SSD Reliability

Silencing hard drives, optical drives and other storage devices

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

quicksilver
Posts: 23
Joined: Sun Aug 16, 2009 8:24 am
Location: VT

SSD Reliability

Post by quicksilver » Fri Apr 09, 2010 12:38 pm

I'm going to be purchasing a new laptop within the next month or so. If I decide to get an SSD I'll likely order the laptop with an HDD, then purchase the SSD separately so I can get one that's higher quality than what's usually bundled.

I do web development, so I do a lot of file searches and a lot of modifications to files, often doing search/replace on hundreds of files at a time. A fast hard drive will really help my productivity. I'd like a 5-year lifespan at a minimum.

I've seen some mentions of SSDs not being as reliable as HDDs. Can someone give me more info about that? I've found very little actual info on the subject so far. Are there any error-checking options?

lm
Friend of SPCR
Posts: 1251
Joined: Wed Dec 17, 2003 6:14 am
Location: Finland

Post by lm » Fri Apr 09, 2010 1:40 pm

If you want it to work for 5 years, get one that has a 5 year warranty. That's how I buy all my PC parts.

I assume you use a version control system, so you would commit your work to a version control server regularly, and thus the reliability wouldn't really be a problem?

quicksilver
Posts: 23
Joined: Sun Aug 16, 2009 8:24 am
Location: VT

Post by quicksilver » Fri Apr 09, 2010 1:52 pm

I do a lot of freelance quick fixes, so I usually just download the whole site, make a copy, then do the fixes on the copy, then upload any changed files. So I'll have a copy of the original site, and a copy of the changed site. Yeah, it's kind of crude but for that type of work it's easier.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Fri Apr 09, 2010 2:50 pm

I would like to know a bit more about this.

I don't think much about warranties. It is not unusual for weak products and week brands to be spiffed up with longer warranties. That is a well known and accepted practice in the auto industry.

I think warranties for these types of products are even more suspect. Most drives that fail under warranty are unlikely to be returned. And if they are, their equivalent replacement will likely cost much less than the original.

And the real damage, loss of data and time, is not covered.

I would expect that SSDs are either inherently more reliable or inherently less reliable than HHD. Even more likely, their failure modes are different and the conditions that contribute to failure are different.

Can anyone contribute anything more to this topic?

Eunos
Friend of SPCR
Posts: 378
Joined: Mon Dec 12, 2005 3:29 am
Location: Melbourne, Australia

Post by Eunos » Fri Apr 09, 2010 3:44 pm

I would question that they are less reliable than hard drives, but the difference is that SSD is a fast-moving area in which a lot of products are brand new and untested. I doubt there is sufficient data to address the original question when nobody would even have a 5 year old SSD.

If you dropped the laptop by accident, only an SSD will protect the data. This is one argument in favour of them. As for wear and tear or random breakdowns, there is no way of saying.

I can only stand by my previous recommendation to prioritise a quality backup system for important data, regardless of whether your main hard drive uses platters or flash cells.

quicksilver
Posts: 23
Joined: Sun Aug 16, 2009 8:24 am
Location: VT

Post by quicksilver » Fri Apr 09, 2010 5:05 pm

It seems the individual memory cells can wear out on SSDs? Many of them employ "wear leveling" to spread it around. Not sure how this translates in the long-term.

I regularly backup to an external drive. Even with backups, data loss or even worse, unknown corruption, would be a pain to deal with.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Fri Apr 09, 2010 6:07 pm

Only someone deep in the industry, doing engineering studies on SSD longevity for a major player could give you anything like a truly relevant answer at this point. HDD likely failures are known: too much heat or physical shock plus the odd random no reason death. W/ SSD, the first 2 of those factors are not likely to be significant (unless you are a real brute w/ them)... so you are on new ground.

