SSD Thread (Benchmarks, Brands, News, and Advice)

The fine print always says:

"Specifications subject to change."

 

In states it might valid, but in Europe we have quite good consumer protection policies.

 

Yes, but they need have some ethics to tell the truth to customers. Anyway OCZ customer relations are below average so I'm not surprised.

 

Read it this way: maximize profit!

I can't believe it, they're ripping of buyers with that ATTO read/write speeds (for all SandForce based SSDs) and now with 25nm proces without lowering price.

 

i was a loyal buyer of ocz in past years( ram and coolers) but this is unacceptable.

ok, i know that the market is evolving every day, the passage to 25 nm was only a matter of time, but..for the damned hell, this is a fraud.

the worse thing is the fact that only recently someone has started talking about it in ocz forum and some post were also censored, according to some users.

here you can see differences between 2 versions

Result Mini-review: Direct comparison between 34nm and 25nm drives using 0-fill highly compressible data and 0/1-fill partially compressible data


this is madness!


goodbye ocz, my next 3rd generation ssd will be Intel, and untill i'll have an internet connection, i will blame them around all forums...

 

OCZ is quite liberal in removing any posts on their support forum that make their product look bad

IMO, it is plain childish. Their forum is populated with mostly OCZ fanboy(well OCZ SF SSD fanboy). Truly frusted users either say it on other channels(newegg, here ...) or through blogs. The last they need to worry is their forum

 

That I agree. But it is the case since day one when they only stand by their ATTO 'spec' which we all know is meaningless in SF drive. Since then, I know they are not a trustworthy vendor.

 

@ splinterpc: I see that You're from Italy, where do You buy SSDs from?

 

Havent been in here in a long time. I just got a new laptop today its an envy 14 I5 580m. My old laptop is a 2007 1.5ghz duo dv2550se with a kingston 64gb ssd. The speed difference is a lot. The dv2550se is faster. Just goes to show you what a cheap ssd will do. I turn the envy on first.

YouTube - bigwood212's Channel

 

Yeah... even my Atom beats my Vaio with T9500 in booting time. But that is completely different story when it comes to processing power.

 

boot time is the most uninteresting number for day to day use though a good feel good/measure of what SSD can do.

My Dell D410 with 2GB RAM can resume from hibernation in 12-15 seconds and that is on a 5400rpm HM160C

 

That is what I thought. But now that I am temporarily back to a spinner, I DREAD booting. I need to plan my freaking morning around it it takes so long. Will get a new SSD as soon as possible. For fast boots and many other reasons Silence of which is a big one. SPEED is another

 

To be honest for a desktop or desktop replacement notebook PC, SSD isn't all that important to me. Where I am finding it is important is with netbooks, or smaller laptops. Battery life greatly increased, no vibration (which 7200RPM's are horrible in little laptops), and more durable. I am getting used to SSD's though and do like the features they provide, however, still think they are way over priced.

 

i ordered it here BPM Power - Vendita on line - fotografia,informatica,computer,musica,gps,navigatore,navigatori gps,mp3,ipod,cellulare,cellulari,telefonini,telefonino,marca,modello,modelli,televisore,televisori,lcd,plasma,audio, video, cavi, adattatori

they have been kind: they have proposed me to replace vertex2 with a coupon for a intel x25M

 

Nb... Slighly less performance n more reliability... I still like ocz products tho, but deceiving customers is flat out wrong

 

this page explains well the situation

Micron's ClearNAND: 25nm + ECC, Combats Increasing Error Rates - AnandTech :: Your Source for Hardware Analysis and News

 

doing it more than once should help you understand that you should not "still like" them. they did such stuff since their first ssds. just don't forget that, ever.

 

i would have taken the vertex2 if they had proposed a sostitution with a 34 nm version for free, but they want even more money to give us what they promised! crazy!

 

Just to point out that Intel doesn't have the best track record either. Nothing in their SSDs, but there were the anti-competition payouts, the "unlocking features through add-on cards", and a few other less than honest practices. I'm not trying to minimize the impact of this situation, and I feel that OCZ will have to do some significant work to repair their reputation, but I feel that, as always, everything needs to be taken into account. After all, this may be the "big event" that drives OCZ to be more honest in the future. For right now, certainly, the OCZ line and name is tainted. The same thing has happened to other companies before, and they've gotten past that. We'll just have to see how things turn out.

 

i agree with you,someone is still angry cause they didn't supported trim in 1 gen but here is worse...as some ocz users said, it's like buying a ferrari and find a 4 cylinder inside instead of a v8...the car still runs, but for 100k $ didn't you want more power?

