Possible reason to avoid OCZ?

@mourningstar62, davepermen: You guys are taking chimpanzee's comment out of context. What he's saying is that it's a waste to store large multimedia files (like 1080p video) on an SSD since a modern HDD is just as fast and much cheaper.

Which I totally agree with. I wouldn't even have brought it up if not for the fact that the multimedia world is the best place to find real-world data on uncompressible writes and also one of the few situations where the Sandforce TRIM bug really hurts.

I guess the moral of the story is that while Sandforce 1XXX drives take a heavy sequential write performance hit once all blocks have been written to, it doesn't really matter in the end because few people uses SSDs for large sequential transfers.

 

That's a point I forgot about. RAID arrays and TRIM don't exactly work together right now.

 

Has OCZ responded to this issue yet?

 

I thought Intel RAID passes long TRIM command?

 

I don't think so.

 

Not sure if this was posted before but here's OCZ's reaction.

Guide 34nm Spectek Nand...see here for specs and info

 

The controller passes TRIM commands to those drives that are not in the RAID array. That's it.

Still doesn't explain Micron's reaction if true. Like I said, I hope the legal fireworks fly on this one because then it will all come out and we'll know the truth. OCZ is a scum company and they've been caught red-handed several times. Unless proven otherwise I'm just going to believe that they were caught in the cookie jar again.

They could have focused on their efforts and their documentation showing that OWC was wrong and taken the high road. Instead, they insulted the competition. So even though they might have been the wronged party initially, their PR team just made a bad move and took the low road.

Either way it doesn't matter for me. OCZ will never enter this household again and I decided that months ago.

 

As far as I know the reaction isn't coming from their PR team, it's from a technical support specialist.

 

They are a group of funny people, full of self cofidence though many times do not know what they are talking about or inconsistent(the 25nm fiasco is another case)

OWC is not completely wrong but is also pushing a bit too far in the challenge.

What happens seems to be that those chips are branded under Spectek which is created to sell bins that fails Micron's standard sample screening process. Those chips then got tested again to see which is truly 'not up to standard' and re-bin. Having to setup a seperate brand for it means Micron doesn't want to use its brand.

The closest analogy I can think of is refurbish materials. I would say they should be perfectly fine and most likely cheaper for OCZ.

 

raid is dead to me since i touched my first ssd. a useless complication it is, nothing else. (and don't talk about raid to me to make reduntant, storage. i have winhomeserver with DE, surpasses EVERY raid solution)

 

SSD is actually multi level RAID 0. And I don't think window home server is in the same league as RAID. RAID is complicated(and it should only be used in corporate environment with good planning) but that is the only solution so far to ensure availability of a storage device.

 

ssd is for me not raid. don't care how it works internally (i know, of course). for me, it's plug in and works.

home server is easy as pie. plug in new hdd, storage is expanded, and storage for those files that i care is always secured onto two drives.

so no, raid is NOT the only solution so far to ensure availability of a storage device. and it's not the best one (too complicated, too slow to manipulate).

DE is the best solution for cheap save storage. ssd is the best for fast storage. combining ssd + DE (windows home server drive extender) is thus the best overall: fast storage, and cheap secure storage.

 

Too much smoke around them, if you know what I mean. Not just recently, and not just since they entered the SSD market either.

While we're in the same neighborhood, I don't really like sandforce controllers either, though not because I think the company is shady. I inherently don't trust algorithms that do anything less than write my exact data. (copy on write is good for example, dedup is not)

Compression is something you let the file system or programs/user control, I don't trust it being done defacto to all data behind the scenes. Its really nice for transmitting/storing video and whatever, but not for keeping safe a copy of my documents.

Dedup is something for SANs where that "single copy" is also protected by some sort of parity and the architecture of the file system. (mmm snapshots)

But thats just me, people can buy what they like and live with whatever happens in the long haul, I won't bother them. They don't have to post some long rambling or angry defense of their choice either

 

@Aluminum

I think in the same line. It takes years for a file system to get stablized and there is no reason to believe SF engineers are super smart that they can beat all those engineers of Sun Micro, IBM, SGI and contributors of linux file system etc. and have the problem solved in a year or so.

 

chimpanzee,

Yeah!!!

Those young SF engineers - what? Next year, I think they'll all be out of their teens finally. The arrogance of it all!

(I really do agree with you though...).