Looking to buy a new Thinkpad T Series

[ThinkRob]

To Thinkrob:
My Blade server friend tells me that the "MLC" drives that are used at an enterprise level are "not complete MLC" and "they are based in tiers." As well as how 32GB of reserve and mixed NAND with engineering can go a very long way.

You realize that a number of consumer-level drives can (and do) use spare/reserve space as well, right? And that IBM offers SSD solutions based on SandForce controllers (yes, the same SandForce that you mentioned before)?

And what does "based in tiers" mean?

I really don't quite see how you could "save hours a day" by using an SSD.

I can't speak for anyone else, but I can answer how the above is true for me: builds. Complex project builds that are very I/O-bound benefit tremendously from SSDs. Even a kernel compile will benefit quite a bit from an SSD, as it's mostly small random I/O -- and with a modern multi-core processor, the speed-up is even more pronounced.

[Navck]

"There are different classifications of SSD drives that OEMs spec by, each one is different per OEM and supplier; they validate SSD drives in their systems and bin in tiers. It is literally like part binning for HDDs
Except SSDs are a little more sensitive to deviations and variations like
Say
....From Bin A to Bin B
In performance of a given component
They do this so they can ensure, as suppliers of servers or whatever, that their systems are integrated smoothly enough for their end user client
In some ways, this is easier with HDDs; SSDs require a different approach
And are typically pretty intolerant of variations *if*, say, your interface engineers did something different and the SoC is doing sub-optimal wear levelling that was a trade off for performance
Because the interface clocking is slightly under-spec compared to what the OEM wants
But if your MLC media has a larger sigma in the thresholds where the pages start depleting from wear
Then that could be significant in actually accelerating wear
And/or appearing as a hit in performance
As the SoC suddenly decides to save a ton of pages and block it just wrote
It gets very complicated
Usually, they have engineering teams doing that
It's a little, in some respects, more straightforward for an HDD
Because even though there are more variables and components
That also means there are more degrees of freedom in changing things with them
And tweaking stuff
To hit spec
That and you can model a HDD with pretty high accuracy and precision with a good turn around time
This is not the case with an SSD; each complete step in this cycle requires a validation
Therefore, each OEM comes up with tiers on their ends to match the performance of what they have validated from their SSD supplier
As an analog of binning parts in an HDD
Whether or not you, as the buyer, know about this is a little bit up to how large a contract you are
And how much you pay
This is like when RAM first became dual parity
And ECC"

Quoted from my friend with the Blades.

[Colonel O'Neill]

I can't speak for anyone else, but I can answer how the above is true for me: builds. Complex project builds that are very I/O-bound benefit tremendously from SSDs. Even a kernel compile will benefit quite a bit from an SSD, as it's mostly small random I/O -- and with a modern multi-core processor, the speed-up is even more pronounced.

Well, wouldn't it be even better to have the kernel compile work on a large RAM cache and just dump the results to disk at the end? It'd alleviate excess write cycle usage as well as give the compiling a massive boost in speed (assuming it scales in direct proportion to random I/O speeds on the storage media). Given how often one would probably need to compile a kernel (I used to dabble in VB6; bugs are a massive pain), wouldn't RAM be more ideal in that particular case?

[uux]

Google's study deviates from what the harddrive industry agrees on: Lower temperatures means lower rate of failure.

You're missing the point. Hard drives fail, even in perfect conditions. No one is kicking the servers over at Google. It's not user error. They fail.

I'm sorry that you don't like how the industry likes to test data retention, but you realize that your computer does generate heat, right? Arrandale processors are rated UP TO 105C, which would be that "oven" you mock. Ever consider why your Thinkpad spins the fan even when the system is cool? Circulating stagnant hot air pockets in the system could be a vital function here.

Who said anything about liking or disliking how the industry likes to test data retention? Don't put words in my mouth. You actually proved my point here. Your ThinkPad is engineered NOT to create an oven. Your examples are getting more ridiculous everytime you post. Are you seriously using heat as a disadvantage to a SSD? Heat isn't favorable to a platter based drive either. Most SSDs are rated to function in more extreme temperatures than your traditional HDD.

