Thinkpads Forum

Hi Y'all,

I have a T410, i7-620M, 4 gigs RAM, Win-7-64bit and a month ago installed a
120GB OWC Mercury Extreme Pro SSD:

http://eshop.macsales.com/item/Other%20 ... /SSDMX120/

Boots fast.
Quiet.
Three year warranty.

Crystal Disk reports 4K writes at 8.747 MB/minute.

What is more reliable/better/faster/bigger/cheaper?

Thanks and good health, Weogo

3ry warranty sounds nice > love to see your "Intelligent Block Management & Wear Leveling " reports after a YEARS USE!!

I have Intel SSD - and the wear = a major issue for me (us) as most of the laptops = under very very heavy RAM&HD USE (both read and write modes) and most Intels last us 7-9mths before they 'death alert' as we call it! We even got some 'pre release' intel SSD's that are 1TB and 1.5TB THEY LAST EVEN LESS at the moment !

So be very keep to see your experiances

Weogo wrote:What is more reliable/better/faster/bigger/cheaper?

I'll give you something equally reliable, just as good, +/-1 as fast if you know what you're doing but bigger and cheaper:
A spinning platter that has a magnetic material below protective coatings and a lubricant layer that is read by a swinging arm mechanism.

But in seriousness, SSDs are not cheap and they will burn you later if you use them intensely.

Otherwise, if you don't care of your system and don't really do maintenance (Defragmentation) then a SSD is for you.

ausmike wrote:We even got some 'pre release' intel SSD's that are 1TB and 1.5TB THEY LAST EVEN LESS at the moment !

I'm sorry, a 1.5TB drive? I have a tough time believing that you've got some (unheard of) Intel hardware that's almost ten times the capacity of their largest released drive.

I have Intel SSD - and the wear = a major issue for me (us) as most of the laptops = under very very heavy RAM&HD USE (both read and write modes) and most Intels last us 7-9mths before they 'death alert' as we call it!

With all due respect, I think you're something of an outlier then. Not only is that well before the warrantied (and tested) lifespan, but it's also quite the opposite of what many users have been reporting: that Intel's drives have exhibited good longevity with many GB of writes per day.

For almost all desktop uses, the X25-M seems to be the current "best blend" of reliability and performance. It's not the fastest (though it's close), but it does have a couple years of (public) vetting, and is rated for something on the order of 10 or 20 GB (I can't recall which) of writes per day for several years. Yes, if you are to write enough to them they'll fail (as will *all* storage devices), but you're very, very unlikely to hit that with normal desktop usage.

the X18-m in my ThinkStation is averaging 18.04 GB worth of writes per day and it hasn't slowed down one bit.

just sayin'.

erik wrote:the X18-m in my ThinkStation is averaging 18.04 GB worth of writes per day and it hasn't slowed down one bit.

just sayin'.

Yeah, I've got a first-gen X25-M that sees at least 10GB of writes/day and is now into its second year of life with no problems whatsoever. Is there a performance drop? Yes -- but you won't notice it unless you benchmark. Anandtech talked about this at great length, and found that Intel's drives suffer performance degradation far, far less than many other designs. It's basically a non-issue unless you need every single last drop of performance... in which case you probably won't be going with a consumer MLC drive in the first place.

NOTE : someone did say " PRE RELASE " ....

ausmike wrote:We even got some 'pre release' intel SSD's that are 1TB and 1.5TB THEY LAST EVEN LESS at the moment !

I thought this was a 'friendly' forum ? whats with name calling folks?

I for one DO KNOW WHAT MIKE IS referring too- eventhough i know he is kinda handtired with 'product secrecy' BS!
but than .... sticking to topic here
SSD = do wear out FASTER than normal PLATTER DRIVE - lots been written on this topic e verywhere and (read writes) speeds does go down on HEAVY PRELONGED USEAGE ie - lot lof loss when compared to "normal" drive!
sorry for my engrish; but then maybe pple dont some notice - [i]or maybe I dont know much nor does mike - but we SURE CANNOT WRITE as elequilently as some others - maybe lack of work in CH NC ? [/i]

StevenC = thanks - you dont need to defend me
but I am able to deal with 'people' who SEEM to know it all..
Just got back > had a nice flight > thanks mate for gread time in S'zen.