 

ocz didn't support trim as well in the first place. which was obvious, trim wasn't yet defined back then.

intel might have made faults, true. but they never lied on their hardware.

 

The "unlocking card" issue could be taken as a lie about hardware (can't find the original thread we had here at NBR, so I'll just link this Endgadget article). And as I said, I don't disagree that this was a poorly done move by OCZ. If they had been open about the shift to 25 nm (with a new model number, or something similar) then people would at least have known what they were getting, and they could have chosen to spend more to get the better performing 34 nm SSDs. And depending on exactly why you got the ferrari, a 4 cylinder might be better than the v8... better fuel economy, less chance of getting pulled over by a cop for speeding, and it still looks great to anyone that sees you on the road in one! And if the price ended up being cheaper, or staying the same while the v8 one went up, then there are obvious reasons to go with the 4 cylinder instead of the v8. Still, as I repeat again, the issue is transparency; OCZ didn't have it this time, and now they're scrambling on the back end since they got found out. And maybe this whole debacle will drive them to be more forthright and honest in the future. Or maybe not, but that we'll have to wait and see.

 

Guys... who are you kinding? EVERY single company out there has cheated their customers on multiple occasions... sony, apple, microsoft... ocz, intel, ibm... i mean... the list is endless...

Any company looking to make a profit, will at one point in their lifetime, cut some corners to save big bucks and maximize profits hoping to get away with it... and im sure they do MOST of the time... and once and a while they get caught...

This is OLD news and doesn't really surprise me anymore... so what they scammed a bunch of ppl selling 34nm ssds as 25nm... if all goes well they'll get a lawsuit going, get sued, pay a couple mil... and theyre back to doing the same old till they get caught again...

Crooked? Yes... decent products? yes... will I think twice before buying from them? Maybe ... will i end up buying from them again? Most probably.

It's life. Get over it. Scam or be scammed. It's a jungle out here.

 

I hope OCZ will learn lessons from this BAD move. Thru the last two years they have climbed up the rank for ssd's. Now, let's see how this will impact them.

 

if they want people to be a beta tester, at least they could pay them for the patience

 

The way I see the SSD market today (and I completely understand that I may have missed the important bits), I think that the move to the smaller die size may be more of an issue than not. Frankly, I've seen a lot of technology changes over the years, and it seems to me that if the change involved affects some core metric, it is highly irresponsible to use the old model name/number.

For instance, I understand SSDs to have some metrics in performance that are "key" or "core" metrics:
Access Time
Read Time (Random and Sequential)
Write Time (Random and Sequential)
Data Transfer (File copy)
Error Rate
Decay Rate (Errors over time, as the chips lose the ability for further writes)
Power Consumption

I'm likely missing some, but those metrics seem vital to me. If I am reading the reviews of the technologies correctly, the change to the smaller die will dramatically effect the Decay Rate, as smaller dies result in fewer writes-before-failure. In other words, the drive just got a lower life-span.

I understand that error correction and other tricks are working to counteract this effect, but I also read that it isn't a 1:1 ratio trade-off yet; Lifespan is decreasing due to decay faster than it is increasing due to ECC technologies.

The short of all that is, in my opinion, to change core metric performance of a drive without changing the model number or the model name is really foul play. I don't think it would really matter if the core metric got better or worse, from the consumer's point of view. We need the ability to know for certain what the performance metrics are for a component, based on what we are given (almost universally a model number or name). With today's post-consumer market (ebay, etc.) it becomes even more important to know what you are buying.

Would it have hurt to put "Rev. B" after the model number or name? Can't cost that much to change the box printing or the label stencils. Compared to the sudden distrust of the consumer base, it should have been a no-brainer.

Did OCZ deliberately try to cloud the change? That's a harder question to answer. Only those in OCZ management will ever really know. I would go out on a limb to say that they would be pretty stupid to have tried that, given the information economy of the world. I can't think that they didn't realize someone would notice the substitution.

If that's the case, then we are left with a radical incompetence in the management of OCZ. Someone, somewhere, signed off as responsible for the decision. Either they didn't understand the impact of the hardware change (which would be very distressing, if true) or they thought the benefits of the change outweighed the liabilities, and that today's enthusiast community would accept a major change in manufacture without altering the model identifier. That, too, is so short-sighted as to be distressing.

Anyone hear of someone getting sacked lately at OCZ?

 

No one knows but their explanation is not convincing. They switched from 32Gbit chip(34nm) to 64Gbit chip(25nm). However, now they are asking you to pay up the difference for 32Gbit(25nm).