Personally, I would treat it no differently than a HDD -- with gentle soft hands -- and keep regular backups. Also go with the best brand you can afford. Right now, that looks like Intel.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Fri Apr 09, 2010 8:05 pm

Every known SSD design (consumer grade 2.5" SSDs) has had failures either in the reviewers hands or in the customers hands. See viewtopic.php?t=57904 for some quotes just one one reviewers experiences.

but for what it's worth I'm using one on my work PC, on a web server, and plan to buy more for use in every PC and server I have access to.

I think if you want the most trusted design you go for Intel generation 2 drives 80GB or 160GB.

The next most trusted design are Indilinx based drives 64GB or 128GB.

Corsair SSDs have a 2 year warranty

OCZ Vertex Series 3 year warranty

Intel X25-E, X25-M, and X18-M SSDs 3 year warranty

Crucial M225 series has a 5 year warranty

Patriot SSDs Offer 10-Year Warranty (doesn't mean they'll work longer just means if they are in business they'll send you a replacement drive when yours fails)

Right now I'd go Intel 80GB gen 2 as the most robust drive and I'm not a fan of Intel. In fact outside of SSDs I avoid Intel products.

If you can live with the loss of capacity formatting the drive to less than full capacity will extend the life of the drive as wear leveling will have more spare blocks to work with. Anand recommends formatting to 80% of the advertised capacity on consumer grade drives for performance reasons but it'll help for longevity as well.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Fri Apr 09, 2010 8:18 pm

dhanson865 wrote:Ithink if you want the most trusted design you go for Intel generation 2 drives 80GB or 160GB.
What exactly causes you to say that? Any specific links or data?
dhanson865 wrote:Anand recommends formatting to 80% of the advertised capacity on consumer grade drives for performance reasons but it'll help for longevity as well.
1. I understand how reduced formating will speed up a mechanical hard drive. I don't understand now that speeds up an SSD. How does that work?

2. Is it that simple to do a small format on an SSD and it will automatically use the remaining to protect the other memory? That seems too easy?

3. Do any of the SSDs report back to you when they are running out of their backup space? That is a pretty important capability. It automatically makes the SSD safer.

Eunos
Friend of SPCR
Posts: 378
Joined: Mon Dec 12, 2005 3:29 am
Location: Melbourne, Australia

Post by Eunos » Fri Apr 09, 2010 9:17 pm

Here is a copy of my X25-M's data, after 5 weeks.

Image

Keep in mind, that 99% is likely to stay at the same number for months, making the anticipated life of the drive probably decades (in theory). Having written the equivalent of once to the entire drive, that leaves 9,999 out of 10,000 writes still good. 8) It's only really used for the internet and a few documents at this point.

For wear levelling to work optimally, it is advisable to have a large amount of free space. It is unclear whether there is any benefit to formatting to a lesser drive size (thus leaving the remainder for wear levelling purposes) or if it's good enough to format to full size, but only use a limited percentage of the drive. All these drives have spare area allocated anyway.
ces wrote:1. I understand how reduced formating will speed up a mechanical hard drive. I don't understand now that speeds up an SSD. How does that work?
It won't speed it up, but it will prevent write amplification from degrading its performance. This makes TRIM very important.

http://anandtech.com/show/2829

frenchie
Friend of SPCR
Posts: 1346
Joined: Tue Jul 08, 2008 4:53 am
Location: CT

Post by frenchie » Sat Apr 10, 2010 1:15 am

I've had my intel SSD for a few month and here is my data :

Image

It's still fast and responsive. My next build will have one if I can afford it :)

lm
Friend of SPCR
Posts: 1251
Joined: Wed Dec 17, 2003 6:14 am
Location: Finland

Post by lm » Sat Apr 10, 2010 3:51 am

quicksilver wrote:I do a lot of freelance quick fixes, so I usually just download the whole site, make a copy, then do the fixes on the copy, then upload any changed files. So I'll have a copy of the original site, and a copy of the changed site. Yeah, it's kind of crude but for that type of work it's easier.
You might be perfectly familiar with what I am going to say, but since I can't know that, how about this workflow:

Download whole site, make it the first commit to a version control repository that stays on your server.

