IS7 - Serial ATA performance - Sandra

[John (77)]

I was wondering what kind of SATA performance others are experiencing with their IS7 boards in the non RAID mode? I ran the Sandra file system test and the result of 30800 KB/S looks low to me, considering I'm using a WD Raptor drive.

The MB is not overclocked and the BIOS was updated to version 13. Properties for the disk controller shows it running in Mode 5. The Sandra memory test results are 4411 MB/S & 4456 MB/S which is about what I would expect for Kingston 3-3-3 memory. The overall performance feels sluggish compared to other P4 2.4 systems I have built. I would have thought that using a WD Raptor drive would make for a much snappier system. Have I missed something?

Sandia correctly identifies the CPU as a 2.4 G and shows FSB of 202 MHz.

System:
P4 2.4G 800 MHz
2 X 256 MB Kingston set up in Dual Mode
Western Digital Raptor SATA drive (XP system drive C)
BIOS 13
BIOS SATA set to Enhanced mode
1 Sony CD
1 Sony CD/RW
1 Western Digital 40 G (NTFS but empty drive D)
Windows XP

Anyone have a Raptor running on an IS7 in the non-RAID mode churning out better results ... or some thoughts?

Thanks,

John

 

[zLaTi]

Hi

Uzytkownik "John" napisal w wiadomosci

I was wondering what kind of SATA performance others are experiencing with their IS7 boards in the non RAID mode? I ran the Sandra file system test and the result of 30800 KB/S looks low to me, considering I'm using a WD Raptor drive.

<snipped out>

You might want to try HDTach instead, since Sandra
(like every women <Yes, I'm married>) often lies...

When you do that, post your results here...

'til then

mgr inz. zLaTi

Anyone have a Raptor running on an IS7 in the non-RAID mode churning out better results ... or some thoughts?

Thanks,

John

 

[TomG]

Sandra's test is notorious for being a poor test... try HDTach for non raid setups or Diskspeed32 for non-raid or raid setups. I carry both on my server...

HDTach log of a raid array that has not been recently defragged running on an add-in version of a Highpoint 372 controller: HD Tach version 2.61
Drive: HPT37x RAID 0 Array 2.00
Access time: 12.6ms
CPU utilization: 5.6%

Burst speed: 96184kps
Average read speed: 54622

HDTach log of a IBM 120GXP on SATA 1 through Serillel-2: HD Tach version 2.61
Drive: IC35L120AVVA07-0
Access time: 12.4ms
CPU utilization: 5.0%
Burst speed: 85414kps
Average read speed: 38072

just to give you some ideas of performance...

 

[ga]

I have the exact same setup with you except a different brand of ram (ocz) and pc3500. I'm getting exactly 30800 KB/S too, so I suspect this is normal? If you find a improvement please post back here or email. Thanks

John wrote:

I was wondering what kind of SATA performance others are experiencing with their IS7 boards in the non RAID mode? I ran the Sandra file system test and the result of 30800 KB/S looks low to me, considering I'm using a WD Raptor drive.

The MB is not overclocked and the BIOS was updated to version 13. Properties for the disk controller shows it running in Mode 5. The Sandra memory test results are 4411 MB/S & 4456 MB/S which is about what I would expect for Kingston 3-3-3 memory. The overall performance feels sluggish compared to other P4 2.4 systems I have built. I would have thought that using a WD Raptor drive would make for a much snappier system. Have I missed something? ...

 

[Jerry Barrett]

Hi, I remember reading on of Toms Hardware tests with SATA drives, you have to change the block size to get peek perfomance, I seem to remember them mentionsing 64K blocks as opposed to 256K blocks....

Have a Google 'SATA block optimis' for it and see what it comes back with or something similar.

Good Luck
Jerry
"John" wrote:

I was wondering what kind of SATA performance others are experiencing with their IS7 boards in the non RAID mode? I ran the Sandra file system test and the result of 30800 KB/S looks low to me, considering I'm using a WD Raptor drive.

The MB is not overclocked and the BIOS was updated to version 13. Properties for the disk controller shows it running in Mode 5. The Sandra memory test results are 4411 MB/S & 4456 MB/S which is about what I would expect for Kingston 3-3-3 memory. The overall performance feels sluggish compared to other P4 2.4 systems I have built. I would have thought that using a WD Raptor drive would make for a much snappier system. Have I missed something? ...

 

[John (77)]

Tom G,

Thanks for taking the time to run HDTach for me and reporting your readings. I couldn't run the version of HDTach you have on your site because I have XP set up as a NTFS. However, I did run Diskspeed32 and this is what it came up with for the Raptor.

Max read 64,256 KB/s
Min read 16,595 KB/s
Average read 48,471 KB/s
CPU usage 4%

I don't know how close the Diskspeed32 readings compare to HDTach. Just thought the Raptor would perform a little better than this. Guess I was hoping for to much.

I did read the other post about playing with the drive block size possibly helping. Unfortunately, I'm on vacation and set this system up for a relative. My stay is about over so further testing will have to wait till I get back home and upgrade my home system.

Thanks to all who replied,

John

TomG wrote:

Sandra's test is notorious for being a poor test... try HDTach for non raid setups or Diskspeed32 for non-raid or raid setups. I carry both on my server...

HDTach log of a raid array that has not been recently defragged running on an add-in version of a Highpoint 372 controller:
HD Tach version 2.61
Drive: HPT37x RAID 0 Array 2.00
Access time: 12.6ms
CPU utilization: 5.6%

Burst speed: 96184kps
...

 

[TomG]

when I get a moment, I run Diskspeed32 to see how it compares...

 

[ga]

"Jerry Barrett" wrote:

Hi, I remember reading on of Toms Hardware tests with SATA drives, you have to change the block size to get peek perfomance, I seem to remember them mentionsing 64K blocks as opposed to 256K blocks....

Have a Google 'SATA block optimis' for it and see what it comes back with or something similar.

Thanks for the valuable tip, however I wasnt able to find where to change blocks size in the bios on my IS7, any ideas?

 

[TomG]

your Raptor is MUCH quicker than the IBM 120GXP I have on the SATA channel. My max is only your average and my average is only 38K using Diskspeed32.

 

[Myron Johnson]

"zLaTi" wrote:

You might want to try HDTach instead, since Sandra (like every women <Yes, I'm married>) often lies...

That's why we Americans prefer sheep.
They don't talk much and hardly ever say anything...baaaaaaahd.

/Myron