[MagicAndre1981]

My Score is 4.7:



But the confusing part is the HDD score. I have a Samsung F3 1T, but get a score 7.4

How can this be?

[puntoMX]

MagicAndre1981, on 25 June 2010 - 10:15 AM, said:

But the confusing part is the HDD score. I have a Samsung F3 1T, but get a score 7.4

How can this be?

I had that too, the AMD ACHI drivers are fooling windows 7 . I see now after some updates that it's back to 5.9.

[MagicAndre1981]

which AHCI driver do you use? Can you post a link?

[puntoMX]

It's just the latest AMD driver version 1.2.0.125, I got it from the Gigabyte site but it should be the same from AMD directly.

[MagicAndre1981]

I'm using the same driver.

Here is the output from the XML:

  <AvgThroughput kind="Sequential Read" units="MB/s" ioSize="65536" score="7.0">135.57187</AvgThroughput> 
  <AvgThroughput kind="Random Read" units="MB/s" ioSize="16384" score="7.9">2426.07007</AvgThroughput> 
  <Responsiveness Kind="AverageIORate" units="ms/IO" score="5.3" factor="0.0">5.38000</Responsiveness> 
  <Responsiveness Kind="GroupedIOs" units="units" score="6.4" factor="0.0">13.52123</Responsiveness> 
  <Responsiveness Kind="LongIOs" units="units" score="5.8" factor="0.0">22.39657</Responsiveness> 
  <Responsiveness Kind="Overall" units="units" score="5.9" factor="0.0">302.82925</Responsiveness> 

Look at the random Read value

[cluberti]

Not sure what's happening there - what is the specific drive / model # / serial # and what is the specific motherboard chipset, driver version, and connection type (assuming SATA)?

[MagicAndre1981]

Samsung SpinPoint F3 1000GB, SATA II (3Gb/s) (HD103SJ) used as AHCI. My mainboard has an AMD SB710 southbridge.

I've installed the latest AMD AiO driver.

[cluberti]

Hmmm - honestly not sure unless Samsung has updated the drive FW to specifically score high in Winsat - can you paste the screen output and provide a log file from running winstat disk -v -log <path and filename of logfile>.log? I doubt it's the AMD controller in this case, but...

[MagicAndre1981]

I can't upload the log.

So I've placed it on my SkyDrive:

// see next post.

This post has been edited by MagicAndre1981: 02 July 2010 - 04:16 PM

[cluberti]

What about the screen output? The log doesn't capture all of it...

[MagicAndre1981]

d'oh. Didn't know that the log doesn't include the data.

I've redirected the output to the txt file:

http://cid-128fc5186...ublic/msfn/logs

But the values are now much lower.

[puntoMX]

Quote

> Disk Random 16.0 Read 3.15 MB/s 4.7

That's about right, so, what did you do to get this output now? I think it was some crazy buffering in RAM.

[MagicAndre1981]

I run the WinSAT tool again and now get this strange value again:

> WinSAT is Official : Yes
> Laufzeit 00:00:00.00
> Wird ausgef�hrt: Speicherbewertung '-seq -read -n 0'
> Laufzeit 00:00:11.00
> Wird ausgef�hrt: Speicherbewertung '-ran -read -n 0'
> Laufzeit 00:00:20.16
> Wird ausgef�hrt: Speicherbewertung '-scen 2009 -drive C:'
> Laufzeit 00:01:16.26
> Wird ausgef�hrt: Speicherbewertung '-seq -write -drive C:'
> Laufzeit 00:00:06.98
> Wird ausgef�hrt: Speicherbewertung '-flush -drive C: -seq'
> Laufzeit 00:00:04.51
> Wird ausgef�hrt: Speicherbewertung '-flush -drive C: -ran'
> Laufzeit 00:00:08.97
> Wird ausgef�hrt: Speicherbewertung '-hybrid -ran -read -n 0 -ransize 4096'
> Laufzeit 00:00:00.20
> Wird ausgef�hrt: Speicherbewertung '-hybrid -ran -read -n 0 -ransize 16384'
> Laufzeit 00:00:00.02
> Disk Sequential 64.0 Read 133.42 MB/s 7.0
> Disk Random 16.0 Read 2409.10 MB/s 7.9
> Responsiveness: Average IO Rate 3.38 ms/IO 6.1
> Responsiveness: Grouped IOs 10.84 units 7.0
> Responsiveness: Long IOs 13.89 units 7.1
> Responsiveness: Overall 150.53 units 6.6
> Responsiveness: PenaltyFactor 0.0
> Disk Sequential 64.0 Write 211.06 MB/s 7.3
> Durchschnittliche Lesezeit mit sequenziellen Schreibvorg„ngen4.293 ms 6.2
> Latenz: 95. Perzentil 16.012 ms 4.9
> Latenz: Maximum 308.790 ms 6.8
> Durchschnittliche Lesezeit bei zufallsgesteuerten Schreibvorg„ngen5.637 ms 5.8

[cluberti]

Either it really is getting that much data moved during the test, or perhaps the firmware's been tuned to perform better on WinSat tests? How old is that drive?

[MagicAndre1981]

How do I see which firmware I use and how old the drive is? Do you know good tools for this?


// Edit
I used HD Tune to get the Data:

Firmware 1AJ100E4 (if this the one with the AMD850 southbridge issues? But I use a 750). I purchased the drive by the end of last year.

//Edit 2
made an update to 1AJ100E5

I'll make new tests. Maybe the new firmware fixes this curious things

This post has been edited by MagicAndre1981: 13 July 2010 - 02:52 PM

[Tripredacus]

Is it at all possible that, at some point during the test, it is reading from the Cache or from memory instead of the physical disk? That would explain a speed increase at least.

I moved all these posts out of the WEI topic.

[MagicAndre1981]

I have no idea left. I updated the BIOS, changed the cable, flashed the Samsung Firmware, installed the latest AHCI driver (1.2.1.197 from Catalyst 10.7).

- <DiskMetrics>
  <AvgThroughput kind="Sequential Read" units="MB/s" ioSize="65536" score="7.0">135.57187</AvgThroughput> 
  <AvgThroughput kind="Random Read" units="MB/s" ioSize="16384" score="7.9">2426.07007</AvgThroughput> <---------------------------
  <Responsiveness Kind="AverageIORate" units="ms/IO" score="5.3" factor="0.0">5.38000</Responsiveness> 
  <Responsiveness Kind="GroupedIOs" units="units" score="6.4" factor="0.0">13.52123</Responsiveness> 
  <Responsiveness Kind="LongIOs" units="units" score="5.8" factor="0.0">22.39657</Responsiveness> 
  <Responsiveness Kind="Overall" units="units" score="5.9" factor="0.0">302.82925</Responsiveness> 
  <Responsiveness Kind="Cap" Reason="FAILED">TRUE</Responsiveness> 
  </DiskMetrics>

When I run the WinSAT command from comandline and redirect the output, the values are lower.

I think this is a bug in WinSAT. I'll ignore it now.

This post has been edited by MagicAndre1981: 30 July 2010 - 08:38 AM

[MagicAndre1981]

the issue is now gone. When I now run WinSAT the highest value is 5.9 again. Maybe an update fixed it.