That means, they have no issue using 32Gbit 25nm chip in the first place(which other than initial yield should be cheaper to produce than 34nm and usually cheaper from say Micron as memory producer don't want to sit on stocks).

That sounds to me like that they want to hide this 32Gbit -> 64Gbit in the same transition of 34nm -> 25nm, i.e. cloud the real change.

 

I don't think that is the same thing. You get what you pay for. They didn't sell you the 'unlocked' capability. This is a practice commonly used for a long while. IBM is well known for that in their mainframe etc.

Microsoft is doing it now too for Windows 7. You only have one set of image and unlock the features depending on how much you pay.

OCZ's action is not the same. Using Intel as an analogy, it is like rebatch an Atom as Pentium. It runs the same x86 instruction set, you get even better power consumption number plus some additional stuff. The catch is it is much slower(clock for clock). I doubt Intel would do this without telling people about it, it can become a huge legal liability.

 

And, to a certain extent, we are getting what we paid for here, too. Prices on Vertex 2s have dropped. The issue has been that the reason for the drop has not been advertised, and then people are finding out that there are other issues as well (the performance drops). As well, in a sense, OCZ has been telling people that 25 nm was shipping (it's been mentioned in every firmware revision since... 1.26, I think). It certainly wasn't transparent enough, as all the furor is attesting to, but in terms of what OCZ has actually published as specs for these drives they still match (except possibly the reduced size from the increased need for RAISE, but the difference between actual usable capacity and marketed capacity has always been a bit fuzzy, compounded by different formats and the GiB and GB difference). The issue here is the discrepancy between what people thought they were getting and what they actually got; a better example might be the WNDR3700 and the WNDR3700v2. The v2 has somewhat poorer performance than the original v1, and if you missed the fact that it's a v2 as compared to the v1, you might have the same disappointment when the v2 didn't perform as well as the reviews on the v1 did. Of course, in this OCZ case, the fact that it wasn't clearly advertised that a model was 25 nm instead of 34 nm made that much easier.

I also don't think using Pentium is the best example, considering how the Pentium name has processors based off of a bewildering number of micro-architectures (Netburst, P6, Core, and Nehalem are all included... and before long I'm sure Sandy Bridge too).

 

That is actually my point. Pentium was initially an architecture then becomes a brand. But there are 'reasonable expectation'. So if Intel use the Atom architecture as Pentium without telling people, it would open a legal flood gate. What you have mentioned so far is better and better(well at least overall performance) architecture for the same Pentium brand.

Try to imagine if Toyota would use one source of Steel for its Camry chasis and do the crash test and later replace it with something 1/2 as strong in order to save cost. They are still the same thing(Steel).

 

And if the 1/2 as strong steel still met the same crash specs as the original, then there would be no (legal) cause for complaint. Please, do read all of my posts on this topic. I agree completely that this was a bad mistake on the part of OCZ. My points to date have been that OCZ is not the only company to have made errors like this, and that while this is a good enough reason to avoid buying Vertex 2s for the time being, this isn't necessarily a reason to write off the company completely. As always, it comes down to knowing exactly what you're buying and paying for (and perhaps the ability to know what you're buying, which OCZ has admittedly compromised with the unclear labeling of 25 nm versus 34nm). If in the future, OCZ comes out with a better drive than Intel for a better price point, I might go back to them.

 

I know if the steel still meet the crash spec, it is non-issue. I implicitly use the example to save me some typing or else we are exchanging legal documents.

The point is, what OCZ has done is equvilent to using softer steel that no longer has the same strength(that was done using the strong steel). Sure, it may still meet the 'minimal' legal government requirement. That doesn't mean they are not opening legal flood gate for class action.

What is your example of other company doing the same thing ?

 

I was saying that the even the newer drives meet the original specs that OCZ published (we won't get into how those inital specs are not necessarily representative, given the way that Sandforce compresses data and thus has issues with incompressible data). This isn't some "minimal" legal requirement, this would be equivalent to the softer steel still letting the car meet the same crash performance as the harder steel (except that now, maybe the car is easier to dent or something due to the softer steel, or that "weaker" crashes are more likely to involve totaling the car, similar to how the addition of "crumple zones" made cars safer, but more... destructible). As for a more recent example, how about the iPhone 4 issues? First there was the antenna. And now there are issues with the glass casing. And yet how many people still buy Apple?

 

We are not talking about whether people will be buying OCZ or not. As I mentioned in another thread, there is one born every minutes.

But the nature of what OCZ has done and why I mentioned the Toyota case. If they take a hardened Camry for the crash test which get a 5* rating then put that in every ad they use. Then later switch it to one that get a 2* rating(that still meets the legal requirement).

