If my guess based on price and reported problem is right, that is probably an OCZ Core V2 120GB SSD. That particular SSD runs on a
JMicron JMF602B controller and does not have TRIM support. More worryingly, it does not support the
ATA_SECURE_ERASE command (
confirmed by users).
That command is an SSD user's last resort for restoring brand-new performance to an SSD. Essentially, it marks all cells as empty, thus indicating to the firmware that all cells can be written into. (Yes, all data will be erased. Surprised?) Note that this is not the same as a normal secure-erase performed by disk-wiping programs. An SSD that doesn't support TRIM or ATA_SECURE_ERASE is doomed to the scrap heap.
This is info I dug up while researching a 30GB OCZ Core V2 that I bought some time ago; most likely I will relegate it to an external enclosure when I'm done with it. The Core V2 hails from a time when the TRIM command had not been implemented in SSD firmware yet, so I'm not surprised that such issues should crop up. This issue has not been reported on the newer SSDs that support TRIM, i.e. SSDs running on the Indilinx Barefoot or Intel controllers. That's not to say they don't have their own issues, but thus far TRIM is sufficient to maintain good SSD performance for most desktop scenarios. Before you look into buying an SSD I strongly suggest reading up on them first; you can easily reduce their write life span drastically without much benefit by faithfully using them like a normal hard disk (e.g. defragmenting them).
It is also highly important that you read up on the differences between the major SSD controllers (which is the
main, but not the only, determinant of performance and reliability). Unlike with hard disks, where performance is much more uniform across different brands and models, SSDs are still a relatively new technology (despite having been around for more than a decade), and can be drastically different especially performance-wise.
Major controller makers in no particular order:
* Indilinx (of which their Barefoot series is quite well-known)
* Intel (in G1 and G2 flavours; go with G2 since
G1 has no TRIM support)
* JMicron (which has at least 2 versions,
a sucky old one and
a better new one)
* Marvell
* Samsung
* Sandforce (fresh newcomer, overturning the incumbent Intel X25 as the poster child of performance - it's pushing SATAII interface limits in both read
and write. Not recommended unless you're an early adopter and won't whine on forums about SSDs breaking).
They each have their own quirks, and are optimised rather differently, so trying to pick the best one for your needs isn't something you can google in an hour. If you don't have time to be reading up and just want a "safe" SSD*, go with Indilinx Barefoot or Intel. (*I say "safe" only because those 2 controllers have been around the longest among the most commonly available SSDs, and up to now the issues reported with them are relatively minor; who knows what 5-year reliability reports for those drives will say?)
If you don't want to spend precious time reading and don't want to fork out for an SSD known to be more reliable, do yourself and other people a favour and don't jump for an SSD just yet. Yes, you'll be doing other people a favour, because at least you won't be spreading misinformation about SSDs based on older experimental hardware that, in more well-read hardware circles, are already known to be lemons. At this point in time SSDs simply aren't idiot-proof enough that someone who knows zilch about them can go out, buy one and expect to be worry-free for a year or two. You can, if it's an SSD like the Intel X25V/X25M or OCZ Agility/Vertex (TRIM support, firmware well-optimised for IOPS as well as random read/write and relatively stable and reliable), but at the same time there're simply too many lemons out on the market that I can't advocate buying one with no prior knowledge.
