Read more.This 96-layer V-NAND uses a Toggle DDR 4.0 interface running at 1.4Gbps.
Read more.This 96-layer V-NAND uses a Toggle DDR 4.0 interface running at 1.4Gbps.
SSD tech is confusing nowadays: V-NAND versus 3D X-point which rules?
HDDs only have price, they have no performance.
They also have wear resistance. If you have a write heavy application, that fills and overwrites your storage many times a day. (Think security cameras, or a DVR), then almost any type of SSD will wear out and fail within months. The magnetic domains in an HDDs will cope fine with getting re-written millions of times, but the cells in an SSD will stop working after about 10k writes.
It can be a lot less than that; the latest QLC NAND is only good for about 1K program / erase cycles. Obviously wear-leveling & over provisioning can extend the life of a drive for quite some time, but as data densities increase, NAND cell lifetime tends to decrease...
Is there any software out there that's free and can look at how much of the overprovisioned space has been used and therefore give you an idea of how long your drive has left? I have an old 80GB Intel SSD that I may want to consider removing from service (although there's nothing essential on it now and it's backed up).
Size? If your pockets are deep enough you can get pretty big SSDs, and in a tiny 2.5in form factor to boot.
And yet things like dashcams continuously write to a Micro-SD card which generally have pitiful endurance compared to SSD. Recording a lot of security cameras onto a single hard drive, yeah that requires spinning rust to work at reasonable cost (which again is price/performance).
Note also that the next HAMR generation of HDD will require a laser to write data, and semiconductor lasers have a life expectancy before they fail which doesn't look that much better than SSD write endurance.
It isn't like the spare sectors on a HDD mapping out failed sectors, it seems to be more down to providing plenty of slack space to make it easier for the flash controller to maintain wear levelling.
Standard SMART tools should tell you how much data has been written. smartctl on Linux says my Samsung drive is 98% good:
177 Wear_Leveling_Count 0x0013 098 098 000 Pre-fail Always - 25
Ah, I dunno why but I can't get my brain to switch from the old HDD sector to SSD pages. That makes sense. I was wondering if the SMART stuff in the BIOS would warn you when a failure is imminent.
Mine was an early consumer SSD and it's getting on a bit now (I'm sure well past the MTBF but by WD Raptor HDD is also....) so I think it's prudent to poke at it and see how decrepit it really is.
Actually don't Intel Ssd's simply go into read only mode ... and that is done regardless of true ability to go on, just as soon as the allowed 1000 drive writes are done, or service life is over.
Media wear out indicator .... (intel kiss of death)
*SOME* Intel SSDs go into read only mode when they wear out, other Intel SSDs, and most other brands just stop working and take your data with them. TheTech Report did a wear out until they are dead test a few years ago.
More recently there was this rant on the Debian-ARM mailing list, from a Linux dev who's options I trust.
Originally Posted by lkcl@lkcl.net
There are currently 1 users browsing this thread. (0 members and 1 guests)