As you make changes to your working copy, make commits every now and then.

When you are ready, you can get the "list of changed files" between the first and last version easily out of the version control system.

And the benefit is that no matter how your laptop gets to a non-working state, you can just checkout a new working copy to any machine from your server and continue work from where you made your last commit.

If your quick fixes are so quick that you would only lose a couple days work in any case, then maybe you could just ignore any risks because the effort of protecting yourself against the risk is more expensive than letting it happen.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Sat Apr 10, 2010 7:06 am

ces wrote:
dhanson865 wrote:I think if you want the most trusted design you go for Intel generation 2 drives 80GB or 160GB.
0. What exactly causes you to say that? Any specific links or data?
dhanson865 wrote:Anand recommends formatting to 80% of the advertised capacity on consumer grade drives for performance reasons but it'll help for longevity as well.
1. I understand how reduced formating will speed up a mechanical hard drive. I don't understand now that speeds up an SSD. How does that work?

2. Is it that simple to do a small format on an SSD and it will automatically use the remaining to protect the other memory? That seems too easy?

3. Do any of the SSDs report back to you when they are running out of their backup space? That is a pretty important capability. It automatically makes the SSD safer.
0 and 1. The things that make me say that are relatively well wrapped up in a series of articles anand has written from 2008 to today. There is more that I've read and my personal experience but if you read these articles you'll have a big chunk of it. I know it's a lot to read but it covers all the ground from the designs, how they work, how reliable they really are, pricing, bugs, you name it.
Intel X25-M SSD: Intel Delivers One of the World's Fastest Drives
by Anand Lal Shimpi on 9/8/2008

The SSD Anthology: Understanding SSDs and New Drives from OCZ
by Anand Lal Shimpi on 3/18/2009

The SSD Update: Vertex Gets Faster, New Indilinx Drives and Intel/MacBook Problems Resolved by Anand Lal Shimpi on 3/30/2009

Intel's 34nm SSD Preview: Cheaper and Faster?
by Anand Lal Shimpi on 7/21/2009

Intel X25-M G2: Dissected and Performance Preview
by Anand Lal Shimpi on 7/22/2009

Intel Forces OCZ's Hand: Indilinx Drives To Drop in Price
by Anand Lal Shimpi on 7/23/2009

The SSD Relapse: Understanding and Choosing the Best SSD
by Anand Lal Shimpi on 8/30/2009

The SSD Improv: Intel & Indilinx get TRIM, Kingston Brings Intel Down to $115 by Anand Lal Shimpi on 11/17/2009

OCZ's Vertex 2 Pro Preview: The Fastest MLC SSD We've Ever Tested
by Anand Lal Shimpi on 12/31/2009

OCZ’s Vertex Limited Edition Review & SSD State of the Union
by Anand Lal Shimpi on 2/19/2010

OWC Mercury Extreme SSD - First Look at a 50GB SandForce Drive
by Anand Lal Shimpi on 2/26/2010

Western Digital SiliconEdge Blue Review: WD Enters the Consumer SSD Market by Anand Lal Shimpi on 3/3/2010

Intel's X25-V & Kingston's 30GB SSDNow V Series: Battle of the $125 SSDs by Anand Lal Shimpi on 3/19/2010

6Gbps SATA Performance: AMD 890GX vs. Intel X58/P55
by Anand Lal Shimpi on 3/25/2010

Crucial's RealSSD C300: An Update on My Drive
by Anand Lal Shimpi on 3/25/2010
Intel has had the SSD to beat since 2008. In the PC industry when a player can sell the same product virtually unchanged year after year that product was a standout and becomes a safe choice.

There are newer drives that compete on the performance side vs the Intel design but they don't have the years of history and the number of drives in the field that Intel has.

