Varying disk test performance

Any new ideas, possible impromevents for future versions.
mr-b
Posts: 28
Joined: 2014.11.02. 13:13

Varying disk test performance

Post by mr-b »

Is it possible to have a slider of Slow to Fast HD testing time vs drive performance impact?
Or are such tests controlled by drive firmware rather than an app?

It'd be handy to reduce the testing performance impact for new disks that I want to start copying data to, but still test thoroughly. Alternatively I may want to test them as fast as possible before copying over data!
User avatar
hdsentinel
Site Admin
Posts: 3128
Joined: 2008.07.27. 17:00
Location: Hungary
Contact:

Re: Varying disk test performance

Post by hdsentinel »

No, there is no such option available - and not really required as this would be not useful at all.

There are numerous tests available and they both work differently (see Help -> Hard disk tests section for more details about them).

Hardware self tests (Disk menu -> Short self test, Extended self test) run "inside" the hard disk, when the hard disk is idle, does not perform any actions, no read/write commands.
This means that these tests can be started any time, even if you actively use the hard disk during the testing for any purposes (copy/move files, antivirus scan, defragmentation, etc...) and the actual performance of the hard disk not impacted - just the test will run (usually much) longer than estimated, as the estimated time shows the test time when there is no hard disk actions during the test.
It is not possible to control priority / performance of this test, we can just start it and the hard disk firmware proceeds and results a progress indicator - and the result after the test (OK or failed by a component of the hard disk).

At the other hand, software self tests (Disk menu -> Surface test functions) are performed by Hard Disk Sentinel and the test runs in fully controlled/monitored way. I mean that during the test, possible change in hard disk health, reported error counters, retries, hardly accessible / slower areas are all detected and reported.
For this, the best is to lock the hard disk completely, to prevent any other actions (any reads/writes from other software or the OS itself), otherwise this performance testing may not be accurate, due to the background activity (any reads/writes) caused by other software, as then the hard disk would need to perform different tasks during the test.
This locking may not be required for READ type test (Disk menu -> Surface test -> Read test), but for other tests, the hard disk need to be locked, to prevent accessing of other software.
So generally it is better (if not required at all) to NOT perform any other actions during the test.
This also (by its nature) automatically ensures that the hard disk test is running at full possible speed, so will complete in shortest possible time.
And this can be easily used to verify possible bottlenecks, eg. if the hard disk/SSD transfer speed is lower than expected, for example maximized by an USB 2.0 enclosure, or if an USB 3.0 external hard disk connected to USB 2.0 slot - or if the motherboard/controller may limit the transfer speed of a SATA 6G device.


This is why the best recommend way is (exactly as http://www.hdsentinel.com/faq.php#tests suggests) to perform the tests BEFORE formatting and copying actual data on the hard disk - and then some of those tests upon problems, for example when the health decreases.

A general guide-line about tests:

1) perform the tests suggested at http://www.hdsentinel.com/faq.php#tests initially, before formatting and copying data to the hard disk

2) perform Disk menu -> Short self test / Extended self test any time (even can be configured to start scheduled, for example at once per week. See Configuration -> Operations page and the Help about creating scheduled projects)
As these does not affect performance and usually fails quickly on any (even minor) problems, if they report no problems, then we're fine

3) if the above tests fail with problem and/or if you see decreased hard disk health, you may try a Disk menu -> Read test -> Surface test, to reveal all possible problems and get further information about damaged, slower blocks

4) and if this also reports problems, the best is to perform backup and then Disk menu -> Surface test -> Reinitialise disk surface test to repair, make the hard disk better (in both terms of reported health % and usability).

mr-b wrote:Is it possible to have a slider of Slow to Fast HD testing time vs drive performance impact?
Or are such tests controlled by drive firmware rather than an app?

It'd be handy to reduce the testing performance impact for new disks that I want to start copying data to, but still test thoroughly. Alternatively I may want to test them as fast as possible before copying over data!
If the hard disk is already used, already contains data, then steps 2-4 can be still applicable.
Post Reply