Problems? See this thread at archive.org.
Excellent post. But to reply in the quote above, how many times do you flash your pc's bios? Maybe twice, three times in it's life? On average. Of course that's going to be reliable, but also look at it's usage. It's meant to store data on a single flash EEPROM chip, not used in an array as a hard disk replacement to be read/ written to over and over and over.
Yes flash memory has been around for quite a long time, but to use it in this manner is new. It's using old technology in a new way, and I'm still not convinced yet that it's going to last much longer than a standard HDD with constant, heavy usage. Flash memory was never originally designed for such a use, but rather for permanant data storage, only to be read. And with windows always caching files, even when idle, that worries me.
EDIT: Only time will really tell. Technology is always progressing/improving, so I don't doubt that it could become a very reliable alternative (if not now, later) to magnetic hard drives.
With my formatting habits, I'm sure I'll kill an SSD in no time.Unfortunate fact.
But see, even with their "advanced wear leveling" technology, no doubt you would eventually come across a data block that's been used far too many times, and it will simply wear out. As in one of my previous posts, I posed the question if that would be seen as irrecoverable bad sectors. If that is true, then eventually your available storage capacity will decrease over time.
-
-
I've had an additional thought. Let's do some cheap math.
Assuming (for argument's sake), you have a 50gb SSD. Now, assuming you do 50gb of write/erase per day, that would mean you would hit each data block ONCE per day.
Now assuming that flash memory can be written to up to 1000 times (or is it 100,000? I can't remember, and I'm too lazy to search for it now), that would mean that you'd reach those 1,000 writes in less than 3 years.
If it were 100,000, that of course would last much much longer.
The inherit problem with flash memory is, it was never intended to be written to so much. As passive pointed out, that it's been known for many years the decay rate of using flash memory in a write-heavy usage environment, that is wrong. Only in the last fews years, have flash memory sticks/cards become available, and those used on PDA's and smartphones the usage is much much less than writing to each block once per day.
Also, a problem arises with the windows pagefile. Windows, when it creates the pagefile when you bootup your computer, it locks the file, and locks the sectors on the hard disk that is used for the pagefile, and writes (fills in the data on the pagefile) as it needs to. The problem with that is, for as long as your computer is on, windows will utilize that and read/write to it as often as it wants. Unfortunately, even for an SSD with "advanced wear leveling", those particular blocks are locked and used over and over, until the next reboot and a new pagefile is created. That could potentially create a lot of wear on an SSD just in that instance.
Flash memory was NEVER intended to be written to so heavily, it was meant for permanent data storage, without the need to have a constant power source to keep the data intact (unlike volatile - normal RAM, where you take away the power and the memory cells are subsiquently cleared).
Those claims by Mtron that it could have an endurance of >140 years, is simply only on paper, but doesn't take into account for the real world usage of a hard drive. Especially those that like to game a lot, encode videos, basically use their pc as a media center, not just use a computer for 30 minutes a day checking email and playing solitaire. Ok that's an exaggeration, but you get the point.
-
I have a question on the IDE flavor of SSDs, as I notice that they have a lower average read/write time than the SATA equivalent. This one in particular http://www.newegg.com/Product/Product.aspx?Item=N82E16820208317
The specs say it supports up to UDMA Mode 4 interface speeds, which is ATA-5. Since conventional notebook IDE drives are ATA-6/7, why is this SSD tethered to an obsolete interface specification? Does this not hamper it's performance?
-
-
Hang 'em!!!!
I've seen that random write speed is attrocious. I wonder, however, if Mtron has overcome this issue. I thought I read somewhere that there is a design that avoids the problem. Mtron may be the one, but not sure. -
-
-
Extenuating circumstances and Im waiting for another drive. A comparison will come out shortly after.
Attached Files:
-
-
-
-
-
-
-
I just wanted to bump this puppy up! Flemenko, where are you????? We need that data
Thanks, Dave
-
I enjoyed it alot. Really interesting about the different speeds. I remember back in the DOS days trying a RAM Drive to speed things up. I have always wanted to speed things up
I figure I will get one of these in about 6 months. Would sure love an Mtron 64GB when it hits the market. Heck, I would use it for my damn desktop drive at that point)) I have a pretty darn busy desktop and it only has 58 GB on it and I could easily remove 20 GB and not notice
Rocketdisk is a very interesting company and Beverly Hills, no less! Thanks, Dave
-
The transcend is the only one available at the moment (i'm from Kuwait). Anything else will take a decade to ship (christmas/new year) and cost me a lot more. It's either this or a HDD
-
Thanks Snap and rep to you
-
An Introduction to the Solid State Drive
Discussion in 'Hardware Components and Aftermarket Upgrades' started by Les, Oct 5, 2007.
