Reliability of SSDs?

[Burner27]

Your reads are lower than mine, though your writes look about like I'd expect of 3x256gb m4's in RAID 0.

Same benchmarking program?

 

[Burner27]

Same benchmarking program?

Here is mine using AS-SSD

 

[bryanW1995]

Here's what I have on a single disk:

And here's what I get in RAID 0 with 2 x 256gb

Why do you only have 650gb? BTW, I have write back cache enabled.

 

[beginner99]

My ideas:

1) Get 2 of the new 1 TB raptors in RAID1.

They are faster than current setup and offer more space. I think you actually have a usecase for these new raptors.
Probably around $600

2) 2x 128 GB SSD (Crucial M4, very cheap) in RAID1 as system drive + HDD for Storage

And then RAID1/5 with HDDs for storage.
Probably around $ 520 with 1 TB WD RE drives.
You could also go for 2x64 GB SSD if its enough space for system and Apps but the price difference is IMHO too small to justify this especially since the 128 GB versions are faster in writes.

 

[Burner27]

Here's what I have on a single disk:

And here's what I get in RAID 0 with 2 x 256gb

Why do you only have 650gb? BTW, I have write back cache enabled.

Write back cache is enabled. The reason you are seeing higher numbers is because there are only 2 x SATA III ports on my board and I am attached to 3 SATA ports. When you do that, it reduces the tansfer rate to the slowest denominator: SATA II. You are hooked up to 2 x SATAII - hence your higher numbers.

 

[Turbonium]

Many around here, including me, regard the Samsung 830 series as the most reliable SSD on the market. If you bought two of those and updated to the latest firmware before RAID'ing them together you will have a stable setup.

Are Intel SSDs considered reliable? I'm not really a fanboy, but I find myself naturally gravitating towards them when considering SSDs.

 

[William Gaatjes]

How is the quality ?
I found this website mentioning some interesting issues about new flash chips.
If it is reliable or not, i cannot confirm at the moment.

http://www.storagesearch.com/ssdmyths-endurance.html

Even after more than 8 years writing about flash SSD endurance this subject remains 1 of the top 3 SSD topics still searched for by our readers. That's because just when the controllers get better, the new flash chips get worse again - and the faster that SSDs get the quicker they will use up those 9 lives (or 3,000, or 10,000 etc).


Flash wear out still presents a challenge to designers of high IOPS flash SSDs as the intrinsic effects at the cell level get worse with each new chip generation. That's in contrast to RAM SSDs - where as long as enterprise users remember to replace their batteries periodically - the memory life is more dependent on elapsed time (classic bathtub reliability curve) and heat stresses rather than directly related to the number of R/W cycles.

Higher SSD capacity, and faster speeds come from progressively smaller cell geometries - which we used to call shrinks. In flash memory small size means less trapped charge holding the stored data values and greater sensitivity to charge leakage, charge dumping and disturbance effects from the normal processes which happen around the cell vicinity during R/W, powering up, powering down etc.

If you're a consumer you don't have to worry about the internals of endurance management - because most new SSDs are good enough (if they're used in the right applications environment).

Exceptions still do occur, however for users in the enterprise SSD market - where I still hear stories of users thinking it's perfectly normal and economic to replace burned out Intel SSDs every 6 to 12 months - instead of buying more reliable (but more expensive) SSDs - from companies like STEC.

Butif you're a systems designer it's useful to know that the longevity difference between "good enough" and the best endurance architecture schemes can still be 2x, 3x or 100x - even when using the same memory.

And new evidence is coming in from research done by STEC that old, heavily written MLC cells - managed by traditional endurance schemes - may get slower as they get older - due to higher retry rates on reads - even though the blocks are still reported by SMART logs as "good". Meanwhile a recent paper by InnoDisk confirms that whereas SLC and MLC memories have often had endurance populations within each chip which were mostly much better than guaranteed (something which SSD makers had been telling me since 2004) - the headroom / margin of goodness - in new MLC is lower than in the previous MLC generation. That's why controllers which used to work well with vintage MLC need something much stronger than a tweak to deliver well behaved SSDs when co-starring with the new brat generation of naughty flash.

http://www.stec-inc.com/

 

[bryanW1995]

Write back cache is enabled. The reason you are seeing higher numbers is because there are only 2 x SATA III ports on my board and I am attached to 3 SATA ports. When you do that, it reduces the tansfer rate to the slowest denominator: SATA II. You are hooked up to 2 x SATAII - hence your higher numbers.

Of course, I should have thought if that. Hopefully haswell will have 4-6 sata 6gb/s ports.

 

[bryanW1995]

Are Intel SSDs considered reliable? I'm not really a fanboy, but I find myself naturally gravitating towards them when considering SSDs.

Intel and Samsung are 1 and 1a in reliability, with crucial in a strong 3rd. After that, most ssd's with a Marvell controller are pretty good.

Intel used to be the unquestioned number 1, but their new 520 and 330 series both use the sand force 2281 controller.

