[SOLVED] What do you make of this HDD benchmark?

[Pimpom]

Shown below is the result of an HDTune benchmark test of a 1TB HDD from my daughter's new-ish Acer laptop. The laptop was always excruciatingly slow from day one. I've temporarily replaced the HDD with a new SSD and it's now blazingly fast. The difference is way beyond the normal difference between operating with an HDD and with an SSD.

SMART and scans don't find any file system, partition or surface errors. The unusual HDTune result remains the same whether the HDD is in the laptop, in my desktop or attached as an external USB drive. HDDScan shows the same results. Error Scan with HDTune proceeds at the same variations in speed as in the benchmark test.

The HDD is a WD10SPZX (SMR). I don't have another specimen of the same model to compare with but I feel that there's no way a 2.5-inch 5400 rpm HDD can run at 200+ MB/s from beginning to end. The dips in the middle to ~75-100 MB/s would be more realistic.

My next step will be to create a rescue disk, wipe the whole disk and restore everything. For now, what do you make of the test result?

 

[fzabkar]

I can't explain the weird results, but when you are ready to rebuild the drive, I would initialise it and then immediately rerun the HD Tune read benchmark against the blank drive. That's because this SMR model supports TRIM. TRIM should result in the drive's CMR cache being cleared, which then means that all reads will be from the SMR regions rather than from the CMR cache. At least that's my theory.

If you still see this strange benchmark, then perform a quick format. This should definitely invoke TRIM.

 

[Pimpom]

Noted, thanks. I've just started creating a recovery drive on a 128GB thumb drive. Progress is painfully slow but this was not unexpected, considering the reason I'm doing it. Looks as if It's going to take at least an hour and a half.

 

[Pimpom]

I started a zero fill on the HDD with Partition Wizard and since it was going to take a long time, I watched TV in the living room, occasionally checking the progress. On the third or fourth check, I found that my computer had crashed - a rare event. I don't know if it was a coincidence or if it was caused by the operation. Progress had reached 17% on the previous check.

A test with HDTune without initializing the disk (HDTune can do that) still showed a weird result but the first 36% is now much more normal than before. I'm guessing that that's how far the zero fill had gone before the crash. If my guess is correct, I take it as an encouraging sign.

It's 2:00 am here now, so I'll restart the zero fill tomorrow. Whatever the outcome, I've already decided to replace the HDD with an SSD, but it's an interesting exercise. Here's the latest HDTune result. The slightly lower speed compared to the previous test is because it was via a USB adapter this time whereas the earlier one was done while connected as an internal SATA drive. I've observed the difference before.

 

[fzabkar]

I would do an enhanced secure erase rather than zero fill. A secure erase executes internally within the drive and doesn't require any data transfers from the host. By my estimation, the whole process should take about 3 or 4 hours.

https://www.bing.com/search?q=1000+GB+/+80+MB+per+second+in+hours