OCZ Vertex 2 with 25nm NAND flash reported slower

[Voo]

Still Indilinx, I think. I guess you meant the old Vertex 2 SSDs?

I thought it was clear that I was talking about the 34nm Vertex 2 drives from context, sorry if I jumped a bit too far there.

 

[jwilliams4200]

I thought it was clear that I was talking about the 34nm Vertex 2 drives from context, sorry if I jumped a bit too far there.

In that case, I have to disagree with your statement that they were perfectly fine. The Sandforce-controlled Vertex 2's have at least two problems -- their sequential read speeds are only about 200 MB/s for most real data, and they have a bug/feature where the sequential write speeds can drop to 60 or 70 MB/s for a week after they determine, in some unspecified manner, that they have had a lot of writes.

 

[Voo]

their sequential read speeds are only about 200 MB/s for most real data

Sure you get about 215mb/s instead of 250mb/s if you use a random 2mb sequential read IOmeter run, but I wouldn't consider that a problem - write is a lot harder affected, but still if you know about the compression and base your opinion of the drive on the real world data you'll end up with a perfectly fine drive.. it's just not the second coming like the marketing benchmarks would want us to believe, but what ever is.

and they have a bug/feature where the sequential write speeds can drop to 60 or 70 MB/s for a week after they determine, in some unspecified manner, that they have had a lot of writes.

Which as far as I'm aware of only happened after benchmarking and/or stresstesting the drive, which is hardly a normal useage szenario.. and would still be about the niveau of a Intel drive.

 

[jwilliams4200]

you'll end up with a perfectly fine drive..

...

Which as far as I'm aware of only happened after benchmarking and/or stresstesting the drive, which is hardly a normal useage szenario.. and would still be about the niveau of a Intel drive.

LOL! You have an odd definition of perfect. To most people, something that has flaws is by definition, imperfect.

And come on! Only aware of after benchmarking? Are you serious? How else would it be measured, by telepathy?

 

[Nebor]

LOL! You have an odd definition of perfect. To most people, something that has flaws is by definition, imperfect.

And come on! Only aware of after benchmarking? Are you serious? How else would it be measured, by telepathy?

Yeah, I think I pointed out that installing Windows 7 from a flash drive is noticeably slower on the new OCZ compared to my Intel SSD.

 

[Voo]

LOL! You have an odd definition of perfect. To most people, something that has flaws is by definition, imperfect.

And you seem to have a strange reading ability, considering that "perfectly fine" is hardly the same as "perfect" in modern english..

And come on! Only aware of after benchmarking? Are you serious? How else would it be measured, by telepathy?

Umn, I assume by writing real data to the drive and seeing that it's not impeded? If you're arguing that you need to benchmark a drive with 4k random data at a QD of 32 to find out how fast sequential write is, you should think about that once more.

 

[Emulex]

geez stop arguing this thread is off topic already.

 

[pkrosenlegal]

I am an attorney with the Rosen Law Firm, P.A.

We are investigating allegations that OCZ may have violated certain consumer fraud laws in connection with their advertising of certain Vertex 2 SSDs. Particularly, it is alleged that recently distributed drives have storage capacities lower than advertised. For example, certain 60gb Vertex2 SSDs have only 55gb of capacity. Likewise, certain 120gb Vertex2 drives only have 118gb or 115gb of capacity.



Other knowledgeable enthusiast sites also claim these 25 nm drives have inferior performance.


As a result of these allegations, the Rosen Law Firm, P.A. is investigating a potential class action lawsuit against OCZ for violations of certain consumer fraud laws.

If you purchased any of the subject drives, you may contact me to discuss your legal rights and our proposed class action against OCZ.


Phillip Kim, Esq.

The Rosen Law Firm, P.A.
275 Madison Avenue, 34th Floor
New York, New York 10116
tel: (212) 686-1060
email: pkim@rosenlegal.com

Attorney Advertising. Prior results do not guarantee a similar outcome.

 

[jwilliams4200]

And you seem to have a strange reading ability, considering that "perfectly fine" is hardly the same as "perfect" in modern english..