1. How the 20% rule works. You are asking me to explain the inner workings of an SSD or cite an existing explanation. I don't have the patience to do either fully but here are a few links that cover part of this
http://www.anandtech.com/show/2899/5

http://www.anandtech.com/storage/showdo ... i=3667&p=3 discusses free space effecting speed on Trim enabled drives:
Performance in a TRIM enabled system is now determined not by the number of invalid blocks on your SSD, but rather the amount of free space you have. I went into a deep explanation of the relationship between free space and the performance of some SSDs here.

TRIM will make sure that you don’t have to worry about your drive filling up with invalid data, but it doesn’t skirt the bigger issue: dynamic controllers see their performance improve with more free space.

My rule of thumb is to keep at least 20% free space on your drive, you can get by with less but performance tends to suffer. It doesn’t degrade by the same amount for all drives either. Some controllers are more opportunistic with free space (e.g. Intel), while others don’t seem to rely as much on free space for improved performance.
presumably this issue is even worse if you don't use Trim (say an XP machine on the Kingston drive without using the Intel toolkit).
2. Again we go back to one or more of the anandtech articles above to see why that is. Use the links from answer 1. and add this to the mix

http://www.anandtech.com/show/2829/7


Uggh I haven't had breakfeast yet so my Googlefu is weak. There should be a better quote out there but I don't know where it is right now.

3. I guess the posts above showing the wear level in the SMART data answers this.

In short don't worry about the unused wear leveling space the drive will take care of itself but

Always backup your data. An SSD or Hard drive either one can brick totally and leave you without access to the data on that drive. Every day is International Backup Awareness Day.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Sat Apr 10, 2010 7:12 am

dhanson865 wrote:0 and 1. The things that make me say that are relatively well wrapped up in a series of articles anand has written from 2008 to today. There is more that I've read and my personal experience but if you read these articles you'll have a big chunk of it. I know it's a lot to read but it covers all the ground from the designs, how they work, how reliable they really are, pricing, bugs, you name it.
I wasn't talking about performance. I know that you can rely on the intel to be a good performer. And I have generally been following it (including Anand) since it first came out. I have just not seen anything about its reliability from a MTBF perspective - good or bad.

The MTBF is not so much a wear leveling issue. It is whether or not the SSD will break and stop working. My understanding is that even worn out 100%, and SSD will still read data, so that even worn out, the data is not lost.

The question is, how often do they break, if so how do they break, and when they do break how retrievable is the data. Do any of the articles you list address this with respect to Intel, or to any other SSD?

Eunos
Friend of SPCR
Posts: 378
Joined: Mon Dec 12, 2005 3:29 am
Location: Melbourne, Australia

Post by Eunos » Sat Apr 10, 2010 7:45 am

I believe SSD data is retrievable, but have heard that the cost to retrieve it is higher than for platter drives.

You may find this Intel file interesting, suggesting MTBF: 1,200,000 hours.

http://download.intel.com/design/flash/ ... 322296.pdf

Not content with their calculations, I will get back to you in 1,199,800 hours and let you know how I went. :D

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Sat Apr 10, 2010 9:31 am

ces wrote:
dhanson865 wrote:0 and 1. The things that make me say that are relatively well wrapped up in a series of articles anand has written from 2008 to today. There is more that I've read and my personal experience but if you read these articles you'll have a big chunk of it. I know it's a lot to read but it covers all the ground from the designs, how they work, how reliable they really are, pricing, bugs, you name it.
The MTBF is not so much a wear leveling issue. It is whether or not the SSD will break and stop working. My understanding is that even worn out 100%, and SSD will still read data, so that even worn out, the data is not lost.

The question is, how often do they break, if so how do they break, and when they do break how retrievable is the data. Do any of the articles you list address this with respect to Intel, or to any other SSD?
Did you read the quotes in viewtopic.php?t=57904 or do you remember them from reading the articles?

