TR Forums

My M.2 SATA SSD drive worked well until a few days ago. Now, the speed varies wildly:



However, the speeds reported by Crystal Disk Mark seem normal to me:



I wasn't running anything else when I ran these tests. I have this problem under Windows and Linux. smartctl and badblocks detect no problems, although the latter runs very slowly. WD Dashboard doesn't find anything. I've run TRIM, removed and reinstalled the drive, etc.

Any suggestions as to whether the drive is defective, or whether something else might be the problem? As I mentioned, the problem appeared suddenly.

Any chance it is thermally throttling?

Are you seeing performance anomalies in normal use, or just in HD Tune?

The drive's temperature is 37 C (99 F), according to HWInfo. Running HD Tune and Crystal Disk Mark simultaneously, the highest temperature was 45 C (113 F). The drive has my games, and I first noticed the problem when a game was taking entremely long to load. The slow speed was very noticeable, and that's why I started exploring.

I've also been having other problems (Windows hanging or rebooting), which I blamed on Windows, but now, after checking everything else, and especially because of the new problem with the SSD, the only possible explanation I have left is a defective motherboard (which worked well for two years). I've tried reflashing the BIOS, by the way.

Have you tried replacing the SATA cable? Worth a shot (and much cheaper than replacing a mobo or ssd).

M.2, so no cable.

Yan wrote:

M.2, so no cable.

I would reseat the drive in the slot then.

Have you installed WDC's utilities, how does the smart data, drive health look?

When was the last time you trim'ed the drive?

Are your chipset drivers up to date?

Is the drive firmware up to date?

To answer some specific questions:

I tried removing and reinstalling the drive. WD's Dashboard utility reports that everything is fine, that "Life Remaining" is 100 %, that the firmware is up to date, that the "Media Wear Out Indicator" is 0.32%, etc. I ran TRIM today and yesterday and the day before that. As far as I can tell, all drivers are up to date.

Now, a more general observation. I was having problems with Windows hanging or rebooting. I was blaming Windows, since I had no problem in Linux, and I was reinstalling Windows. Then, the problem with the drive appeared, and since the drive was slow also under Linux, I began to think that the problem must be the hardware. I've checked the memory, the CPU, the physical health of the drive itself, and I can find no problem. I'm about to conclude that the motherboard must be defective, but if so, it's defective in a way I haven't seen before.

Use the error scan to read the entire disk, see what the speed map looks like.

Then run it again and see if it's different.

(reading stale data on an SSD forces it to rewrite it to fresh cells, it's possible the firmware on that drive is dumb/old and doesn't do this in the background when powered)

I did a "quick" error scan using HD Tune. Then I started a surface scan using Windows' CHKDSK, but I gave up after 15% because it was so slow.



The result is very similar to the first one, and there's no improvement at the beginning where I did do a surface scan.

Now I'm seeing some weirdness with my other disks (the access time test only does the first half of the disk; with the error scan, the "position" first goes up and then gets into negative numbers).

If it affected only Windows, it could be a driver problem, but since it also affects Linux, I can't see how it could be anything other than a bad motherboard.

Quick doesn't read every sector, or even close to it.

After a full error scan, the result is the same:

So to be clear:

The drive is slow in both windows and Linux?

or

the drive is slow in windows but fine in Linux?

The drive is slow under both Windows and Linux.

It guess it's clear there's a hardware problem. The question is whether the only problem is with the SSD, or whether I'm right to suspect the motherboard because I'm having simultaneously several problems that aren't obviously related. (The system hang and rebooted under Windows, now my other drives are starting to show problems, etc.)

Is the SSD your boot drive or a data drive? If it is your boot drive it can cause all sorts of strange system behavior.

The drive is a data drive; my games are on it.

I've removed my M.2 drive, reinstalled Windows, and HD Tune now reports problems with my mechanical drives. The access time tests only cover half the disk, and the error scan starts giving negative numbers for the position after about half the test:





I'd like to do a similar test under Linux, but I can't find an equivalent to HD Tune.

Since the problems continue after removing the M.2 drive, and since the RAM and CPU seem fine, the only possible problem left is the motherboard, I think.

The negative numbers look more like a problem with HD Tune. From the changelog on their site it looks like you need v4.00 or later to support drives 2TB and up, and you're running 2.55 (which was released all the way back in 2008)?

I've finally found the problem and a solution and I wanted to mention it.

HDDScan clearly showed the problem, with some blocks taking over 500 ms to read. I've found that I can "repair" these blocks using badblocks with the non-destructive read-write test. (And also presumably with the destructive test.)

I'm not sure whether this means that the drive is dying anyway or whether the "repairs" are permanent.

As for the computer hanging, I finally determined that the problem was caused by the AMD Adrenalin driver, specifically the AMD external events utility service. It apparently caused the computer to hang whenever the monitor was turned off after a period of inactivity. Disabling the service solved the problem. It seems to be the problem also mentioned here. I've filed a bug report.

Thanks to everybody!

Still sounds like the drive is starting to fail. If blocks are "fading" to the point where the error correction logic is frequently taking 500 ms to recover the data, the flash is not retaining data reliably.

Crucial has a decent 1tb ssd you can get as replacement for this one.

once you have cloned the drive you can see if the problem is still there.

If it is not you have found the source of the issue.
If it is then you can start looking elsewhere such as mainboard.

Aranarth wrote:

Crucial has a decent 1tb ssd you can get as replacement for this one.

Do you mean the MX500?