• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

* Do not buy: OCz Vertex 2 and Agility 2 - Beware!

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Just a nickname

Member
Joined
Aug 28, 2010
I think it wouldn't be a bad idea to post this as no one did.

Many on these forums including myself have recommended OCZ Vertex 2 in the past.

This is just a head's-up that the current model Vertex 2 (25nm version) is noticeably diminished in write speed, lifespan, and capacity than the previous 34nm version. This affects models 180GB or smaller.

The speed is slower for writing non-sequential data that is not compressible (fallen from 130ish to 80ish MB/s). The lifespan has fallen from 5000 P/E cycles to 3000. The unformatted capacity has fallen by about 4-5GB (for example, a 120GB model is reduced to 115GB, a 4% loss). Obviously this means a larger percentage is lost from a smaller drive (60GB to 56GB is a 9% loss!).

OCZ is indicating that none of this could be helped, as the NAND manufacturer is forcing 25nm on them. I agree. However, particularly damning is that there was no model change, no warning, no decrease in the advertised capacity, and rumor is they are actively deleting complaints on their forums.

http://www.ocztechnologyforum.com/fo...-NAND.-Q-amp-A

http://www.guru3d.com/news/ocz-verte...ported-slower/

There is no visible method for determining whether your Vertex/Agility 2 is 25nm or 34nm other than opening the case, a warranty-ending move. Perhaps it can be checked by way of reported capacity. Does anyone know what the formatted capacity is in OS X of a drive with 120GB unformatted capacity?

source: http://forums.macrumors.com/showthread.php?t=1095121
 
Last edited:
http://www.ocztechnologyforum.com/f...8-Vertex-2-(Extended)-Poor-Performance-25nm-2

Wowzers is all I can say after wrapping up the benchmarks. To give a taste on IOMeter sequential R/W alone...

34nm 60GB Vertex 2= 248/246 R/W
25nm 60GB Vertex 2= 182/215 R/W

At this point it is blatent false advertising if OCZ doesn't change its marketing information on all sites they sell their drives. They still claim it has "285/275MB/s" speeds and a 60GB capacity which is just lying plain and simple.

source: http://forums.storagereview.com/ind...-you-need-to-know/page__p__267009#entry267009

http://forums.macrumors.com/showthread.php?t=1095121
 
Hmm... The performance drop was bigger than I had expected. And that OCZ doesn't differentiate between 3xnm and 2xnm models is bad. I suspect we will see a revision of the controller to fix the performance issues. That is what Indilinx had to do to get performance on IMTF's 3xnm stuff up to par.
 
Major revision to the silicon that revises write performance and capacity down without notice. These guys are in hot water. Their thread has gotten hotter, rather than cooling down. It's a pity. I was looking at their products earlier today for a purchase about a month out. I guess I'll be waiting to see what falls out of this situation.
 
I bought two OCZ V2s about a month ago. One is the new variety. It would be a HUGE hassle to change it back out for a replacement 3Xnm part. OCZ can't afford to pay me for my time and the trouble.

A huge screw-up by OCZ, I just dropped them off my acceptable manufacturer list. :mad:

EDIT: They've closed the OCZ forum thread.
 
Last edited:
Same here. I've been looking at their SSDs hoping for a good price to bite but after what I've read, I'll look at another manufacturer.
 
Hmm... is this means its time to hunt for new but old stock of SF based ssd ? They might still using 34nm flash.
 
Thanks for posting this J.A.N. (Just a Nickname)!
Keep an eye out on the OCF front page, as we'll have an article posted shortly.

Thanks folks! :thup:
 
This is seems to be a super-sensitive issue for OCZ and to be honest, I'm a little surprised at how they're handling it. They're being rather aggressive against the consumers, though I must admit that after reading the forums, it seems like most of the people complaining are utter idiots who just want to troll.

I understand that the process shrink was forced on them, but I'm sure they tested the new drives before releasing it to the market. I would be extremely surprised if they didn't run benchmarks and notice the performance difference, but its worth looking at from their perspective:

NAND prices are going up at 34nm, and their suppliers are effectively forcing them to use 25nm. Now with no power over the situation, they see that their drive performance has declined. Saying "hey guys, our product is now worse than it was before sorry" is a terrible PR move, and they'd probably lose just as many sales. It is definitely a sketchy move not to disclose it, but for the large majority of users, the write speed degradation will not make a noticeable difference...

Most people buy SSDs for the low seek times and for quick boots/application loads. I knew very well when buying my SSD that program installation times, for example, might not be super fast. They're better than mechanical drives for sure, but at that point who cares if I have 180MB/s or 220MB/s write bandwidth.

