Do SSD disks do not last as long with Windows XP as with other OS?

Question

We do software development (Visual C++) on Windows, and already had a few cases where developers using SSD disks with Windows XP had to replace their disks after a (one, 1) year of usage, because they were broken. (Timeframe 2010 - 2012)

Obviously, compiling a lot means a very high number of writes - Visual Studio Compiler likes to write a lot of temp files in addition to all the build artifacts created by a normal C++ project anyway.

Now, I know that the TRIM command is not supported on XP, but I always understood that to be a performance thing, not a longevity thing for the drives???

Also, given that some claim that a modern(ish) SSD drive should last 51 years with full write utilization, how can it be that a developer, even doing many compilations during an 8 hour working day, can trash his SSD -- and what has this to do with Windows XP (vs. Win 7)?

Note: This is a developer shop, so naturally everyone has his own clever explanation of this and that. But this is a developer shop, so the expertise of the people here lie with SW development, and not with HW reliability.

And given all the myths about SSD disks on the net, I really have a hard time finding reliable infos on why an SSD should fail earlier (or anyway) on Windows XP ....

DISCLAIMER: Note that I do not necessarily claim that these SSD disks really broke because they were used on Windows XP with a write heavy usage pattern. I'm asking whether there is any existing evidence (because my co-workers claimed so), that an SSD drive will fail earlier on XP than otherwise (because of missing TRIM or other reason) ...

---

Having followed a few links, I would like to especially highlight the most upvoted answer for the question SSD on Windows XP. (Note that this answer (from 2010) is a quote of the article -- from 2007 (!) -- it links to. The gist of the article/answer seem to be that SSD drives can go bust with lots of write operations and that, somehow(?), XP is worse in this regard. Looking at the 51-years-claim I linked to above, the statements in this answer don't make any sense to me.

Also, there's the MS article from 2009, where the only reference to TRIM and wear is:

As an added benefit, the Trim operation can help SSDs reduce wear by eliminating the need for many merge operations to occur.

But the same article states under Flash wears out:

At some point, a flash cell simply stops working (...) If frequently updated data (e.g., a file system log file) was always stored in the same cells, those cells would wear out more quickly (...) Wear leveling logic is employed by flash controller firmware to spread out writes (...) most devices will last years under normal desktop/laptop workloads.

So, while MS seemed it worthwhile to mention additional wear-time benefit under the TRIM command section, the also mentioned the drives firmware as the main factor to maintain the drive by spreading out writes.

Really, this leaves me quite confused !?! :-)

Answer 1

One other big problem is partition alignment - Windows XP aligns to old disk style by default, but SSDs need 4KB alignment internally. Otherwise you're doubling the volume of writes because a lot of things will overlap internally. (I'm not really qualified to talk about this at extensive length, so maybe somebody can elaborate at this - it's not really my area of expertise.)

That said, I am still fairly surprised if your disks are failing this fast. SSDs can be pretty failure prone unless you're careful about what you buy - this wouldn't surprise me nearly as much from OCZ or Crucial SSDs as it would from Intel ones (which I can vouch for in dozens on XP machines with no problem).

My personal opinion? XP may not be helping, but it's probably mostly a red herring.

Answer 2

TRIM is also good for keeping the drive health up. The reason is simple: On writing a SSD swapps blocks very often for limiting the maximum count of write operations for each block.

On a SSD never trimmed the only available blocks for swapping are those from the reserved free space (blocks that are not directly accessible because the SSD has more flash memory available internally than it offers to the outside).

On a SSD well trimmed the pool of available blocks also contains those that are free on disk. Therefore the write operations can be spread among more blocks.

Answer 3

I cannot visualize how different flavors of an OS can reduce the MTBF of an SSD! The whole idea of an SSD is to increase storage reliability, access times and life expectancy of the storage medium. How many SSD's has this happened to? Have you experimented with different SSD manufacturers?