 

[Rubycon]

RE: Intel 520/330 series: I don't believe the use of SF controllers is as big an issue as a proper firmware AND quality flash.

As for SATA / SAS 6Gbps channels they are already too slow. Bring on 25+ Gbps fast ipath already!

 

[bryanW1995]

RE: Intel 520/330 series: I don't believe the use of SF controllers is as big an issue as a proper firmware AND quality flash.

I would have disagreed with you 3 months ago, but so far so good. And intel has such a good track record that I'm inclined to give them the benefit of the doubt. Of greater concern to them should be that samsung has a similarly stellar quality reputation.

As for SATA / SAS 6Gbps channels they are already too slow. Bring on 25+ Gbps fast ipath already!

I agree!

 

[Rubycon]

520s are strong performers on the test bench but I've got my 830s earlier. They just flew through the torture/reliability testing. Plus no penalty for compressed writes which is always a plus.

 

[Coup27]

Are Intel SSDs considered reliable? I'm not really a fanboy, but I find myself naturally gravitating towards them when considering SSDs.

I personally think that Intel have lost one notch of their previously impeccable reliability record, impeccable for SSDs circa 2009-2010 at least.

If you go back to the original X25-M, then the X25-M G2 and X25-E, those drives were rock solid when they came out. A few firmware updates later and they were rock rock solid. I recently sold an X25-M G2 and it was totally bullet proof. The 510 series never even received a firmware update, one of the very few SSDs to ever achieve that.

Amazingly the 320 series which was built on the G2 has not proven to be as stella and was/is plauged by the 8MB bug which there is more than enough evidence online to suggest has not been completely cured, and with the launch of the 330 series, is not likely to be cured now either.

With Intel completely outfoxed in the controller market, it turned to SandForce. Granted it may have sorted out a lot of the problems with the firmware (there is evidence to suggest Intel actually caused a lot of these as well) but there is far too many reports online of BSOD's with the 520. I also dislike how Intel ignores it's customers on their "community" support forum. The company even has reps who post here so how that happens I don't know.

I have owned a G2 myself and bought 2 320's and all have worked fine, but I just don't believe Intel are as good as they were. The #1 reliability torch has to go to Samsung.

 

[jwilliams4200]

I also dislike how Intel ignores it's customers on their "community" support forum. The company even has reps who post here so how that happens I don't know.

I have owned a G2 myself and bought 2 320's and all have worked fine, but I just don't believe Intel are as good as they were. The #1 reliability touch has to go to Samsung.

Do you have a link to the forum where owners can interact with Samsung SSD technical support? :whiste:

While I agree with most of what you wrote, I cannot agree with disparaging one SSD company over another with regard to their forums. I haven't seen any manufacturer's forums for SSDs that are worthwhile. Samsung and Plextor do not have any (AFAIK), Intel has one that is rarely helpful, Corsair has one that is rarely helpful, OCZ has one that is very annoying but rarely helpful.

The best bet is to buy the most trouble free SSDs and hope for the best, since the manufacturer support forums aren't likely to help much (well, maybe they can help novices, but I am talking about help with real problems, not user error or ignorance).

And I think Plextor just nudges out Samsung for #1 in reliability.

 

[Coup27]

Do you have a link to the forum where owners can interact with Samsung SSD technical support? :whiste:

I cannot, because they do not have a forum. However having no forum is better than having a forum and then ignoring your users.

I also question your comment that Plextor nudges out Samsung in reliability. Plextor don't design or make their SSDs, they buy them in from Lite-on who also sell the design to others such as Corsair. This leaves Plextor has little more than a firmware house in relation to SSDs.

 

[jwilliams4200]

I cannot, because they do not have a forum. However having no forum is better than having a forum and then ignoring your users.

Hardly. Both are equally unhelpful.

 

[Coup27]

Hardly. Both are equally unhelpful.

I disagree.

Samsung has no forum, so it's users will fill out a support ticket and will receive a taylored response to their question.

Intel has both options, but only one is manned. That leaves users who opt for the forum approach first, waiting days and days thinking that somebody from Intel will answer their query only to be ignored to the point of either seeking help elsewhere or filing a support ticket, wishing they had done that in the first place.

I talk from first hand experience.

 

[npaladin-2000]

While I agree with most of what you wrote, I cannot agree with disparaging one SSD company over another with regard to their forums. I haven't seen any manufacturer's forums for SSDs that are worthwhile. Samsung and Plextor do not have any (AFAIK), Intel has one that is rarely helpful, Corsair has one that is rarely helpful, OCZ has one that is very annoying but rarely helpful.

I frequent OCZ's forum, it's actually pretty helpful, and they have a lot of their support techs and engineers frequenting it, actively. And I realize I'll get flamed for saying something positive about OCZ.

 

[jwilliams4200]

I frequent OCZ's forum, it's actually pretty helpful, and they have a lot of their support techs and engineers frequenting it, actively. And I realize I'll get flamed for saying something positive about OCZ.