What, you mean a full RAM disk setup with a backup battery? SLC drives are not cheap.

They are cheap. I take it you're not familiar with that segment.

Whatever, for some reason when I make any valid criticism that comes from STORAGE ENGINEERS, every fan on the bandwagon feels the need to elevate SSDs onto the pillar of "best storage option."

Quote me where I said SSDs are the best storage option. They have their benefits, champ. Like it or not, there are applications they excel at for a much lower cost than traditional platter based drives.

PS: Dust in the breather hole? HA! Some drives have multiple breather holes AND they have filters in them, good luck plugging it up somehow. I won't address the airplane issue because of modern fly height adjustment with the head and how it can actually INTENTIONALLY contact the platter for short durations. Didn't know that? Guess what, modern drive platters have a thin layer of lubricant as part of the platter, the head CAN touch the disk without damage ("Hydroplaning") during reads and writes. Go find out some information from actual engineers in the storage industry about harddrives.

I'm aware of the filters. They can get damaged quite easily. You spoke of the photographer who doesn't know jack about computers. Do you think he keeps the inside of his computer dust free? Yes, I did know that about modern fly height adjustment with the head. I'll throw your own argument back at you. Why do you think that's needed? Does this remind you at all of your argument about wear leveling being needed to point out the disadvantage of limited write endurance in flash NAND cells? It's kind of comical to read your arguments. In your examples, HDDs will always work perfectly as intended, but SSDs will always fail because they're always running into the worst case scenario. You keep contriving ridiculous examples as arguments.

Did you ever own an early Maxtor drive? An abundance of lubrication on the platter actually caused the lubrication to coagulate and cease the head from moving. They're not perfect, with perfect quality control, etc. You can say what you want about Maxtor, but all of the manufacturers have had quality issues and bad batches at one time or another. I don't trust their reliability anymore than a SSD. Most modern HDD manufacturers stopped using liquid lubricant as well. See here. Even if the right amount of lubrication was used, the repeated heating and cooling as well disks not being level could cause coagulation. Fun stuff. I don't know what your engineers tell you, but liquid lubricant on the HDD platter was a bad idea.

Also, any damage that PHYSICALLY damages the magnetic media? Anything there is GONE, you cannot "repair" that damage.

That's right. That was my point. You'll be lucky if that doesn't happen. Head crashes are one of the more common failures in laptop HDDs.

[uux]

I really don't quite see how you could "save hours a day" by using an SSD. Either you're clearly overexaggerating the benefits, or somehow, this SSD makes your 24 hour day a 30 hour day.

You said it yourself. You don't quite see how. Yet, you call it an exaggeration anyway. You do understand that accessing random data is WAY faster on a SSD due to the lower seek times, don't you? Now put a 5GB tree of source code (random small files) on your disk. Use CVS or SVN to update said tree. My SSD does it at a minimum of 20 minutes faster. The computer also remains completely responsive due to the ability of the SSD to handle higher IOps. Yes, a fast SAS or SATA drive can do it in comparable time. The difference? I can't get a disk that fast in my laptop. The SSD actually sped things up so much, that I keep a mirror of the local CVS repository on my SSD, and simply check out from that. It's extremely painful to do that on a consumer level laptop HDD. I used to have multiple computers to gain the same benefit. They're not needed anymore.

I also don't quite see how you only manage 1GB of writes a day; I did a rough sampling of file writes using Process Monitor, did some calculations, and ultimately ended up doing approximately 80MB of file writes in 20 minutes just doing some moderately intensive web surfing and MSN chatting. If I extrapolate this to using my computer for about 5 hours a day (on average), I get 1.2GB a day. Those conditions are really light to me. I just did a 7x86 install in VMWare to check out that weird new KB update that was pushed out this morning. There goes another 8GB in an hour.