I don't think anyone outside of Intel, Crucial, OCZ, etcetera knows the exact failure rate of their drives.

We do know for a fact that they can fail in any number of methods. If seen references to:

1. totally dead doesn't respond to power at all
2. powers up but no data visible
3. the drive misreporting its size in MB instead of GB and data not being accessible
4. the drive working OK but data being wiped

I've heard people talking about how flash is supposed to fail to an unwriteable read only state but I've never seen anyone who actually knew of a drive in that state just references to flash documentation and Wikipedia articles.

I think if you ever have an SSD fail on you it will be all or nothing. I'm saying you'll never see a drive in the state where all the blocks are 100% worn out and the data is still readable. It's just extremely unlikely that you'll get into that state without hitting another problem first.

You could get there on purpose maybe if you setup lab conditions and sent writes to the drive in a consistent controlled manner and worked at keeping the drive from getting into a state the controller doesn't handle well but by the time you do it you'll have either spent thousands of dollars worth of equipment and staff to do so or years will have passed and people will scoff at your data at the end of the test.

You can mine sites like newegg for reliability data and you can mine the support forums but in both cases the manufacturer can get a message deleted or edited.

Assuming newegg data means anything to you we have (sorted by most reviews and only picking drives worth buying performance wise)

Intel 80GB retail 134 reviews 1 drive failed.
Intel 80GB OEM 193 reviews 6 drives failed.
combined we have 327 reviews and 7 drives failed or 1 failure for every 47 reviews (but we have no idea how many sold on newegg that didn't submit a review).

OCZ Vertex 30GB Retail 353 reviews 4 drives failed and another 17 reviews with what I'm guessing are compatibility issues. Hard to say what percentage of those are the SSD controllers fault.

OCZ Vertex 60GB retail is chock full of so many negative reviews I find it hard to categorize them all.

OCZ Agility 60GB retail is a similar story (it's the same controller as the vertex but just as importantly it's the same ODM)

Patriot Torqx 128GB retail 118 reviews about 50 dead drives (worse than the vertex and agility but easier to count)

Intel 160GB OEM 61 reviews 4 drives failed

OCZ Agility 120GB retail 54 reviews 4 drives failed

Intel 160GB Retail 52 reviews 0 failed some reviews about a preproduction firmware batch accidentally escaping the lab.

Crucial M225 64GB Retail 50 reviews 2 failed

Partiot Torqx 64GB retail 35 reviews too many failures for me to care about counting. Call it one in 3 reviews with some reviews mentioning multiple failed drives

OCZ Vertex LE 100GB retail 17 reviews 0 failed (not exactly a large sample size and the are all reviews less than a month old. These drives could start failing any day now but for now the record is clean).

Crucial C300 128GB retail 11 reviews 0 failed (again too early to call)

Crucial C300 256GB retail 10 reviews 1 failed (again too early to call)

Corsair Nova series 7 reviews 0 failed (again too early to call)

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Sat Apr 10, 2010 10:21 am

dhanson865

1. Yes I remember reading that thread. I in particular remember reading your posts in that thread. Seems like you may be one of the more knowledgeable people around on this subject, other than perhaps the engineers designing and manufacturing these SSD drives.

2. So basically, based on the imperfect self reporting data, the Intel clearly has fewer complaints. And the difference is not small (5% or less compared to 10% to 50%) though the 30G Vertex seems about the same or even better than the Intel.

The data you have, while imperfect, is likely to be the best available outside of the manufacturer.

3. I need IDE compatibility, so Intel (which tells you to use ACHI) doesn't work well for me.

The vertex works for me. What troubles me is that the difference between the 30GB Vertex and the 60GB Vertex. They seem to be almost on opposite sides of the bell curve.

4. I wonder how these statistics compare to those of HDD. That may be the only way to assess whether these SSD failure statistics are good or bad.

