Discussion Forums > Technology

New Workstation Rig

<< < (4/4)

kureshii:

--- Quote from: TMRNetShark on September 17, 2011, 01:56:34 AM ---ANYWAYS, Life Endurance on an SSD today is meaningless. Why?

Cause SSDs today have a write endurance of well over 5 million write cycles... That means if you constantly "writing" on your SSD, you would never be able to write enough on your SSD for it to fail in your computer's natural life. (or so they say... continue reading)

Why?

Your harddrive doesn't write constantly, but it writes a lot, right? In an SSD, when you (re)write even 1 bit, you are essentially rewriting an entire block, right? So a 256 KB block is changed all to just change 1 bit... a waste it would seem, right? You just took up one of those "endurance" cycles away because of that. Let's look at your average SSD that has a endurance "life cycle" of 3 million. That's 3 million times that your SSD can write before blocks start going boom. What if you were writing at the constant speed that an SSD can write at until it dies? That's 3 million times the capacity of the harddrive (so let's say 64 GBs) and let's say that 70 % of those blocks last until the "3 million life cycle endurance". At the constant writing rate of 165 MB/s, that 70 % of those blocks will last 26 years. That's if 1 bit was being changed or the entire harddrive was constantly being written on day in and day out for 26 years.

"Yeah, TMR? Then why are there so many examples of people's SSD's failing less than a year after they were bought?"

Well, to answer your question... It's either that SSD companies are lying to us about having write cycle endurance limits in the millions (where we thought the technology was at) and the reality is that the write cycle limits are only as far as 100,000 write cycle endurance (which puts it at a theoretical ~314 days limit before it goes bust under the same conditions as the 26+ years example). OR "that if a failure ever does occur, it will not occur in the flash chip itself but in the controller."

So theoretically, with the numbers we are given... SSD's SHOULD last 20+ years. What we are actually seeing is quite different though... and I doubt everyone's "SSD's controller" is failing because they used it wrong.  ::)

Would I buy an SSD? I'm not convinced one bit by the technology.

--- End quote ---
… I hope this is more of your so-called “sarcasm”, because if I facedesk any harder I’m not going to have a face any longer.

TMRNetShark:

--- Quote from: kureshii on September 17, 2011, 03:05:03 AM --- (click to show/hide)
--- Quote from: TMRNetShark on September 17, 2011, 01:56:34 AM ---ANYWAYS, Life Endurance on an SSD today is meaningless. Why?

Cause SSDs today have a write endurance of well over 5 million write cycles... That means if you constantly "writing" on your SSD, you would never be able to write enough on your SSD for it to fail in your computer's natural life. (or so they say... continue reading)

Why?

Your harddrive doesn't write constantly, but it writes a lot, right? In an SSD, when you (re)write even 1 bit, you are essentially rewriting an entire block, right? So a 256 KB block is changed all to just change 1 bit... a waste it would seem, right? You just took up one of those "endurance" cycles away because of that. Let's look at your average SSD that has a endurance "life cycle" of 3 million. That's 3 million times that your SSD can write before blocks start going boom. What if you were writing at the constant speed that an SSD can write at until it dies? That's 3 million times the capacity of the harddrive (so let's say 64 GBs) and let's say that 70 % of those blocks last until the "3 million life cycle endurance". At the constant writing rate of 165 MB/s, that 70 % of those blocks will last 26 years. That's if 1 bit was being changed or the entire harddrive was constantly being written on day in and day out for 26 years.

"Yeah, TMR? Then why are there so many examples of people's SSD's failing less than a year after they were bought?"

Well, to answer your question... It's either that SSD companies are lying to us about having write cycle endurance limits in the millions (where we thought the technology was at) and the reality is that the write cycle limits are only as far as 100,000 write cycle endurance (which puts it at a theoretical ~314 days limit before it goes bust under the same conditions as the 26+ years example). OR "that if a failure ever does occur, it will not occur in the flash chip itself but in the controller."

So theoretically, with the numbers we are given... SSD's SHOULD last 20+ years. What we are actually seeing is quite different though... and I doubt everyone's "SSD's controller" is failing because they used it wrong.  ::)

Would I buy an SSD? I'm not convinced one bit by the technology.

--- End quote ---
… I hope this is more of your so-called “sarcasm”, because if I facedesk any harder I’m not going to have a face any longer.

--- End quote ---

About 95% of it was sarcasm... The link is the only non-sarcastic remark (seeing as how it really IS a debate).

kitamesume:
^ you first need to confirm if that "millions of write cycles" is per NAND-Flash Chip or the whole SSD. if its the whole then its too little.

Edit: still though, even if its the whole's limit, it`ll last 20 years.

formula:
per disk: ([Write Endurance] x [Capacity]) / [Write Speed] = wear out in seconds of continuous use.
per flash: (([Write Endurance] x [Capacity]) x [NAND-Flash #]) / [Write Speed] = wear out in seconds of continuous use.

example per disk: ([5,000,000 Writes] x [61440MegaByte(60GB)]) / [500MB/s (Sata3 has it good)] = 614,400,000 seconds of continuous use or 19.48 years.

PS: continuous use i meant running the SSD @ it's rated speed for 24/7.

NaRu:
The thing that fails in SSD is the controller, not the write endurance. I had a SSD fail on me twice because the controller failed and Bios couldnt see it anymore.

Navigation

[0] Message Index

[*] Previous page