1GB on average, over a period of about two years. That's what the SMART attributes say. Some days I would write more than 40GB, others I may have only wrote 100MB or so at best. I work with that 5GB tree of source code everyday, multiple times a day. Most of that function is reading, not writing. As I said, people here will probably call it light usage. But you know what? That only solidifies my decision to use a SSD in my laptop. It works for my application (low writes), and I actually gained from the advantages the SSD had.

[ThinkRob]

Well, wouldn't it be even better to have the kernel compile work on a large RAM cache and just dump the results to disk at the end? It'd alleviate excess write cycle usage as well as give the compiling a massive boost in speed (assuming it scales in direct proportion to random I/O speeds on the storage media). Given how often one would probably need to compile a kernel (I used to dabble in VB6; bugs are a massive pain), wouldn't RAM be more ideal in that particular case?

In that particular case (kernel compilation), yeah, tmpfs might be a good idea. I was actually referring more to my day-to-day dev. work which involves a couple of very ungainly source trees which aren't really good candidates for that approach.

[ThinkRob]

Quoted from my friend with the Blades.

It's almost as though various levels of specification, testing, verification, error correction, and R&D are applied to magnetic and solid state storage so as to allow a range of products to be produced to satisfy the varying reliability and performance needs of both the consumer and enterprise spaces...

Let's face it: enterprise storage makers clearly think that SSDs are viable. They ship both SLC and MLC-based drives, in some cases using controllers from the very same company used by manufacturers of consumer SSDs. A number of server manufacturers offer SSD storage as a supported option, using both MLC and SLC-based drives.

Now either you know something that these folks don't, or a large and traditionally rather risk-averse portion of the industry believes that SSDs are ready for mainstream usage.

[Navck]

UUX: I'm sorry to say this... But your knowledge about harddrives is very wrong.

You pointed at Google's study that was flawed. I was pointing that oue.
My T43 has been in operation for *five* years with a full interior cleaning after ~ 3 years in service, yet the harddrive's breather hole has not "clogged" in such a manner to cause the drive to malfunction.
The lubrication layer on harddrive platters used TODAY is closer to an epoxy like coating that STILL can evaporate and then condensate back onto the platter again when you exceed temperature thresholds. Some drives from Seagate tried to mitigate the lubrication buildup on the head by doing "cleaning" when the drive is idle.
I hope you realize your SVN is also limited by the efficiency of the client, server and the connection between those two too, as with any other coding that would cause unnecessary cycles to be wasted making the user BELIEVE the harddrive is at fault.

Apparently the engineers I speak to are employed by the industry and are paid to do what they do daily. Do you engineer harddrives? I just checked with my Blade server friend and he feels you're throwing "chaff and flares" in this discourse.

[uux]

UUX: I'm sorry to say this... But your knowledge about harddrives is very wrong.

You pointed at Google's study that was flawed. I was pointing that oue.

I pointed out the Google study not due to the contents of the study, but to show you that even in perfect conditions that a traditional hard disk drive will fail. I even spelled it out for you in my last post. It was due to your comment about banging on the computer. I'm sure some people do, but you paint the picture of every hard drive that has failed is due to it. I'm sorry, but as much as you hate the marketing of SSDs, and people who think they're perfect for everything, you've shown repeatedly to be fanatically against them to the point of ignoring reason. You are at the extreme opposite end of the SSD fanboys.

My T43 has been in operation for *five* years with a full interior cleaning after ~ 3 years in service, yet the harddrive's breather hole has not "clogged" in such a manner to cause the drive to malfunction.

You are not the photographer who knows nothing about computers. You're example keeps changing dramatically. I gave a T42 to my sister. Perfectly clean. Within six months, it was completely filthy inside. She is the photographer who knows nothing about computers. She sat it on her carpet while it was running, slid it under her couch when it wasn't being used (while running), etc. It happily pulled in every dust partical that it could. The HDD is fine. It's well covered and isolated. Your T43 won't easily get dust inside of the HDD area. But how about those generic towers that manufacturers sell? Many do not have filters over the intake fan. They happily suck in dust, hair, and whatever.