Cheers everyone

No offense intended mate. I certainly don't know what sort of hardware you can get your hands on. It's just that I've come to be a little skeptical over the years of large claims regarding what hardware one does or does not own. (I still remember a discussion around 2001 or so in which a friend swore to me that he had access to the PlayStation 3, featuring a whopping 128MB of RAM...) You must admit, it is pretty incredible to think that Intel's SSD line will soon jump from 160GB to 1.5 TB (150% of the capacity of the largest consumer SSD on the market.)

But leaving all that aside:

How about we make the discussion more quantitative? There certainly were some issues with performance degradation with first-gen X25-M drives running early versions of the firmware -- but those problems were fixed. Furthermore, any performance loss that is observed is not permanent. If this is indeed still a major issue I'd love to see a write-up on it, as I use a number of SSDs in both home and work settings. Do you have any links you could share?

I suggest that anyone who's interested in SSDs check out Anandtech's articles on the matter, as they not only dispel many uncertainties about the various technologies but do so using a number of well-designed benchmarks. (They've also got a great explanation of how SSDs work, specifically how various page/block mapping schemes can lead to [or avoid] various problems.)

I'd like to note with the smaller litho processes, you will start seeing very high capacity SSDs with very low endurance (~3000 writes MLC)

The good news? Well, besides quantum tunneling killing your data, hey, you have 22-25nm NAND that lets you either have huge capacity or longer lifespan through a massive reserved zone.

As for the capacity? OCZ Colossus being the 1TB drive which is actually several subdrives all bashed into one through controller magic (Controllers for the controllers? Enjoy your overhead)...

But hey, all guessing here, Intel might actually have a 1TB+ prototype that doesn't have a reserve zone set that happens to be in ausmike's hand.

lol NOV* ...you cheecky mate !

However - maybe this thread might throw some 'light' to my continued frustrations with TEAM-SSD and their related Project H'ware/Sware etc issues: blah blah ,,,,
http://forum.thinkpads.com/viewtopic.php?f=58&t=90342

edit > corrected links

Navck wrote:I'd like to note with the smaller litho processes, you will start seeing very high capacity SSDs with very low endurance (~3000 writes MLC)

It is indeed going to be interesting to watch what happens as NAND makers continue to shrink their process size. I know that Sandforce's approach is very much geared towards coping with low-grade NAND. Of course based on my sole experience with their SF-1200 controller, I wouldn't really say they cope all that well. Anandtech also managed to kill a couple of SF drives as well, which really makes me question their ability to work around NAND issues in the controller...

And yeah, I'm not really a fan of the multi-controller approach that OCZ has taken, but I guess if space is their priority it kinda makes sense.

Sure, if you want space that has compromised wear leveling algos. I wouldn't touch OCZ's offerings at all, especially with all the horrible items I've heard about from an engineer friend.

Oh right, you know as they shrink the process size, you're going to see extremely awesome things happen, like your electrons deciding to appear in another cell... Or your insulator and semiconductor material deciding to trade places...

This is why I keep telling people that when they buy a SSD, they're paying for a beta product's R&D, they don't even have models predicting data retention yet. (Here, put the NAND media in an oven and see what you can read after 1 hour of that. Welcome to SSD R&D "best case.")

Navck wrote: This is why I keep telling people that when they buy a SSD, they're paying for a beta product's R&D.

I don't agree with that. Or rather I do, but I think that almost all of the computing field is like that: the industry changes so rapidly that products are evolving constantly. I don't think that SSD tech. is any different in this regard. Just as with CPUs, WLAN, and even LCDs, what is today "cutting edge" will invariably be laughably overpriced, buggy, and unreliable when compared with current generations of the same technology a few years from now.

Are SSDs less reliable, slower, and pricier than they will be in a few years? Absolutely. But that's the case for pretty much any component.

Thinkrob, I'll wait for them to introduce SSDs with racetrack memory. Infinite rewrites, fast, radiation hardened. Runs hot too.

The problem with SSDs right now is NAND will wear and the smaller the process size, the faster you run into walls of problems that will take a whole lot of funding and time to solve.