That is ridiculous. If you do not mean perfect (or functioning perfectly) then do not write "perfectly". Write fine, or okay, or satisfactory, or sufficient. If you write perfect when you refer to something that is flawed, then you misusing the language.

As for benchmarking, it means measuring the performance. If you do not benchmark something, then it is difficult to know the performance. If you meant specifically benchmarking 4KB performance at QD=32, you should have written that. Do you want to be misunderstood?

 

[Voo]

As for benchmarking, it means measuring the performance.

No actually "benchmark" is defined as "a standardized problem or test that serves as a basis for evaluation or comparison (as of computer system performance)" which is at least for hard drives usually a bit more than one particular metric, so it was pretty clear from context. If you want to misinterpret things well nobody can stop nitpickers anyhow.

That is ridiculous. If you do not mean perfect (or functioning perfectly) then do not write "perfectly". Write fine, or okay, or satisfactory, or sufficient. If you write perfect when you refer to something that is flawed, then you misusing the language.

Sorry if english is a bit more complex than that, I'll try not to confuse you any more with phrases. God forbid someone uses "perfectly normal".

 

[jwilliams4200]

You can keep pretending that you did not write something which was misleading and incorrect, but we know the truth. A flawed SSD, as are all Sandforce SSDs, is not "perfectly fine". They may be adequate, but they are far from perfect.

As for your absurd claim that Sandforce SSDs only throttle the performance after they have been benchmarked, since you are only aware of the lower speed being measured after someone actually measured it, well, enough said.

 

[Voo]

As for your absurd claim that Sandforce SSDs only throttle the performance after they have been benchmarked, since you are only aware of the lower speed being measured after someone actually measured it, well, enough said.

It's well known that if you hammer SF drives with a large amount of random 4k writes in a short period of time (which you realistically only do when benchmarking a drive; no client usage pattern is that stressful) it'll reduce it's performance.. that you can easily check by writing a large sequential amount of data to it. If you get that performance degradation without running any benchmarks (or stresstests, or any other unrealistically stressful application/test - sorry I almost forget I have to spell everything out for you) then come back.

You can keep pretending that you did not write something which was misleading and incorrect, but we know the truth. A flawed SSD, as are all Sandforce SSDs, is not "perfectly fine". They may be adequate, but they are far from perfect.

Yeah I'm terribly sorry for using such.. umn advanced phrases and not limiting myself to basic english :/

 

[jwilliams4200]

It's well known that if you hammer SF drives with a large amount of random 4k writes in a short period of time (which you realistically only do when benchmarking a drive; no client usage pattern is that stressful) it'll reduce it's performance..

Again you misuse the language. In this case, you wrote "it's well known" when you meant "I don't really know and don't have a reference or specification". Come back when you have some actual numbers and facts.

 

[kevinsbane]

The phrase "perfectly fine" does not imply perfection. It is a standard phrase used descriptively; eg, "perfectly normal" means that something is as normal as normal is. It does not mean that it is perfect nor does it try to say that it is perfect. The phrase only serves to highlight the "normalness" of what it is trying to describe, not the object's perfection, which it makes no statement about.

As a result of these allegations, the Rosen Law Firm, P.A. is investigating a potential class action lawsuit against OCZ for violations of certain consumer fraud laws.

This sounds interesting. As much as I dislike OCZ's handling of this matter, I also dislike frivolous or borderline frivolous lawsuits. I'm wavering on whether this is frivolous or not. In the end, OCZ's gonna pay in reputation, no matter whether this lawsuit wins or loses, and the loss most consumers will face is, for the most part, negligible. Well, put it this way. Fraud, I can see. But as a class-action lawsuit? That sounds... unnecessary.

 

[jwilliams4200]

The phrase "perfectly fine" does not imply perfection. It is a standard phrase used descriptively;

It is misleading slang, not a standard phrase. Perfectly fine does imply perfection. Words have meaning. If perfection is not implied, the phrasing should be "fine" or adequate, or satisfactory, or sufficient, or any of a number of other words that might describe the quality. If perfection is not being described, then the word perfectly should be omitted.

