Do SSDs wear out?

Samsung SSDs are second to Intel SSDs. It's probably fine but you should always, always make backups because anything can fail. If it's valuable data then follow the 3-2-1 rule.

>How many writes can an SSD handle?
More then you can possibly throw at it. Google around for a forum thread where people are doing tests on SSD just read/write continously for years and haven't killed an SSD yet.

>Everything wears out eventually.

Even CPUs? Honest question. Will a 10 years old CPU underperform, compared to when it was new?

dban

No.
This is a problem unique to flash memory, even RAM is immortal.

CPUs should last a damn long time. Put it this way..if you fire up a Commodore 64, an IBM XT with an 8086 CPU, an NES, or an Amiga, they will all run just fine in terms of their CPUs. There might be other failures but the CPU won't be one of them unless they were operating in extreme conditions. GPUs are another story.

Basically the problem is there is a part of every flip flop in flash that is meant to stay electrically isolated. Every time you write a small amount of electrons leak in, if there are too many electrons trapped in the gate it stops working. No other type of semi conductor IC has this type of gate as far as I know so nothing else should fail in this way.

Attached: Flash-Programming.png (460x382, 13K)

>Even CPUs? Honest question. Will a 10 years old CPU underperform, compared to when it was new?
Heat is what kills CPUs, and even then, it's usually an either/or thing - either it works fine, or it's totally dead.
If you fire up a 10+ year old system that's always kept it's core temps down, the CPU will perform exactly as well and as quickly as it did when new.
Which is to say, not all that great if it's over 10 years old...
But no, the CPU won't degrade with time.
Flash memory does, because it's physical state changes in order to store the data. You can only change it's state so many times before it stops being able to go back to the way it was, so you loose that cell.
However, it takes a while. SSDs run for a good, long time - though they'll run even longer if you try and minimize the number of writes to the drive (this is more of a problem with the early SSDs though - less so for the new ones today).
A great example - use an SSD for your game installations, NOT your OS.

Continued

Once you install the game(s), which obviously takes quite a few write operations, the game mostly reads it's files from that point forward. Sure, some little things may change like config files, statistics, caches, that sort of stuff, but the activity on the drive will overwhelmingly be read operations as the game loads it's data files as you play.
This is the perfect usage scenario, as reading from a flash memory cell doesn't degrade it at all.

Fill the drive with random 1s and 0s then wipe it
Now you can only recover meaningless data

When your typical OS deletes a file, it doesn't actually delete the file - it deletes the pointer, or reference to that file, and declares that the space that file formerly occupied is now available for use.
Unless and until some other process writes data to that exact same spot the deleted file can be recovered, to varying degrees.
In order to securely -wipe- a drive, which is what you want, you need to run something that will overwrite the space on that drive with junk or blank data. Look for a drive wiping tool or a data / file shredder. There's multiple programs floating around that do this.
Do your research first - some are better than others. I'm not gonna recommend one because then we get into what OS you're using and how secure you need the wipe to be and how much trouble you're willing to go through and a bunch of other variables.
The important thing is that you understand that you need something that will overwrite that space.

You have like 1980 more TBs to write. Worry not.

sudo yes fuckthepolice > /dev/sda

Will fill the entire disk with 0x6675636b746865706f 6c6963650d0a over and over again.

A decent consumer SSD will be rated for 300-500TB and last much longer realistically. It's literally too much to care about. Maybe it will be a concern when QLC or later is the standard.

That would be true on a HDD but on SSDs they're constantly taking switches out of their accessible range when they become error-y. There's no guarantee that the regions of SSD memory you have written to in the past will be overwritable in the future. You'd have to go into some serious controller level shit to actually guarantee it. Your best bet is a hammer or microwave or something.

Someone tested it. Samsung 840 Pro can take 2PB before exhausted.

This. The controller hides complexity by mapping OS references to phisical locations and writes important data to new cells constantly to prevent wear on specific blocks. It even has cache RAM and predictive stuff in the controller.

Hdd also has a few techs like this to prevent biterrors but nothing on the scale of ssd controllers

>750 EVO
>10TB written
>sites claim it's supposed to last for 70TB

Attached: 1527189764282.gif (326x326, 255K)

What about leaving your machine turned off and the SSD without power for an extended period of time? How damaging could that be?

I had a SSD die on me but i put it in the oven and surprise it came back from the dead

Actually, you don't need to do that. The worn out sectors simply become read only.

>racetrack memory
... magnetic tape?

3 backups
2 something
1 offsite

I forgot what the 2 is

Okay well, acktshtually

Transistors can wear out over time. Can't remember what it's called.

It's highly dependent on manufacturing errors which are essentially random. Some CPUs might live for centuries, others might only last a few decades or so. There are a ton of individual parts that can fail in any computer component.

But their lifetimes are so god damn long that it's not really a concern in any real world scenarios.

It will change to read only mode but that happens after much more writes depending on the model and NAND type.

Attached: 127 - CmHW59o.png (326x232, 16K)

Performance will not change, there are tons of videos on this. However, it may die eventually. In my experience CPU's usually last a really long time, I've never lost one to age, motherboards and RAM on the other hand... Occasionally happen but aren't too bad.

youtube.com/watch?v=44JqNJq-PC0

Ill share an anecdote.
An old coworker shared a story of when he wanted to see how fast he could burn out an SSD. So he booted his favorite linux distro and made the entire ssd into a swap partition. He then pulled up a bunch of really memory intensive programs. He watched it go for a bit before going to get himself some coffee. By the time he got back it had kernal paniced.

Short answer: yes

wot

didn't he run a surface test after? there are many reasons why the kernel panicked, we don't know if the ssd died

HDDs usually give me lots of warning before dying. SSDs fail suddenly and totally.

how long do SSD last on average thou?

HDD I think die after maybe 3years of constant use seeding etc and maybe 6 years of casual use.

I remember hearing SSD last like 5 years or some thing but I'm still using one from 2011 and it seems mostly fine some times done some odd stuff but I think that was mostly installing windows bugs.

Online

Just use SATA secure erase.
Takes ~5 seconds for 100Gb SSD.
Impossible to recover data.

Attached: Screenshot_20180703-201115.png (1920x1080, 590K)

every GPU that's ever died on me has died due to extream conditions thou. like running in hot cast in SLI in Australian summer or literally damaging it with conductive thermal compound or it being defective NVidia an ddieing 2-3 years later.

I do think if you keep a GPU cool it will last forever just like a CPU its just people don't reseat heatsyncs on old GPUS and overclock them more aggressively. Ive killed a core2duo that ran at 4.5ghz for 9years and pushing it to 5ghz in last fue hours of its life (still stable thou).

GPU last people just stop taking care of them when they are not relevant. while CPU that's old people still clean and put new thermal compound on so they appear to last forever.