OCZ Vertex 2 with 25nm NAND flash reported slower

[tweakboy]

http://www.guru3d.com/news/ocz-vertex-2-with-25nm-nand-flash-reported-slower/

 

[jimhsu]

This is all over the net now.

I'm wondering if this is indicative of an underlying problem with the 25nm process? Intel has been delayed for a few months as well with their 25nm release.

It's quite possible that we might see two NAND processes in market for 2011 ... the 34nm "performance" product, and the 25nm "value" product. Oh well, we'll see with the release of the other 25nm drives.

 

[tweakboy]

Their having fabrication issues big time. This then the Sandy B2 stepping SATA bug. Intel has been stinking it for a while now, IMO.

 

[jimhsu]

On Corsair's side, the Corsair Performance 3 also uses 34nm NAND (http://forums.bit-tech.net/showthread.php?t=202245 ), but the C400 uses 25nm. Wonder how that will turn out ...

I think most of the performance issues simply comes with using LESS NAND chips on the design -- going say from 8x 8GB 34nm NAND to 4x 16GB 25nm NAND is unquestionably going to be slower, even if the 25nm chips are as good as the 34nm.

 

[frostedflakes]

The new drives supposedly have more over-provisioning as well. For example I think I read the 120GB (111GiB) drive now ships with 115GB (107GiB) available to the user.

edit: Crap, I see the article already mentions that, sorry.

edit 2: The performance difference shouldn't be that huge (if anything, isn't NAND supposed to gain a little read/write performance when you shrink? I didn't think performance was supposed to go down). Probably a firmware issue or something, I'm sure OCZ will get it worked out.

 

[Voo]

The new drives supposedly have more over-provisioning as well. For example I think I read the 120GB (111GiB) drive now ships with 115GB (107GiB) available to the user.

edit: Crap, I see the article already mentions that, sorry.

Yeah mentioned already in the article and I personally think that's just plain fraud. If I pay for 120GB I should get 120GB, not less. If they have to overcompensate more because they use smaller flash (which is MUCH cheaper for them) they should just include more flash on the drive.

Oh and the performance differences ARE huge, one of our members got a new drive and benched it - 37mb/s sequential write isn't acceptable even if it's not the most important stat.

 

[sawtx]

Was wondering why the Vertex 2 I got last week only showed 107 GB.

 

[frostedflakes]

Yeah mentioned already in the article and I personally think that's just plain fraud. If I pay for 120GB I should get 120GB, not less. If they have to overcompensate more because they use smaller flash (which is MUCH cheaper for them) they should just include more flash on the drive.

Oh and the performance differences ARE huge, one of our members got a new drive and benched it - 37mb/s sequential write isn't acceptable even if it's not the most important stat.

Well technically you are getting 120GB.

120GB = 111GiB. Just like 10Mbps internet = 1.25MBps. Both are accurate, just different ways of measuring it.

I just don't see how there's any way that the shrink from 34nm to 25nm could cause that much of a performance drop. It has to be a bad firmware or something like that, or as jimhsu suggested maybe they're using fewer flash chips in the lower capacity models, which means fewer channels and lower performance.

 

[jwilliams4200]

Well technically you are getting 120GB.
120GB = 111GiB.

Nope, the available space on the 25nm Vertex 2 SSDs is only 115GB = 107 GiB. There may be 128 GiB of flash on the circuit board, but the available space to store your data is what counts, and that is no longer 120GB.

 

[frostedflakes]

Well yeah obviously advertising it as 120GB when it's actually 115GB is messed up, OCZ should change all the specs, packaging, etc. to reflect this. Was just pointing out that before the switch to 25nm, advertising it as 120GB was correct.

 

[darckhart]

well if the lawsuit against WD taught you anything, it's that the manufacturers don't care to change the way they advertise drive capacities.

 

[jwilliams4200]

Apples and oranges. WD and the other HDD manufacturers are correctly using the metric system to report drive capacities. Actually, you get a little more than they advertise. All the 2TB HDDs should have a minimum of 2 x 10^12 bytes of available capacity, since T is the metric system prefix for 10^12. All of the 2TB HDDs actually have 3907029168 available 512 byte sectors, which comes to 2000398934016 bytes, 398934016 bytes more than advertised.

OCZ is selling 25nm flash SSDs with only 115GB of available capacity, with a 120GB label.

 

[jimhsu]

Using higher density, but less NAND chips will most certainly cause performance to go down. SSDs essentially scale linearly (disregarding any firmware effect) with the number of NAND chips (that's why of course larger SSDs are faster). With 25nm, manufacturers can now halve the number of NAND chips, reducing costs significantly. What OCZ has done is basically taken the performance of a 60GB model and sold it as a 120GB model. That is what's unfair to customers. And the reduced lifecycle thing (25nm is good for about 3000 cycles, compared to 5000 for 34nm). And the 5GB of space lost due to overprovisioning to counteract the lifecycle.