As I said, it is not about the 'technicality', OCZ only officially mention ATTO so 'technically' then are ok. But 'reasonable expectation'.

Apple's case again is different. It is just a bad product quality issue.

anyway, I think we both have expressed enough of our view.

 

I disagree in that, to use your example, it'd be more like they had a hardened Camry that got a 5* rating in frontal impacts, and advertised that fact, and production Camrys still get a 5* rating in frontal impacts, but now have lower ratings in, say, low speed impacts (which they never advertised for in the first place). As for reasonable expectation, most of the complaints so far are over benchmarks; I'm not sure that anyone has seen any noticeable difference in actual usage (apart from, as mentioned previously, the space issue). From benchmarks I posted earlier, I've apparently lost 30% of my sequential and 512K write speeds on my Vertex 2 (34 nm)... which I never even noticed.

The iPhone 4 antenna issue is bad design (is that what you mean by bad product quality?), while the glass casing appears to be weaker than that of the iPhone 3GS. It's not a direct comparison, no, as there is an obvious model change involved, but the marketing launch of the iPhone 4 claimed that the new glass was even stronger now. So in this case, "reasonable expectation" would involve the new iPhone 4 glass being less breakable, no?

Either way, I agree with you that we're both just arguing relatively minor points back and forth now. We both agree that this is a bad move on OCZ's part, and although my contention is that while this taints the OCZ name and line for the present, there may be hope for them in the future, I'm not as certain what point you were arguing.

 

The analogies are not really helping and are serving to confuse the matter further.

This wouldn't hurt OCZ so badly if they had a unique product but they don't. They package SF-1200/1500 controllers with NAND ICs, stick them on a pcb and sell them with a markup, just like Corsair, Mushkin, G-Skill etc. If one company screws up you just go to the next alternative that doesn't. The reason why OCZ even managed to separate themselves from the masses of other ex RAM companies pairing disk controllers with cheap NAND is that they had a highly active and visible forum and the company was seen to care about firmware and post sales support.

It is this relationship with their consumers that is partly responsible for Vertex 2 being the default Sandforce recommended buy on every internet forum I've been on. What this debacle does is show that OCZ doesn't take this consumer relationship seriously.

Without this, and with the toolbox in the state that it is in right now, there is literally no reason to buy a 60gb SSD from OCZ when you could do so from Corsair or G-Skill or ADATA or whoever, because:

1) You don't know whether you are getting a 34nm, 16x32Gbit drive or the hamstrung 25nm, 8x64Gbit drive. The warranty is void if you try to open it up to find out.
2) Trading the slower drive in for a 25nm, 16x32Gbit drive costs $10.00 + shipping on top of the retail price. This makes it pricier than the competition which essentially offers the same drive as the older, faster, longer living Vertex 2 without the lucky dip part.
3) The OCZ toolbox is a really horrid piece of software. If your SSD is your primary system drive (and lets face it, thats what you bought it for), you cannot use the toolbox firmware updater. You need to boot from another system disk and flash the firmware, or run the updater in Linux (!!!), or boot from Parted Magic, which is a long and complicated process.

Toolbox secure erase is horrid. It cannot perform a secure erase as long as a partition exists on the disk, which means that you need to wipe the drive anyway. Good luck getting rid of the system reserved partition which cannot be removed through Windows Disk Management.

 

I got a 240GB G.Skill Phoenix Pro a couple months ago for my Asus laptop. Being my first SSD, it was quite awesome. But I have that stupid sleep bug (Windows 7 x64). I use the sleep function a lot, but I've been forced to hibernate since I got the SSD. For some reason, waking up from hibernation takes longer than a cold boot...actually, that makes sense because there's more to load back into the RAM, right? Whatever the case, I wish I'd gone with Intel's G2.

I looked around to see if the Intel G3's were out yet and it seems like no one knows when it'll come out. By the time it does, SF2000 will follow behind.

Don't know if I should put up with what I have or sell my SSD and get an Intel G2 (or even the Momentus XT) to hold me over until the next generation of SSDs.

 

You may be right in what distinguished OCZ from the others, but for me, what attracted me to OCZ was very specifically their price point. When I purchased my Vertex 2 last December, it had what I believed to be the best price point ($270 for a $240 GB after MIR, promos, and $100 gift card from Newegg) for the performance that I wanted. I knew about their forums (especially their warts), but had no real plans to resort to them, unless an issue came up that we couldn't resolve here.