No, it is not helpful for real problems. The replies to real problems are either to claim that it is your fault, or else they give a whole list of useless things to try, hoping that you will eventually get tired and leave them alone.

 

[jwilliams4200]

I talk from first hand experience.

So, because you didn't realize that the Intel SSD support forums are as unhelpful as all other manufacturer SSD forums, and because you didn't bother to call Intel for support, that means that Intel's forum is worse than Samsung's lack of a forum?

Impeccable logic there

 

[groberts101]

No, it is not helpful for real problems. The replies to real problems are either to claim that it is your fault, or else they give a whole list of useless things to try, hoping that you will eventually get tired and leave them alone.

here we go again.. "flame on" Johnny Storm! LOL

I'm guessing that I may know a tad more about OCZ and their forums than you do. I also read many others too.

And with that in mind.. OCZ helps more people on a weekly basis than many of the others combined.. or compared to much longer periods of time. In fact.. many of those "others" actually lurk around or even openly admit to owning another mfgrs hardware and still come there for help with their issues. Of course the smart ones just fill in.. "a similarly controlled drive to OCZ" with the corresponding models actual name.

The funniest thing that I see with many "expert opinions" around here?.. is the lack of understanding or just plain ignorance of the data that exists in plain site for all to see if they choose to do the legwork to find it. Many of these so called "drive related problems" are sorted out to be software and/or bios config related in the end.

I see dozens of posts every week with initial post comments like "I see many others with similar issues, so why can't you guys get these drives sorted out?". Then by the end of that same thread?.. "I apologize for being rude and blaming this hardware only to find out it's a driver".. or "a bios update".. or something as simple as "a software conflict in my install".

An F4 bsod or freeze?.. does not a bad SSD make.

It's far more rare than many realize, when we see a hardware combo's that literally pinpoints the drive as the main problem. Maybe 10 percent at times and usually much less than that.

Helps to have enough information to accurately assess the situation.. and clearly.. you do not. :thumbsdown:

 

[jwilliams4200]

Helps to have enough information to accurately assess the situation.. and clearly.. you do not. :thumbsdown:

Yup, that's it. It couldn't possibly be that you are an irrational OCZ fanboy who posts a lot of nonsense.

 

[npaladin-2000]

Yup, that's it. It couldn't possibly be that you are an irrational OCZ fanboy who posts a lot of nonsense.

Yeah it couldn't possibly be that you have an irrational knee-jerk hatred for all things OCZ.

 

[jwilliams4200]

Yeah it couldn't possibly be that you have an irrational knee-jerk hatred for all things OCZ.

I don't think that word means what you think it means.

Here are a few reasons for disliking OCZ:

1) Biggest reason is the poor quality control and incompatibility problems:

OCZ SSDs have a 7% return rate, compared to less than half that (2.93%) for the next worst SSD manufacturer, and only 0.82% for Crucial. What is more disturbing, though, is that the only manufacturer with any models with higher than 5% return rate is OCZ, and OCZ has TEN models on the list with greater than 5% return rate. Ten models!

http://www.hardware.fr/articles/862-7/ssd.html

2) OCZ claimed to have used their own Indilinx controller in the Octane and Vertex 4, but has actually used a Marvell controller. This type of dishonesty and cover-ups is typical of OCZ.

3) OCZ changed the flash used in some of their SSDs so that the capacity is lower than advertised and the performance is lower than it used to be, and some of the flash they used was not standard IMFT quality, but they do not change the model name and do not inform anyone of the changes.

4) OCZ has a long history of dishonesty and scams. I'm not going to bother listing them here but you can find information on it if you search for it (although the MadOnion forums are gone). Here is a link to get you started:

http://www.tweaktown.com/news/1655/the_real_ocz_uncovered/index.html

 

[Rubycon]

I don't think that word means what you think it means.

Here are a few reasons for disliking OCZ:

1) Biggest reason is the poor quality control and incompatibility problems:

OCZ SSDs have a 7% return rate, compared to less than half that (2.93%) for the next worst SSD manufacturer, and only 0.82% for Crucial. What is more disturbing, though, is that the only manufacturer with any models with higher than 5% return rate is OCZ, and OCZ has TEN models on the list with greater than 5% return rate. Ten models!

http://www.hardware.fr/articles/862-7/ssd.html

2) OCZ claimed to have used their own Indilinx controller in the Octane and Vertex 4, but has actually used a Marvell controller. This type of dishonesty and cover-ups is typical of OCZ.

3) OCZ changed the flash used in some of their SSDs so that the capacity is lower than advertised and the performance is lower than it used to be, and some of the flash they used was not standard IMFT quality, but they do not change the model name and do not inform anyone of the changes.

4) OCZ has a long history of dishonesty and scams. I'm not going to bother listing them here but you can find information on it if you search for it (although the MadOnion forums are gone). Here is a link to get you started:

http://www.tweaktown.com/news/1655/the_real_ocz_uncovered/index.html

Don't forget the destructive firmware updates.