Momentus XT, Momentus, Momentus Thin, MobileMax, and other notebook drives

[AntonK]

Hi all,

I've got a new hard disk Momentus 500 GB (ST9500423AS) to replace 2-years old WD Scorpio Blue 320 GB (WD3200BEVT-22ZCT0, 5400rpm, 8MB) in the laptop Lenovo B550 running Windows XP SP3 32-bit.

After copying all partitions (the first system partition was copied as is - without resizing) I made some benchmarking with CrystalDiskMark 3.0.1... and unexpectedly got very bad results in writing - the writing speed was awfully low and the new HDD (despite 7200rpm, 16MB) turned out to be slower in writing than the old WD.

Here are the results of the Seagate Momentus:

Sequential Read : 111.219 MB/s
Sequential Write : 15.284 MB/s
Random Read 512KB : 40.616 MB/s
Random Write 512KB : 11.625 MB/s
Random Read 4KB (QD=1) : 0.547 MB/s [ 133.4 IOPS]
Random Write 4KB (QD=1) : 0.259 MB/s [ 63.2 IOPS]
Random Read 4KB (QD=32) : 0.997 MB/s [ 243.4 IOPS]
Random Write 4KB (QD=32) : 0.307 MB/s [ 75.0 IOPS]

The results of the WD Scorpio Blue:

Sequential Read : 31.416 MB/s
Sequential Write : 23.817 MB/s
Random Read 512KB : 17.128 MB/s
Random Write 512KB : 16.904 MB/s
Random Read 4KB (QD=1) : 0.344 MB/s [ 84.0 IOPS]
Random Write 4KB (QD=1) : 0.295 MB/s [ 72.1 IOPS]
Random Read 4KB (QD=32) : 0.891 MB/s [ 217.5 IOPS]
Random Write 4KB (QD=32) : 0.692 MB/s [ 169.0 IOPS]

More details:

Test : 1000 MB [C: 48.4% (13.2/27.3 GB)] (x5)
OS : Windows XP Professional SP3 [5.1 Build 2600] (x86)
 

In both cases the disk was connected directly to the laptop. The write caching was disabled in the system policy (when the caching is enabled, the results look better but still slower if compared to the old HDD).

The firmware: 0001SDM5.

Any comments? Has anyone encountered similar troubles?

Best regards,

Anton
 

[WiseDrive]

Do a fresh install, ensure the drivers are all new.

Are you running SATA 6gbs? or earlier?

Crystalmarks do not give a good indication --- condition the drive first.

[AntonK]

Thanks for the hints.

I installed Window 7 Pro 32-bit on the same partition I had used for benchmarking, and after I have cleaned it and defragmented the results started to look as bad as they were in Windows XP:

Sequential Read : 109.215 MB/s
Sequential Write : 15.375 MB/s
Random Read 512KB : 38.004 MB/s
Random Write 512KB : 11.311 MB/s
Random Read 4KB (QD=1) : 0.512 MB/s [ 125.0 IOPS]
Random Write 4KB (QD=1) : 0.249 MB/s [ 60.8 IOPS]
Random Read 4KB (QD=32) : 0.947 MB/s [ 231.3 IOPS]
Random Write 4KB (QD=32) : 0.307 MB/s [ 74.9 IOPS]
Test : 1000 MB [C: 38.8% (10.6/27.3 GB)] (x5)
OS : Windows 7 [6.1 Build 7600] (x86)

BTW In my original post I mentioned the results of WD Scorpio Blue for the wrong partition. Actually they look as follows:

Sequential Read : 61.608 MB/s
Sequential Write : 36.984 MB/s
Random Read 512KB : 27.347 MB/s
Random Write 512KB : 23.006 MB/s
Random Read 4KB (QD=1) : 0.397 MB/s [ 97.0 IOPS]
Random Write 4KB (QD=1) : 0.342 MB/s [ 83.4 IOPS]
Random Read 4KB (QD=32) : 0.991 MB/s [ 241.9 IOPS]
Random Write 4KB (QD=32) : 0.763 MB/s [ 186.3 IOPS]
Test : 1000 MB [C: 48.4% (13.2/27.3 GB)] (x5)
OS : Windows XP Professional SP3 [5.1 Build 2600] (x86)

I'm not sure about drivers version, but apparently there were no improvements in them between Windows XP and Windows 7.

The drive works in SATA/300 transfer mode.

What else can be done to improve the writing speed?