I can't argue against your first point, and have no basis for argument on the third (haven't needed to use the toolbox for anything other than the SSD SMART values), but assuming that what OCZ has said about the reasons for the shift are true, your second point won't be true for much longer. OCZ says the reason for the shift is that 34 nm NAND is no longer sourceable except at inflated prices, and this will be true for everyone that doesn't make their own NAND (in other words, just about everyone but Micron, Samsung, and Intel). This means that as the other Sandforce manufacturers run out of stock, they'll have to shift to 25 nm SSDs too, which will pretty much perform just as the OCZ ones do. The difference being, of course, that they'll make it (the shift) more obvious.

 

seriously, they do that all the t ime, and lie to customers when they report that their ssds are flawed, and deny till they get openly from media shown to have flaws. then they "fix them" to get trough the benchmarks in the media, only to have media show again, that it was no fix at all.

i'll never forget the core series. and how they've shown to not be trustable at all, ever. and so far, they haven't changed, as seen right now.

 

So I am a broke college student thinking about buying an SSD for my y560 laptop and i was wondering if the Intel x25-m 120gb SSD is a good first choice to transition to the ssd over hdd storage. I dont know how to maintain or tweak ssds so I thought the intel software would handle all that. That is my main reasoning for picking this ssd over others. Any input would be appreciated.


Newegg.com - Intel X25-M SSDSA2MH120G2K5 2.5" 120GB SATA II MLC Internal Solid State Drive (SSD)

My y560 laptop also supports msata as well so i do have the option of buying the new msata intel 310 ssd and just install the rapiddrive technology from lenovo which combines ssd+hdd into one big drive.

http://www.newegg.com/Product/Product.aspx?Item=N82E16820167040&cm_re=msata-_-20-167-040-_-Product

just confused on which would yield better performance as space is not a concern.

 

Intel is a good choice. You can't go wrong with them really, and don't need to do anything if you do a fresh install of Windows 7 on it. Just install and enjoy.

 

While I would be tempted to buy just the mSATA version, the 120GB G2 you have linked will give you better performance and is a better value.

You will benefit by using less of the drive (% filled will be lower) and getting higher performance... in addition to the already higher performance (similar to a 160GB G2) because of the additional channels in use compared to the 80GB version.

I would buy both.

But for $230, get the 120GB G2, do a clean install of Win7 and enjoy your college life with probably this single computer/install to get you through.

(Btw, how do you like your y560?).

 

They have no problem shifting to 32Gbit 25nm chips so this reason is not convincing. They are saying, pay 10 bucks more and you can get that.

 

If you're broke and capacity isn't an issue, you can get a 60 GB G.Skill drive for $97. It's half the space of the Intel, but less than half the price. While the G.Skill has faster sequential writes than the Intel, its 4K read/write performance isn't as good. It's a worthwhile trade-off for the price, IMO. There's nothing you would need to do to tweak this either. Just install the drive in your notebook, install Windows 7, and you'll be good to go.

 

I've been offered a Samsung Series 470 128GB drive from a supplier who owes me a favour, are these drives any good? (Gosh this is a very long thread to find info in!!)

Thanks

 

^ I have the 256GB Samsung 470 and I haven't had any problems with it.

 

when the heck are these new SSDs gonna start to be released?

intel, sandforce... when are they coming?

i noticed newegg added the new plextor ssds today, and the corsairs were released a few weeks back, but i want the good stuff already!

 

Right, because 32 GBit 25 nm NAND chips are more expensive than 64 GBit 25 nm NAND chips, and 34 nm 32 GBit NAND chips are even more expensive than that. My point is that unless the other vendors can get those 32 GBit 25 nm chips more cheaply than OCZ, their price points are still going to be about the same, with 16x 32 GBit 60 GB drives being more expensive than 8x 64 GBit 25 nm 60 GB drives. Note that as far as I know, noone else is yet offering the 25 nm drives (or if they are, they're doing the exact same thing OCZ is, so...), so the comparison in Hayte's post is about 34 nm drives to 25 nm drives. A quick look at Newegg's 60 GB SSD drives reveal that the Vertex 2 is still cheaper by about $10 (or more!) than most of the other non-OCZ drives (the 2 cheaper beating it out by $5 or less), and don't forget that largely the whole point of the Vertex 2 is that it's supposed to have extra IOPS through OCZ-only firmware compared to all other Sandforce controllers.

Edit - Oh, and as pointed out in the other thread, they're now waiving the $10 fee, and they've stopped production of the 8x 64 GBit drives. You still need to pay shipping through.

The 470 series are currently probably the best drives out there (except maybe the C300s, which really only shine on a SATA III interface).

 

- check this out -

Western Digital now has 32 and 64 GB *SLC* SSDs

kind of expensive though ...

 

Cannot find them on the crappy WD site, link please!