See, if they sold the drive as "Vertex 2 LE" or something, and reduced prices, people wouldn't be complaining. In fact, they'd probably appreciate the price segmentation (making it more affordable to all). Unfortunately what we have here are 2 distinct products with the same exact model number.

 

[Quad5Ny]

Does anyone know if the non "E" versions still have 34nm Flash?

Man this really sucks, I just purchased 2 Vertex 2's off of newegg and they're the 2x nm versions. =\

Maybe my local microcenter still has the 34nm drives? Hmm, looks like I'm driving out there later today.

 

[Nebor]

It seems like I was in the initial wave of people figuring it out. Write speeds are WAY slower than the 34nm models. Most people's on the OCZ board seem to be capping out around 37 MB\s. Of course, OCZ is locking\deleting threads about it now...

 

[Aristotelian]

Great find. I'm not sure how many of you care about Corsair's new drives, but there's a thread on these forums (here: http://forums.anandtech.com/showthread.php?t=2141031) where a number of people were trying to figure out why the Corsair drives bench significantly more slowly than Crucial's older c300 (also, why the Corsair is 20% more expensive, at least here in Germany).

I think this Guru3d article hits the nail on the head, but even worse now is that the drives will be cheaper for the manufacturers to make, and slower, yet more expensive? I mean, honestly.

 

[jwilliams4200]

Great find. I'm not sure how many of you care about Corsair's new drives, but there's a thread on these forums (here: http://forums.anandtech.com/showthread.php?t=2141031) where a number of people were trying to figure out why the Corsair drives bench significantly more slowly than Crucial's older c300 (also, why the Corsair is 20% more expensive, at least here in Germany).

That has nothing to do with the Corsair P3 SSDs performance. For one thing, Corsair P3 SSDs are not using 25nm flash. But more importantly, the sequential write speeds for the P3 SSDs are very high. The 25nm OCZ Vertex 2 SSDs are displaying very low sequential write speeds. So clearly 25nm flash has nothing to do with low 4K P3 performance.

 

[PowerYoga]

I was one of those people that jumped on the newegg sale wagon. What sort of speeds am I supposed to be seeing anyways? I did a quick AS SSD benchmark on a clean install and I'm getting at least 75 write, but I'm not certain I have AHCI enabled correctly. (Seeing something called JRAID instead of iaStore...)

 

[Aristotelian]

That has nothing to do with the Corsair P3 SSDs performance. For one thing, Corsair P3 SSDs are not using 25nm flash. But more importantly, the sequential write speeds for the P3 SSDs are very high. The 25nm OCZ Vertex 2 SSDs are displaying very low sequential write speeds. So clearly 25nm flash has nothing to do with low 4K P3 performance.

That teaches me to post before doing the relevant fact checking. Sorry for the mistake - and thanks for clearing it up!

 

[Nebor]

I was one of those people that jumped on the newegg sale wagon. What sort of speeds am I supposed to be seeing anyways? I did a quick AS SSD benchmark on a clean install and I'm getting at least 75 write, but I'm not certain I have AHCI enabled correctly. (Seeing something called JRAID instead of iaStore...)

Yours sounds better than most.

 

[jimhsu]

75 MB/s write for a 120GB (*cough* 115GB) model sounds about right for the 25nm process. The 37MB/s results were reported for the 60GB model.

 

[PowerYoga]

ouch. Looked at some of the benchmarks for the 34nms and those were way faster.

What a bait and switch on OCZ's part... Not very happy right now but at least its better than a HDD. Wish my first SDD could've been the intel refresh lineup but that just kinda disappeared from the news.

 

[BTRY B 529th FA BN]

i don't care, as long as they get it straight before they release a beta version to the public

 

[Numenorean]

UGH!

This means that my laptop drive is the 25nm version....I was wondering why it had less space than the one in my desktop.

Oh well, at least the performance drop in the laptop isn't as big of a deal to me.

Now I'm tempted to return it for an Intel though....

 

[XLNC]

ouch. Looked at some of the benchmarks for the 34nms and those were way faster.

What a bait and switch on OCZ's part... Not very happy right now but at least its better than a HDD. Wish my first SDD could've been the intel refresh lineup but that just kinda disappeared from the news.

Yeah this pretty much confirms what a few other threads have been suggesting. I'm pretty sure other manufacturers will follow suit. I hope atleast, that they'll distinguish 25nm products separately. OCZ will not see another cent from me, ever.