It's interesting that system RAM uses the same arrangement, a single capacitor to store bits. The difference is that it's not floating as in flash memory, and because of that it must be refreshed many times a second. Flash also loses data over time due to charge leakage.
Another interesting thing is that flash uses error correction, a form of redundancy. So they can have bad bits, ones that lose data more quickly, etc. without data loss. Still QLC (4 bits per cell) is the worst.
Still QLC (4 bits per cell) is the worst.
True. They've sacrificed quality and speed for larger capacities.
With more error correction information they could make its reliability closer to TLC. Hard drives have the same kind of more and more fragile storage methods, with lots of error correction added. It's interesting how they can blur the line of reliability between different densities.
Yeah, but the bigger the density, the more redundancy you need to add to keep it reliable. This has a cost in R&D and production. Especially at the current scale.
(post is archived)