The lubrication layer on harddrive platters used TODAY is closer to an epoxy like coating that STILL can evaporate and then condensate back onto the platter again when you exceed temperature thresholds. Some drives from Seagate tried to mitigate the lubrication buildup on the head by doing "cleaning" when the drive is idle.

/shrug. I won't pretend to know what is used today. I simply know traditional hard disk drives are fragile. They're an engineering marvel, but still fragile nonetheless. You recognize above one unpredictable method in which your hard drive can just randomly fail. Don't get me wrong either. I'm just playing an adversary to your arguments because just like SSDs, HDDs have many shortcomings. You can't fault one and ignore the faults of the other.

I hope you realize your SVN is also limited by the efficiency of the client, server and the connection between those two too, as with any other coding that would cause unnecessary cycles to be wasted making the user BELIEVE the harddrive is at fault.

Quote me saying the hard disk drive was at fault. Can't? It's because I never said it was. CVS and SVN are terribly slow and inefficient. Though, that's what I have to work with. I also acknowledged the SSD has no benefit over faster HDDs for this purpose. But to reiterate, that's not possible in a laptop. Also, importing the 15 years of history into something more modern is non-trivial. None of that is the point, which you're missing again. I see a beneficial gain from using a SSD. I actually need less overall equipment to do the same task, finish the task quicker, and it doesn't cause an interruption (I don't need to wait for it to complete, the computer remains responsive). It is not in any way detrimental. At it's current rate (if you simple go by the SMART attributes and ignore other random failures), it will last a LONG time.

Apparently the engineers I speak to are employed by the industry and are paid to do what they do daily. Do you engineer harddrives? I just checked with my Blade server friend and he feels you're throwing "chaff and flares" in this discourse.

Do I engineer hard drives? Do you? You stated above that all your knowledge is second hand. You obviously do not. Am I no longer qualified to speak about my experience with consumer based SSDs because I do not engineer hard drives? Where are you going with this, guy? Shall I list my storage engineering friends? It's an impressive list. This isn't a contest. Here is a question for you. Have YOU actually ever used a consumer based SSD? Or is all your knowledge of their shortcomings and failures due to horror stories on the internet and what your engineering friend is telling you? You know, kinda like how you make it seem like SSD users are mindlessly and blindly following the advice of marketing and sales people.

Ask your storage engineer friend what equipment he would throw together for 200,000 IOps. Price it out. You brought up the enterprise segment. Look at the cost of a traditional platter based solution vs something solid-state. You don't even have to answer it here.

It's good that you're out there pointing out the disadvantages of SSDs. They're not for everyone and not for every purpose. Though, they can be quite useful and beneficial when used for what they're good at. You just seem fanatically against them at all costs.

[ThinkRob]

Another benefit of SSDs: I can use them on a bumpy train ride without HDAPS losing its little mind.

Seriously though, the ability to use your laptop normally and with full performance whilst in a moving vehicle is a big feature in my book. With a conventional drive, I'd either have to disable HDAPS, turn the sensitivity down to the point where it's basically off, or simply live with it parking the heads every couple of seconds (which isn't exactly great for the drive's lifespan...) With an SSD? No problem -- it's completely impervious to the bumps and shocks of your average train ride.

[Navck]

APS -> Disable
Car/Train suspension + "Human padding" = Enough to prevent the HDD from experiencing 300g shock AND they're more vulnerable to high frequency vibrations.

The photographer that doesn't know about computers uses a Macbook (Earlier generation with removable battery compartment, C2D), he went through several drives from what I suspect was thermal issues (Definitely not vibrational, that computer is babied in a case) and I am going to watch him kill his brand new X25-M soon. My T43 had a nice coating of microdust all over the interior with chunks of dust that built up into little pieces of "fur" that needed tweezers to remove from the PC card bay. That was a fun experience. My desktop (Closer to workstation) with FOUR harddrives in service also gets dusty even with mesh AND foam filters layered together due to the amount of airflow that is inhaled and exhausted out by the eleven fans now.

