Weird HDTune results?

[cpals]

Ran HD Tune on my two drives and get these results... from what I've seen online, it looks like the graph should be smooth and not choppy like mine. Bad hard drive or normal results?

 

[VirtualLarry]

Not every HD graph is smooth. In fact, using HDTach 2.61, my graphs are rather jagged.

I miss HDTach 2.61 (no longer being developed), because you could use it as a graphical diagnostic of a HD's health. The more jagged drops in the graph, the more likely the HD is to go out.

 

[sub.mesa]

The more jagged drops in the graph, the more likely the HD is to go out.

Sorry but there's no basis for that kind of remark. In fact, RAID arrays and other complex setups often end up with a edged 'sawtooth' pattern on HDTune, which is the result of a low queue depth of just 1 when reading from the device or RAID array. It's a signal that results can be much higher if multiple queue depth used; so a lack of queue depth can cause sawtooth patterns. But the two screenshots here are not sawtooth.

The first HDTune screenshot looks great, the second looks as if it's your system/boot disk; can that be right? If it is, the higher seeks and fluctuating speeds are normal. You are benchmarking a disk you're running your system on; any background I/O will lower the HDTune scores. Particularly, you can see the yellow dots being much more fluctuating.

Only the second HDTune screenshot would be suspicious if it isn't your boot/system disk, which i think it is. If it's not, show us the SMART data as well, located on the Health tab. If you do, make sure to enlarge the box so we can see all SMART values.

The Files benchmark in HDTune Pro does test with higher queue depth, which can be used on RAID arrays. Other benchmarks i recommend are CrystalDiskMark and AS SSD.

 

[VirtualLarry]

Sorry but there's no basis for that kind of remark. In fact, RAID arrays and other complex setups often end up with a edged 'sawtooth' pattern on HDTune

I said HDTach. And I wasn't talking about a normal sawtooth-like pattern, I was talking about spots that show a noticable performance degradation, which can indicate problems reading that portion of the disk.

 

[sub.mesa]

That's just background I/O done in all likeliness. It may even be his C-drive in which case it's only logical you would see fluctuating results and some much higher seek rates. If a pristine benchmark gives clean results, then there's nothing wrong with the drive.

SMART values also check the amount of time required to seek, for example. If that time gets higher, SMART values will lower towards the threshold. When they go lower than the threshold value (20 for example) then it is considered 'failed' or 'prone to failure'.