Is there a way to test an SSD (similar to memtest)?

[zambz]

Hey guys, was wondering, is there a way to test an SSD to prove it is faulty (from a hardware perspective compared to a software perspective)?

 

For example, are there tools similar to memtest available for SSDs which stress test the entirety of the drive to confirm that it is fully consistent and stable?

 

I just want to know when an RMA is necessary vs a firmware problem. I have had my drive for about two and a half weeks and only found stability yesterday after a secure erase and 1.3.2 upgrade (of course I now plan to apply 1.3.3 tonight too) and am just unsure if the BSODs I experienced in the past will return or whether they are due to a real hardware issue with my drive.

 

Cheers

Zambz

[Synthohol]

why try to kill the drive early?

just speed test with ATTO and thats all u can do.

[parsec]

Hey guys, was wondering, is there a way to test an SSD to prove it is faulty (from a hardware perspective compared to a software perspective)?...

 

Cheers

Zambz

 

I am not aware of anything like this being available, that is, being specifically meant for SSDs.

 

IMO, this is a good idea. But we need to be more specific about what we want.

 

One test would be analogous to the DRAM memtest program, which would check all the NAND chips for "errors" or faults. I assume that would be at the byte level, not at the MLC or SLC cell level. Or would the block level be appropriate?

 

When testing DRAM, the testing is done usually on anywhere from say 1GB to 16GB, but could be well over ten times that for SSDs. Also given the speed of SATA interfaces, which is much slower than any DRAM memory interface, the test would take a while.

 

Since access to DRAM is accomplished in several ways, from "Northbridge" memory controller chips, to memory controllers in the CPU, the controllers are being indirectly tested, but nothing beyond being used during the test. SSD controllers would be used in a similar way while testing NAND chips, but do we also want to somehow test the SSD controllers as well? If that could be done, what are the parameters to be tested?

 

The existing programs created for checking HDDs for errors seem to work with SSDs, but I imagine they are not optimal for them at all. I wonder if they are actually almost worthless for use with SSDs. One SSD manufacture has as part of their SSD Toobox program an error checking option, but I don't know what it does, it simply reports a pass or fail.

 

But, when a SSD "does not work", which is a very general statement, what is it that is not working? Or do we want a simple pass/fail test as you implied in your post. Still, what is actually being tested to reach that determination is an interesting question.

 

A pure hardware test is likely much easier to do than anything else, but will a "bad" SSD pass its hardware test and still not function correctly?

[Synbios]

Actually check the raw value of SMART 05. As long as the reallocated sectors are still 0 then all your blocks should still be functional.

 

If you want to check sector integrity there's always chkdsk

 

Run:

chkdsk /r

 

HDtune might also be something more along the lines of what you want, though I'm not sure how well they work for SSD. Personally, I don't use HDtune.

http://www.hdtune.com/hdtune_errorscan.jpg

http://www.hdtune.com/main.html