------------------------------------------------------------------------------------

aidanjt · Posted: Sun Mar 11, 2007 1:17 pm Post subject:

NeddySeagoon · Posted: Sun Mar 11, 2007 3:32 pm Post subject:

drbenway,

I am not suggesting that the output of hdparm is not important, just that it may not be right.

The problem with 'benchmarks' and 'speed tests' is that they rarely reflect real world performance.
They are good indicators that something has changed but thats all. In this case, I'm suggesting that its hdparm itself that has changed, not your hardware. To get even useful comparative results, identical tests must be run under identical test conditions.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

albright · Posted: Sun Mar 11, 2007 4:00 pm Post subject:

Yes, it is hdparm change. Here's a quote from hdparm homepage
about the upgrade to 6.9: