Hard Disk Monitoring

Hi all, I'm using HD Sentinel on two new WD Easystore 14TB drives. I'm running a Surface Test against both drives at the same time. I started the test about 18 hours ago, as I'm expecting the test to take a few days to complete.

I've noticed today that the write speed has been dropping gradually since the test started on both drives. Both drives started around 220MB/sec, but both have now dropped to around 127MB/sec. The drop was steady over the 18 hours (>5MB/hr). At first I thought temperatures might be related, but since I started I have put a fan in front of both drives which has helped cool them down to normal operating temps (55C to 46C). There was no difference in performance when I started using the fan.

So far, everything has come up as light green across both drives, so I don't think the drives are at fault here. My concern is that the speed is going to continue dropping to a point where the test will never fully finish. Is this normal behavior?

Generally it is normal and expected that the speed is dropping gradually on any hard disk drive: we usually see approx. the half of the speed at the "end" of the disk drive compared to the beginning.

Please check the Screenshots page:
https://www.hdsentinel.com/hard_disk_se ... creenshots

The 5th image in the first column shows a similar example, when the speed slightly changed from over 150 MByte/sec (at the beginning) to approx. 70 MByte/sec (at the end).

This is normal behaviour caused by the internal characteristics of the hard disk drive.

I was about to open a thread about this as well, so let's chime in here...

While it is generally true that spinning hard disks slow down on read and write speeds from the outer rings to the inner ones for pure geometric reasons, I have observed generally slow writes in the last couple of weeks as well (currently testing a lot of drives). Using version 5.70 Pro on an isolated W7 hardware test machine.

For example, I'm currently writing sector numbers to a Toshiba DT01ACA200. While these can read the entire 2TB within 5 hours (and do so in HDS), the current write cycle is estimated at 13.5h (13% after 105 minutes), so roughly 40MB/s. That is way too slow and other tools like h2testw that can also write basically the entire disk do so at much faster speeds very close to read speeds. I have tested other options as I was worried that the previously used random patterns are for whatever reason slowing down the process at large parallel write operations on up to 10 disks at a time; that isn't the case. Same with a zero pattern. Using the device non-exclusively also does not change write speeds. I have observed this on two additional machines with onboard Intel SATA ports (C602) as well as LSI SAS controllers (SAS2008) now via swapping their OS disk. Different cables, different mounting options. All slower than they should be with the current Toshiba drives only reaching a quarter of the expected transfer rate. Others were like at half or 2/3 of what they should deliver. Strangely enough, the decrease in transfer rate is exactly the same as on any other drive, meaning it's not like capped on a certain value over the entire range of the disk, but rather the disk is 1/4 the expected speed on the fast outer area, and also 1/4 of the speed of the inner area with the typical decline linked above.

Any idea what's causing this? I haven't deleted and reinstalled Windows and/or HDS yet, that's one of the few constants over the entire range of tests.

Write and subsequent read cycle, same drive.

Generally this is independent from Hard Disk Sentinel (so a re-install will probably does not change) but can be caused by the disk controller or its installed driver.
I'd check if there is an updated driver possible first.

There are some changes specifically for some controllers in recent versions, so I suggest to try the latest beta 5.70.8 just to verify if there is any difference.
Also I suggest to use Report menu -> Send test report to developer option as then I can check the actual disk controller and its driver - may give some thoughts.

With other words, if you'd use a different connection for the drive (eg. USB 3.0 enclosure/dock if currently connected to an internal port) the situation may be different.