If CrystalDiskMark is not a representative utility to benchmark a drive, how the performance can be measured/compared in an objective (quantative) way? The system just runs a bit slower, but it is just a feeling...

[AntonK]

It seems the mentioned slow writing speed is caused by the unaligned sectors (the partition was created long ago in Windows XP).

I've encountered similar issue with the new WD Scorpio Black 500GB (7200rpm, 16MB), which was rather slow after I cloned the system partition to it, but after running the Align Tool the drive's writing speed increased 2 (two!) times (it was a miracle indeed! .

Is anyone aware of any free tools to perform the sectors alignment for a Seagate drives?

[WiseDrive]

Partition alignment is a major issue in Win XP but startiing with Vista and then Win 7, it is automatically done. Hence the fresh install recommendation. It fixes many issues, including bad drivers, alignment, etc. However, people INSIST on cloning, no matter what the consequences could be.

[AntonK]

That is the point! Unaligned sectors are the cause of the slow writing speed.

Doing a fresh install of Windows 7 on a clean drive will probably solve the problem, however this option is not acceptable for me: I prefer to keep using Windows XP at least until April 2014.

So I need a tool which is able to align the sectors after I clone my old partitions onto the new drive. 

That's very strange that Seagate doesn't provide any tools for that, thus saying to Windows XP users: we don't care...

[Cantbecanit]

Anton

Fresh install or face many wasted hours messing about.

[AntonK]

1. Some of my software doesn't work on Windows 7/Vista, that is why the fresh install option isn't acceptable for me.

2. I continue this discussion to discover other options, and let other users be aware of the issue they may face when buying Seagate's drives.

PS: I've already replaced the Seagate's drrive a similar one from Western Digital (which provides free Align Tool) , and I'm completely satisfied with its characterstics.

[KrypteX2]

AntonK,

I'd want to clear up some things for you, maybe you'll manage to do a correct cloning + get normal functioning of the HDD.

1. ST9500423AS seems to be a 4k sector drive. Check here:

http://www.seagate.com/files/staticfiles/docs/pdf/datasheet/disc/momentus-family-ds1701.7-1109-us.pd...

and here:

http://www.seagate.com/files/staticfiles/support/docs/manual/notebook/momentus/Momentus%205400_7200/...

Create a backup file with Acronis 2011 or 2012 of your partitions. Then Restore from it one partition at a time, on the CLEAN 7200.5 drive (no partitions). The first partition should have a 1 MB offset before it. This will ensure a 4K alignement on the new drive for all following restored partitions.

Attention: If you clone the whole drive directly, it will NOT align it to 4K, keeping it misaligned as on your old drive (512-byte sectors) ! You need to first do the backup file (.tib file) and then restore from it one partition at a time.

2. Does it have a 0005SDM1 or 0001SDM5 firmware ? If it is the former, I think it is still a 512-byte sector drive, hence 7200.4. If you check carefully the pdfs, you'll see that this model comes both in 512 and 4K sectors (7200.4 and 7200.5).

If it is the 512-byte, 7200.4 drive, then 0005SDM1 is your latest firmware (I had a 7200.4 drive with 0002SDM1 firmware, updated to 0005SDM1 and worked much better)

If it is a 7200.5 drive, I don't think you'll find a firmware update.

3. Try to run a Checkdisk on the C partition before backing up and restoring your partitions.

4. Install the proper SATA AHCI drivers before cloning.

5. I've found that Seagate has SMART Align on this drive if it has 4K sectors. I think it checks where are the partition boundaries and realignes them logically... however, if you full-clone a drive (aka sector-by-sector), SMART Align will not realise the position of the partitions...). That's why you need to restore one partition at a time, aligned to 4K , I think Acronis actually CREATES each partition on the new drive (aligned to 4K) and then copies back the files from the partition image file to this newly created partition (Smart Align might be aware of it this time).

I personally think that SMART Align is not so smart after all....

Best wishes,

KrypteX.

[AntonK]

KrypteX2,

 thanks for the hints. They are useful in general.

In the same time I have some doubts about the restoring approach...

> Create a backup file with Acronis 2011 or 2012 of your partitions. Then Restore from it one partition at a time, on the CLEAN 7200.5 drive (no partitions). The first partition should have a 1 MB offset before it. This will ensure a 4K alignement on the new drive for all following restored partitions.

I guess, restoring a partition to a different offset will eventually make the file system unusable...

Anyway I'll try this approach some day, if I have enough free time for that