5. If the failures are easy to classify (ie the SSD is bricked) that is easy to understand.

If they are difficult to classify, does that mean perhaps it may be a lot of them have more to do with the user than with the hardware?

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Post by ~El~Jefe~ » Sat Apr 10, 2010 11:38 am

ces wrote:I would like to know a bit more about this.

I don't think much about warranties. It is not unusual for weak products and week brands to be spiffed up with longer warranties. That is a well known and accepted practice in the auto industry.

I think warranties for these types of products are even more suspect. Most drives that fail under warranty are unlikely to be returned. And if they are, their equivalent replacement will likely cost much less than the original.

And the real damage, loss of data and time, is not covered.

I would expect that SSDs are either inherently more reliable or inherently less reliable than HHD. Even more likely, their failure modes are different and the conditions that contribute to failure are different.

Can anyone contribute anything more to this topic?
Like Hyunai and 10 year warranties.... they break a lot in 10 years....

rust too. They suck, according to korean friends of mine. But, they havea huge warranty for a cheap price so it works for many.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Sat Apr 10, 2010 11:45 am

ces wrote:dhanson865

1. Yes I remember reading that thread. I in particular remember reading your posts in that thread. Seems like you may be one of the more knowledgeable people around on this subject, other than perhaps the engineers designing and manufacturing these SSD drives.

2. So basically, based on the imperfect self reporting data, the Intel clearly has fewer complaints. And the difference is not small (5% or less compared to 10% to 50%) though the 30G Vertex seems about the same or even better than the Intel.

The data you have, while imperfect, is likely to be the best available outside of the manufacturer.

3. I need IDE compatibility, so Intel (which tells you to use ACHI) doesn't work well for me.

The vertex works for me. What troubles me is that the difference between the 30GB Vertex and the 60GB Vertex. They seem to be almost on opposite sides of the bell curve.

4. I wonder how these statistics compare to those of HDD. That may be the only way to assess whether these SSD failure statistics are good or bad.

5. If the failures are easy to classify (ie the SSD is bricked) that is easy to understand.

If they are difficult to classify, does that mean perhaps it may be a lot of them have more to do with the user than with the hardware?
1. I'm just a concerned party. I buy a lot of hardware. I doubt I know as much as I should know. If I had more drives in hand I'd be glad to dig in and learn more. See 3 below.

2. This is up to a lot of interpretation see 5 as well. Reviews on newegg are hardly scientifically valid statistics. I do think Intel SSDs are better than the average SSD but I'm not basing that opinion on the newegg reviews.

3. Either you know something I don't or this AHCI thing isn't much of an issue. Admittedly the only PC/servers I've put an Intel SSD in are all Dells but I've never had to change a bios setting. Maybe I'll hit more issues as I try SSDs on machines with newer or older chipsets or with a different brand of motherboard but so far I don't know anything firsthand about this AHCI issue you mention.

4. Lets compare a couple of common drives:

wd3200aaks 1042 reviews I lost count after about 50 or 60 dead drives I wouldn't be too surprised by a 2% or 3% failure rate in the first year on these. I need to find something with less reviews I guess.

wd6400aaks and wd6401aals both have over 1000 reviews

WD5000AAKB 500GB 386 reviews about 40 dead.

WD20EADS 2TB 316 reviews about 70 dead.

I don't think you'll find any drive with a perfect record.

5. It's all about the hardware with SSDs. I don't think user error is causing SSDs to brick. The hard to classify stuff is when it could be the motherboard SATA controller not playing nice with the SSD sata controller.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Sat Apr 10, 2010 12:08 pm

dhanson865 wrote:Either you know something I don't or this AHCI thing isn't much of an issue.
It is highly unlikely that I know something you don't know about this subject.

I read someplace that the Intel website tells you to use the Intel SSDs only with ACHI settings.