Omit needless words!
--Strunk and White

 

[Voo]

It is misleading slang, not a standard phrase. Perfectly fine does imply perfection. Words have meaning. If perfection is not implied, the phrasing should be "fine" or adequate, or satisfactory, or sufficient, or any of a number of other words that might describe the quality. If perfection is not being described, then the word perfectly should be omitted.

Yeah sorry we really didn't know that we couldn't use standard english phrases that have been in use for decades. It's really horrible that words can mean different things in different context, if you send us your personally approved canon we'll make sure to only use that. My ba.. sorry ungood

 

[Hayte]

Playing the semantics game is fine if it actually gets you somewhere and it means something. As it stands, you know what he meant. He knows what he meant. This is pointless. So really we should be able to move on at this point.

 

[jwilliams4200]

Playing the semantics game is fine if it actually gets you somewhere and it means something. As it stands, you know what he meant.

Semantics may be a "game" when we are talking about fine shades of meaning of a word, but semantics is hardly a game when someone writes a word but means the opposite.

I only know what he meant after he clarified that although he wrote "perfectly", he actually did not mean perfectly at all.

And your continued protestations that we should let the matter drop would be less disingenuous if you actually stopped posting about it.

 

[Hayte]

Are you for real?

 

[boozie]

Semantics may be a "game" when we are talking about fine shades of meaning of a word, but semantics is hardly a game when someone writes a word but means the opposite.

I only know what he meant after he clarified that although he wrote "perfectly", he actually did not mean perfectly at all.

And your continued protestations that we should let the matter drop would be less disingenuous if you actually stopped posting about it.

You are acting pretty childish.

Perfectly is an adverb not an adjective. Learn the difference. (the fact you keep posting about "perfectly" instead of "fine" means you don't understand the difference)

I'll fill you in on a 7th grade tip:

Adverbs can describe adjectives.

 

[jwilliams4200]

You are acting pretty childish.

How do you know whether I am pretty if you have never seen me?

 

[boozie]

How do you know whether I am pretty if you have never seen me?

I should have known not to use another adverb.

 

[Concillian]

OCZ has caved to the pressure.

They will pull off the larger 64Gbit flash chips on your drive and replace them with smaller 32Gbit flash chips to populate all channels.

Still 25nm, but by populating all channels, size will be the same as the original Vertex 2s and most if not all of the performance should come back too. I think people have labeled this as a 25nm flash issue, but the main reason for both the size difference and the performance difference was due to the way they implemented that 25nm flash... by using large chips with half the channels. It'd be like changing a 256 bit memory bus graphics card into a 128 bit bus graphics card mid-product cycle.

The issue wasn't really the 25nm chips so much as they were using the larger size chips and not populating the full board like Micros does with the C300 (64 GB drive slower than 128 GB drive). Doing this mid product life cycle was obviously a bad move, customers were rightfully outraged and now they have said they're paying shipping both ways and the total cost to end users is $0.

This was announced yesterday, but, there hasn't been a lot of spread of the news. Those with 25nm Vertex 2s should be going through OCZ's process to get your capacity and speed back.

That sticky post in the OCZ forums is what was updated.

I wasn't affected and haven't owned an OCZ product since I had an AthlonXP, but it seemed people here weren't aware of their change in policy, so I thought I'd make a little PSA.

 

[jwilliams4200]

You would have known if you had graduated from grammar school and went on to learn about style.

 

[Voo]

I should have known not to use another adverb.

Ah you shouldn't waste your time to argue with him. I mean his whole argument is based on "a word can have only one meaning - independent of it being a adverb or adjective or used in different context" - but then I'm hard pressed (note: No, I wasn't killed by a press) to believe he was serious with his last posts.
Just lean back and enjoy the first advocate of Newspeak ;-)

@Concillian: We already discussed this in another thread, but it can't harm to have it here as well. Though "the total cost to end users is $0." is only true if you don't count all the work involved.