Solid State vs Disc

[Dr-Kiev]

Sandforce has nothing to do with it. Intel uses Sandforece. I have two SSDs that are Sandforce; a G.Skill Sniper with a tera byte of data written to it and an Adata. So far they have been going strong.

SandForce is the main culprit . MCU's sf-12xx and sf-22xx families are affected . Those tons of FW updates they have released, couldn't fix issue with read\write errors on nands . Long before, Sandforce claimed , their core could deal with cheap and bad quality nands . A lot of developers including OCZ took that risk, and started to solder the cheapest nands. SSD prices became to decrease due of this . What do we have now? OCZ, who sold more than others sandforce based ssd drives, now bankrupt.

Intel released only two SSD drives based on SandForce core-Intel 520 and 330. It has opportunity to write its own FW for its ssd, nobody can, even Apple and Toshiba. And Intel also develops its own Nands chips, and can decide what quality chips to solder. Intel SSD isn't so cheap like OCZ .

 

[sequoia464]

SandForce is the main culprit . MCU's sf-12xx and sf-22xx families are affected . .

I haven't really followed anything at OCZ since Vertex3 (SandForce), do you know what the returns and failure rates have been on Vertex 4's (think this is Marvell based) and the Vector's?

 

[Dr-Kiev]

do you know what the returns and failure rates have been on Vertex 4's (think this is Marvell based) and the Vector's?

No, i don't. It is too early to discuss failure rate. But should be much better compare with older (SandForce ones) .
Marvell MCU is ARM based core, used for Micron C400/500/Crucial drives.

 

[John Connor]

i wouldn't be so quick to dismiss the "SandForce factor" with OCZ SSDs. Sure, SandForce controllers may be (relatively) safe to use, now, but back then, with immature firmware, SandForce was a pure disaster.

Both of my drives are two or more years old and I haven't had a problem...

 

[Dr-Kiev]

Both of my drives are two or more years old and I haven't had a problem...

Those one, probably, Vertex Pro drives, based on Indilinx (Barefoot) controllers.

 

[Cerb]

Both of my drives are two or more years old and I haven't had a problem...

And many others haven't, either, nor will they. Most of the core problems were compatibility problems, of some kind--they were just numerous, tough to track down, and SF was not prepared.

The problem is that things like error rates are statistical probabilities over a wide population. Get one SSD with a couple marginal dies, and there goes all the fancy error correction magic: a copy of any important data (including any usable error-correction data) really needed to be on another die, as well. Meanwhile, using better quality NAND, which is likely all that Intel, Kingston, Sandisk, et al have been doing, in comparison, with their SF drives (all but Kingston have serious fab interests, and all have made some perfectly OK SF drives), would make that a rare problem and/or EOL problem only, as the chances of related data being bad on two or more dies (assuming RAISE, or something like it) will end up being very low.

 

[sequoia464]

And many others haven't, either, nor will they. Most of the core problems were compatibility problems, of some kind--they were just numerous, tough to track down, and SF was not prepared.
.

If I remember correctly, one of the things that bothered me the most was the problem that the Vertex 2's had with not waking up from sleep mode. I think this was eventually corrected, but by then I had retired my Vertex's and was using M4's