The fragility of harddrives is extremely overstated, ask a data recovery professional about how many logic board swaps they do. Or how about that Seagate firmware bricking drives? The drive still works mechanically.

As for headcrashes in laptops, I suspect those laptops have a head crash and probably some "other problems" like a "cracked chassis." Harddrives are incredibly robust (See: Specsheets) against shock these days, so you would have to impart some significant forces to kill one (Ex: 3 inch fall onto nice, hard Teak. Not like 3 inches suspended within rubber assemblies into carpet)

You might have your storage friends but they seem to be awfully silent to you about harddrives. As for my "SSD experience" I get to deal with the people who feel the urge to "prove something" using their wallet only to let me mock them when they come up short on their claims. Later some of these people will come back to me with their "investment" and ask for me to help them fix it when they do something very stupid.

For what you and zhenya have described for usage habits, SSDs would more likely be for the premium consumer market. But hey, there is a company that actually agrees with that: Apple.

[uux]

As for headcrashes in laptops, I suspect those laptops have a head crash and probably some "other problems" like a "cracked chassis." Harddrives are incredibly robust (See: Specsheets) against shock these days, so you would have to impart some significant forces to kill one (Ex: 3 inch fall onto nice, hard Teak. Not like 3 inches suspended within rubber assemblies into carpet)

You make a lot of assumptions. You know the saying about that, right?

"These days". I still have hard disk drives in use that do not have all those protections that you enjoy "these days". I still have laptops in use that do not use APS. I wouldn't even consider them old. Why do you think they're trying to improve those qualities in hard disk drives? Perhaps it's due to some shortcomings of the technology? Like it's a fact that mechanical failure will happen. Just like in a SSD. The write endurance limit will happen. The data retention problem will happen. There are more important things to worry about.

You might have your storage friends but they seem to be awfully silent to you about harddrives.

*sigh* Personal attacks don't make you or your argument look any better. I don't need a storage engineer to consult with to be able to respond with you about flash based SSDs in the consumer market. It's funny how your storage engineer went from being a storage engineer to a blade server guy. Has your storage engineer guy ever engineered a hard disk drive? I already know the answer. Apparently I don't know anything if I haven't, which means your storage guy doesn't either. So, at best, you're information is third hand. Let me guess, he sells blade servers. A SSD based solution would mean he sells less servers. Am I right? You're a victim of the marketing machine and you don't even realize it.

Hey, here is one for you and your storage engineer. Who do you think designed and engineered the X25-M I'm using? I'll give you a guess since you're having trouble with it from your posts. It wasn't me. Do you think the people who designed it are less qualified than you, or your storage engineer? Do you think it was designed by people who are not paid to do so and who are not in the industry?

As for my "SSD experience" I get to deal with the people who feel the urge to "prove something" using their wallet only to let me mock them when they come up short on their claims. Later some of these people will come back to me with their "investment" and ask for me to help them fix it when they do something very stupid.

So, no, you've never used a consumer flash based SSD as a general storage device. You have no experience on your own with your own device. None. Absolutely none of your criticisms are from experience. Yet, you're going to pretend to be authoritative on the subject based on hearsay and what some fool has done. You know what they say about a fool and their money, right?

For what you and zhenya have described for usage habits, SSDs would more likely be for the premium consumer market. But hey, there is a company that actually agrees with that: Apple.

And hey, you're wrong. Guess what else? You're still wrong. Oh, you know what? You're wrong.

I spent less money overall using the SSD. It's working for me. It's working for zhenya. ThinkRob is enjoying his. Why does that bug you so much? It's comical. I feel like posting pictures of me holding up my SSD and cheering, etc, just to bug you. Maybe I can get a shot of the office staff giving it a group hug.

