Why does ssd have limited number of writes ?
>>53268819
So do HDD, it's just that the point of failure and the time to failure cannot be determined in HDDs.
Just because it can't be determined doesn't mean it won't happen.
>>53268838
But why what's the reason behind it ?
>>53268895
why do people have to die?
>>53269322
because the hourglass runs out eventually, anon...
>>53268819
Because it is new technology. Wait a few more years when they work better, are more polished and last longer.
>>53268838
Agreed. Depends on your use case, and SSDs are just better characterized than HDDs. I bet a workload where a drive gets parked often would easily kill an HDD before even a basic consumer SSD.
>>53268819
jews
>>53268838
>>53269322
>>53269572
>>53269621
>>53269663
>>53269937
You people all fucking suck at answering the question.
He asked WHY they die. You answered, in order:
>Do HDDs die too?
>Could you please respond to this with a question, and please also assume that SSDs and human organics are even remotely connected?
>Why do hourglasses need to be continuously turned upside down?
>Why is the lifetime of current SSDs low?
>Should I use HDDs or SSDs?
>Who controls the press?
Notice how none of these are “Why do SSDs have limited number of writes?”
>>53268819
Laws of conservation
>>53270022
Also, a few google results for that exact phrase pick up:
>It would be impossible to explain P/E cycles in detail without using technical jargon, but basically, flash cells wear out a little bit each time you write to them. In a way, this is similar to writing on a piece of paper using a pencil and an eraser. You can only erase the pigment so many times before the piece of paper becomes worn out or even torn and can't be written on anymore.
Still not really an answer, though.
>>53268819
Because of entropy. Every system will eventually get rekt.
>>53268819
Because every drive in existence has a limited number of writes.
>>53268819
http://spectrum.ieee.org/semiconductors/processors/transistor-aging
>>53270459
>why does it have a limited amount of writes
Your answer:
>because it does
>>53270598
His question:
>why does this specific thing have the same problem as all other things like it
>>53270598
ok, a gamma ray burst can hit the earth at any moment destroying the drive
>>53270628
You're just making shit up, now.
>>53270668
He asked why do SSDs have writes. All drives have a limited writes. An SSD is a drive, so that's why it has limited writes.
Fuck, how hard is it to understand.
>>53270679
Your answer is still essentially "because it does."
Not anything that actually answers the question posed, like >>53270521
>>53270718
So what is op asking then? Are SSDs supposed to be revolutionary and make entropy non existent?
>>53270736
With an HDD, they usually only run out of writes due to mechanical failure.
With SDDs having no mechanical parts, it's easy to see why someone might think they might have unlimited writes.
>>53270760
Oh, well that's because the drive doesn't move at all. So the 1s get caught in the circuitry and slowly clog it up, slowing it down and reducing speed. Eventually they block some paths completely, and once enough has been blocked the drive is unstable.
Pic related. The red blood cells are the 1s, and the artery is inside the circuit
>>53270736
Is "why" such a difficult concept to grasp that the question makes no sense to you?
ssd's use nand based flash chips. to flip the bits in nand you pulse large amounts of current though very small paths. Doing so stresses the paths and they will eventually fail.
>>53270898
>Doing so stresses the paths and they will eventually fail.
‘Stress’ to me implies mechanical tension being applied.
I doubt that's the case here. Why are they being stressed?
An electron doesn't give a shit if it's flowing through a circuit a billion times, it doesn't change. There must be some underlying reason.
>>53271016
This >>53270521 describes the problems in detail. It's meant for processors, but applies to ssds as well.
>>53271016
The current generates heat which causes mechanical stress in the form of thermal expansion which has chance to break the path during the shrinking/cooling phase. Any other questions?
>>53271016
entropic degradation affects literally everything in existence bro
>>53270886
Everything breaks, every falls apart. Some things fall apart quicker than others. Asking "Why does the thing break" is pointless and stupid because the answer is because entropy.
You should instead ask How it breaks. In that manner you can avoid the product, work to make the product last longer, improve it's design to make it last longer or think of something else.
>>53270736
>>53270760
My RAM doesn't give shit how many times I write, read and erase it. I can flip a bit a billion times per second for years on end and it will still be just as durable as all the other bits.
My CPU's cache and registers don't give a shit how many trillion times I rewrite them.
Why is NAND flash so fundamentally difference? They're all just electric circuits, no? What part of the NAND gate physically deteriorates?
>>53268819
>Why does ssd have limited number of writes
Everything on this world has a limited number of (random bullshit)
This isn't magic land where everything lasts forever
>>53271066
Does this mean NAND flash dies more quickly because write cycle is like turning it on and off again over and over again?
I know that doing this is the primary reason for failure of virtually every component in a PC: Stress due to temperature changes, especially for power cycles.
>>53271081
>My RAM doesn't give shit how many times I write, read and erase it. I can flip a bit a billion times per second for years on end and it will still be just as durable as all the other bits.
Than whats the point of ECC and buffer RAM dipshit
>>53271116
Do you even know what ECC is?
ECC isn't some magical technology to increase the lifetime of your RAM.
ECC is just an extra error-correcting code to prevent against random bit flips, which are a _routine_ part of RAM's operation.
It's not like bit flips get any more or less likely the more you “wear out” the hardware. They always happen at a roughly constant rate, due to stuff like cosmic rays or other random fluctuations that you can't really isolate from the environment.
ECC really has no relevance in this discussion at all.
>>53271081
16GB Flash RAM vs 500gb, 1TB+ NAND. Thats alot more management for the drive, more room for degradation. Packing that much more in a similar size form.
>>53271106
NAND flash bits can only cycle so many times. they're like tiny switches in silicon. you toggle it so many times and they break but you can look to see what position they're in (about) forever.
>>53271016
>An electron doesn't give a shit if it's flowing through a circuit a billion times, it doesn't change
your understanding of electrons needs work
>>53271147
Oh, also, every SSD has error-correcting codes as well - because bit read errors are part of the normal, expected operation of a flash cell and it has nothing to do with the cell degrading.
High-reliability server-grade SSDs will often have shittons of parity slapped on top of them to get the required figure for the total number of non-recoverable read errors per X bytes read.
>>53271279
Electrons are physically indistinguishable from other electrons, other than through their configuration.
Electrons don't “wear out” over time, anon.