I have read, and it has been confirmed to my by Vertex on a telephone support call, that I should use their SSD drives only with IDE settings. That included the Vertex LE!

Now,m strangely enough I have read that the 1156 H55 chip set does not implement ACHI (though the H57 chipset does). That seems strange and it may be that many motherboard manufacturer's just add that capability on their own.

I also know someone (who had an 1156 H55 board) who couldn't get an Intel to work properly with it. He even had data corruption problems. He ended up returning and exchanging the Intel SSD for something else. I wonder if his problem was the ACHI vs IDE problem.

psyopper
Friend of SPCR
Posts: 336
Joined: Fri Mar 28, 2008 1:15 pm
Location: Portland, Oregon

Post by psyopper » Sat Apr 10, 2010 2:51 pm

As for H55 AHCI, my Gigabyte UD2H supports it in the BIOS, I think... I installed Windows 7 with the drive (6400 AAKS) in IDE mode then switched the controller to AHCI mode and it found the drive but windows failed to boot.

Eunos
Friend of SPCR
Posts: 378
Joined: Mon Dec 12, 2005 3:29 am
Location: Melbourne, Australia

Post by Eunos » Sat Apr 10, 2010 4:11 pm

I think switching to AHCI mode is more complex than changing BIOS settings. You may need to fully replace all data on the drive in the process.

My motherboard doesn't even support AHCI, but my research indicated that there is only a slight performance penalty under some circumstances. Saying ACHI is essential to an Intel SSD is very misleading. Mine scores a 7.5/7.9 on WEI which is about as good as a non-RAID set-up can get.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Sat Apr 10, 2010 4:17 pm

Eunos wrote:I think switching to AHCI mode is more complex than changing BIOS settings. You may need to fully replace all data on the drive in the process.

My motherboard doesn't even support AHCI, but my research indicated that there is only a slight performance penalty under some circumstances. Saying ACHI is essential to an Intel SSD is very misleading. Mine scores a 7.5/7.9 on WEI which is about as good as a non-RAID set-up can get.
1. So you are using the intel in IDE mode?
2. Are you telling me that if I move an esata HHD or SSD drive between an am IDE configured computer and an AHCI configured computer, I can expect problems?

Eunos
Friend of SPCR
Posts: 378
Joined: Mon Dec 12, 2005 3:29 am
Location: Melbourne, Australia

Post by Eunos » Sat Apr 10, 2010 4:35 pm

Yes, and I think my X25-V is also in IDE mode. The loss of functionality is absolutely minimal compared to the benefits.

I cannot comment on an eSATA configuration. I'd suggest doing some general internet searches with keywords like 'SSD AHCI'. Here is one good discussion:

http://www.overclockers.com/forums/show ... p?t=639214

Cryoburner
Posts: 160
Joined: Sat Dec 01, 2007 4:25 am

Post by Cryoburner » Sat Apr 10, 2010 10:23 pm

dhanson865 wrote: You can mine sites like newegg for reliability data and you can mine the support forums but in both cases the manufacturer can get a message deleted or edited.

Assuming newegg data means anything to you we have (sorted by most reviews and only picking drives worth buying performance wise)

...
You didn't mention the 40 GB Intel X25-V, which is another somewhat lower-cost option. It's currently out of stock at Newegg, but of the 144 reviews, I didn't notice anyone mentioning outright failure of the drive. There were a handful of negative reviews about apparent compatibility (or user) issues when trying to install Windows, and one person who reported blue-screens when going into sleep mode (although he didn't have a tag indicating he actually bought the drive from Newegg). If you don't need much data storage, the 40 GB drive might be a decent option, even if its performance is a little lower than the higher capacity versions.