Here is an excellent site for anyone that wants real information about the issues. link

[Navck]

You realize that there are IBM RAMACs in museums that could still be spun up and used long after their life in a business operation... Right?
Given mechanical failures "happen," harddrive lifespan is much, much longer than you believe.

I'm going to ignore your response toward my sources of knowledge because you seem to be offended by me.

Hours of weeding out crapware from people's computers with SSDs? Oh I guess that never happened to you anyways.

I don't know how you "use less money overall" but I'm not going to bother to talk about that.

That website is filled with strong amounts of SSD marketing material, by the way. Ever consider looking at it critically and realizing the issues I described which, still happen to exist BECAUSE NAND IS THE MEDIUM USED IN YOUR SSD is there? Go look up the specsheets on NAND used within some SSDs and realize some of the MLC has only 3k write cycles on it.

Go ahead, post your pictures if you need that "self pat on the back" to feel confident to continue this discourse with me. I'll merely just respond when I have the time to.

[uux]

You realize that there are IBM RAMACs in museums that could still be spun up and used long after their life in a business operation... Right?Given mechanical failures "happen," harddrive lifespan is much, much longer than you believe.

You don't get it. You can't predict the lifetime of the hard disk drive. It can fail for any number of reasons. It can also exceed the lifetime of a flash based SSD. That doesn't make a SSD a less viable solution. Just be aware that it's limited lifetime is a matter of when, not if.

You brought up data recovery several times. Have you ever been quoted on the cost of recovery for say something that sounds like a simple mechanical failure? The SSD is cheaper. Don't take my word for it. Call up a data recovery specialist and ask them. In fact, I can quote you as saying that is one of your sources.

I'm going to ignore your response toward my sources of knowledge because you seem to be offended by me.

Not at all. As I"ve stated, I find it comical. You are the one who made credibility an issue. I'm calling you out on yours. If you don't like it, then perhaps you shouldn't have brought it up?

Hours of weeding out crapware from people's computers with SSDs? Oh I guess that never happened to you anyways.

No, that does not happen to me. I do not fix other peoples' computers. So, you have first hand experience with flash based consumer level SSDs failing while you're weeding out said crapware? And you've had these peoples' computers in your possession for a long enough period of time that you consider yourself "experienced" with using a SSD? You've had one in your possession for "years" like zhenya, ThinkRob, and myself? In fact, you're usage of them exceeds zhenya's or my own?

I don't know how you "use less money overall" but I'm not going to bother to talk about that.

You don't know? Even after I already explained it in several posts? Really? It took more equipment to do what the SSD does. Plain and simple. That adds up to a higher cost than the SSD. Shall I say it again? This is a written forum, you know. You can read it more than once if you need to. Did you ask your storage engineer on several solutions to the 200,000 IOps question? That should have shown you where I was going with less cost.

That website is filled with strong amounts of SSD marketing material, by the way. Ever consider looking at it critically and realizing the issues I described which, still happen to exist BECAUSE NAND IS THE MEDIUM USED IN YOUR SSD is there? Go look up the specsheets on NAND used within some SSDs and realize some of the MLC has only 3k write cycles on it.

Read the site. The author is quite critical of flash based SSDs, especially earlier models, the notebook segment, using them as long term storage and archival purposes, etc. Of course you might disagree with him, but then again he's not fanatically against them and also views them as being useful. Nothing on that site really disagrees with what you have stated. It's just written more reasonably.

[ThinkRob]

I think we're slowly moving from a discussion of SSDs to a flamewar.

[killer]

I think we're slowly moving from a discussion of SSDs to a flamewar.

Good thinking, matey. Probably this has been done to death.

[Colonel O'Neill]

Or more bananas.

[killer]

Not only a great comment, Colonel, but also an apposite photo.

[JaneL]

I think we're slowly moving from a discussion of SSDs to a flamewar.

I agree. Navck, Uux, you guys know better. Take it outside.

Thread locked.