Question Sudden abrupt change in HD Tune ?

[je1983]

https://imgur.com/a/CYxnetm

View: https://imgur.com/a/CYxnetm


I'm looking for some input on this so i can understand this better, the readings on my HDtune benchmark go steady at around 100MB/s up to around 800gb, then abruptly jump to 250MB/s for the remaining of the 2gb.

I've done tests with Seatools and there doesn't seem to be an issue. I did a quick error check with HD Tune, the green blocks show up at a steady pace for the first couple of lines, then also abruptly finished very quickly, at around the same 800gb area.

The drive is a little over a week old. Initial benchmarks were "normal". I did accidentally leave it on the bed and toss my phone in the direction of the drive, resulting in a minor impact, not a direct hit more of a partial glancing hit. I'm wondering if that might have anything to do with the reading and what this particular type of benchmark may mean. I've done three back to back and they are identical and always abruptly change at the same area.


EDIT: I'm currently moving files around between several drives and I'm using this drive as sort of a holding drive. I've not noticed any differences in transfer speeds so far since the impact or test, but it is only half full.

 

[helpstar]

if it is half full, maybe the test overwrites the first half then continues the other half more quickly
or
one half 900GB of the 1.82TB drive is slower than the other half

 

[fzabkar]

Your drive is an SMR model. The area beyond 900GB has been TRIM-ed, so the drive is returning zeros without actually reading the TRIM-ed sectors. 250MB/s appears to be the maximum transfer rate of the USB port or SATA interface.

A tool such as CrystalDiskInfo will report the model number of the HDD behind the USB-SATA bridge. Then you can look it up here:

https://hddscan.com/blog/2020/hdd-wd-smr.html

 

[je1983]

Your drive is an SMR model. The area beyond 900GB has been TRIM-ed, so the drive is returning zeros without actually reading the TRIM-ed sectors. 250MB/s appears to be the maximum transfer rate of the USB port or SATA interface.

A tool such as CrystalDiskInfo will report the model number of the HDD behind the USB-SATA bridge. Then you can look it up here:

https://hddscan.com/blog/2020/hdd-wd-smr.html

Thanks.

Can I ask if they are all expected to behave this way? I also have a Seagate 2tb, st2000lm007-1r8174, which from a search, is also SMR. It's around 80% full, it's a faulty drive that I've copied but I'm able to still run it and access the data. The HDtune benchmark for this drive shows up "normal" and not like the one above. Just curios since I'm planning on buying a few more in the next months.

 

[fzabkar]

I expect that the drives are delivered in a TRIM-ed state. After you fill them up, they behave normally. Also, not all SMR drives support TRIM commands.

You can check the Features reported by CrystalDiskInfo.

Edit:

TRIM doesn't appear to be supported by the st2000lm007-1r8174.

https://i.imgur.com/dYtNIcA.png

 

[je1983]

I expect that the drives are delivered in a TRIM-ed state. After you fill them up, they behave normally. Also, not all SMR drives support TRIM commands.

You can check the Features reported by CrystalDiskInfo.

Edit:

TRIM doesn't appear to be supported by the st2000lm007-1r8174.

https://i.imgur.com/dYtNIcA.png

Thanks again.

I did a check and it seems to show up on my particular drive.

https://imgur.com/a/wlM9qtd

View: https://imgur.com/a/wlM9qtd

 

[fzabkar]

I did a check and it seems to show up on my particular drive.

Different firmware version. Interesting.

BTW ...

https://www.cheadledatarecovery.co.uk/2018/02/st2000lm007-st1000lm035-worst-hdd-ever/

Be aware that not all USB enclosures can handle UNMAP commands. This is the SCSI equivalent of TRIM.

First search result ...

https://docs.oracle.com/en-us/iaas/Content/Block/Concepts/reducingbackupsizewithtrim.htm