Of course, as was said, it's difficult to say how accurate user reviews from online sites actually are. In addition to potentially being able to remove negative reviews from a site, it wouldn't take much for a drive manufacturer to post a bunch of positive reviews for their product at various online stores. For a popular site like Newegg, they could even buy up a bunch of their own drives to get the review tag indicating they bought them there. They could even buy new drives from competitors to write negative reviews harming their reputation. Maybe they even have accounts at this site, to provide helpful feedback when needed. It doesn't seem all that improbable. : )

In any case, I probably wouldn't worry too much about the lifespan of a decent SSD compared to a traditional hard drive. While not much long-term data is available, thus far many SSDs appear to have a comparable or lower failure rate than most hard drives. In the case of notebooks, the increased shock tolerance is a definite plus.

dhanson865
Posts: 2198
Joined: Thu Feb 10, 2005 11:20 am
Location: TN, USA

Post by dhanson865 » Sun Apr 11, 2010 7:18 pm

Cryoburner wrote:
dhanson865 wrote:
+1 for what you said. not fully quoted to save space in the thread.

FWIW out of stock items are the bane of newegg searches. They drop to the bottom when I sort by price or rating or whatever and I am often in a hurry when I do a comparison.

The 40GB Intel X25-V and it's sibling the Kingston SNV125-s2/40gb are slower performing but presumably identical in reliability as the 80GB Gen 2 drives. I have 2 PCs and a Server at work using the Kingston variant and if the Intel 40GB ever dropped to the $50 price point while the Gen2 80GB is still above $100 I'd probably buy a stack of them.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Sun Apr 11, 2010 8:25 pm

Cryoburner wrote: Of course, as was said, it's difficult to say how accurate user reviews from online sites actually are. In addition to potentially being able to remove negative reviews from a site, it wouldn't take much for a drive manufacturer to post a bunch of positive reviews for their product at various online stores. For a popular site like Newegg, they could even buy up a bunch of their own drives to get the review tag indicating they bought them there. They could even buy new drives from competitors to write negative reviews harming their reputation. Maybe they even have accounts at this site, to provide helpful feedback when needed. It doesn't seem all that improbable. : )
I think that is unlikely. That is the kind of thing you can't hide for long. Once it gets out, the self inflicted damage would be immeasurable. The person who did it would have to be sacrificed.

The greater source of inaccuracy is that more of the people who buy them don't report anything. And I could easily see the reporting rates differing from brand to brand.

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Sun Apr 11, 2010 8:28 pm

dhanson865 wrote: The 40GB Intel X25-V and it's sibling the Kingston SNV125-s2/40gb are slower performing but presumably identical in reliability as the 80GB Gen 2 drives. I have 2 PCs and a Server at work using the Kingston variant and if the Intel 40GB ever dropped to the $50 price point while the Gen2 80GB is still above $100 I'd probably buy a stack of them.
My understanding is that the 40G Kingston/Intel does not have trim while the Intel/Intel does. And also that the 40G Kingston/Intel is no longer available.

ascl
Posts: 279
Joined: Tue May 05, 2009 1:15 am
Location: Sydney, Australia

Post by ascl » Mon Apr 12, 2010 4:16 am

I am currently in the midst of issues with my OCZ Vertex 60. One thing to keep in mind is that the firmware is very new (relative to HDDs) in these things, and it borked firmware can and will cause data loss and corruption.

While my sample size of one is not exactly conclusive, these types of drives are relatively new, and it will be a while before the firmware is properly reliable.

In the meantime... backups are mandatory (and should be anyway).

ces
Posts: 3395
Joined: Thu Feb 04, 2010 6:06 pm
Location: US

Post by ces » Mon Apr 12, 2010 12:35 pm

ascl wrote:I am currently in the midst of issues with my OCZ Vertex 60. One thing to keep in mind is that the firmware is very new (relative to HDDs) in these things, and it borked firmware can and will cause data loss and corruption.

While my sample size of one is not exactly conclusive, these types of drives are relatively new, and it will be a while before the firmware is properly reliable.

In the meantime... backups are mandatory (and should be anyway).
What version of the firmware is giving you problems?

Post Reply