People are also complaining that the drive lifetime is going to decrease because the NAND shrink causes the NAND to have a lower lifetime. This debate was huge when SSDs started coming out, and everyone concluded that the wear leveling means its *extremely* unlikely that you use up all of the read/write cycles. Lowering the rated write/erase cycles on the NAND flash will not effectively reduce the lifetime because the wear leveling will still keep you far from the limit.

On formatted drive capacity: people really need to stop whining about this. I bought an 80GB SSD, and have <74GB in Windows. This is less (albeit slightly) than what you get in the conversion from decimal to binary storage sizes. Any time you get an SSD, you'll get less storage than what's advertised. This is how its always been. For the people who expected 120GB (in all fairness, thats what the formatted capacity was before), and got 115GB formatted, they're offering refunds or they'll add NAND chips to the drive AT COST to bring the capacity back up. This is the new product they're offering. If you don't like it, you can buy another drive.

Yes, OCZ didn't handle this situation the best possible way. In fact you can hardly say they handled it well. However, the people complaining on the forums continue to complain about things that are either trivial in a practical sense or that really don't matter because they can just send the drive back if they're not satisfied. Given that OCZ is now offering a myriad of solutions to the consumers, I think its ridiculous that people continue to complain to the degree that they are...
 
I like this part:

For those customers that have already purchased an “E” part and wish to move to a 32 Gbit die
based drive solution, OCZ will offer a program in which customers can trade in their “E” drive
and receive a credit towards the more expensive 32 Gbit die‐based drives. Customers only need
to pay the difference in the raw NAND price and OCZ will ship the replacement drive once the
original “E” drive is received.


So you already bought the drive, and for the swap YOU the consumer have to pay the difference in order to get the "more expensive" drive of the same size to get the replacement :rolleyes:

Smooth move exlax :mad:
 
I used to be an OCZ customer. I was just about to buy a new Vertex 2 until I stumbled across their 2xnm "scandal" all over the web. It's hard to believe that someone at OCZ would knowingly and consciously decide to sell a severely inferior product in an already established superior product's packaging and specs at the same price to unsuspecting consumers. FAIL! :shrug:
 
So you already bought the drive, and for the swap YOU the consumer have to pay the difference in order to get the "more expensive" drive of the same size to get the replacement

Plus you throw in your time and effort to replace and restore whatever setup you had on the SSD to start with AND the down time for your system. Adding insult to injury. That's what really ticks me off. :bang head
 
Capacity issues aside, since that's not really what we are truely after.

It falls into the territory of misleading advertising. To have a hardware revision that reduces performance without notifying anyone in the supply or retail chains is very borderline. The appearance created is more along the line of hoping that no one will notice. If the performance and capacity had improved, this would be of no concern.

Given that the advertised specs say "up to 275MB/s," I would be less than pleased with a product that maxes out around 180MB/s. A 35% drop off of the potential advertised maximum is non-trivial.

The next application I was considering building was a caching/storage server with a tiered storage system and using IP over DDR Infiniband for replacement of a Lustre distributed storage system. This requires very fast IOPs and STR on a continuous basis. The change in silicon, without notice and keeping the same product designation, is troubling to say the least. The reduced cycles due to the die shrink is also concerning. Needless to say, my proof of concept is on hold until this shakes out completely, as other vendors will be making the switch as well.

The bottom line is that the moderators in their forums are doing as well as they can. The management should have responded to this quickly and forthrightly. Checking their site, the news section, 02/15/2011, trumpets the transition to 25nm, without mentioning anything about performance, other than standard boilerplate.
 
Xaotic, there also appear to be lifecycle (increased failure rate) problems. And again it probably means more work for the costumer (even under warranty) to replace them in a machine. This is added maintainance cost if a company or OEM uses them. We OCers are small fry compared to that situation.
 
Xaotic, there also appear to be lifecycle (increased failure rate) problems. And again it probably means more work for the costumer (even under warranty) to replace them in a machine. This is added maintainance cost if a company or OEM uses them. We OCers are small fry compared to that situation.

I agree completely. I have to look at this from both sides. I'm finally upgrading my home workstation to SSD for boot with a SAS RAID 10 for apps and smaller data. Increased failure rate would be a downside for either situation. Granted, I have hot spares in larger arrays. It's always inconvenient to have failures even in this situation. For a home user, it can truely be disasterous and could be one of the major impediments to full channel adoption down to the end user.
 
Back