3 ms·
That sounds like NAND degradation (retention failures) which can only be partially worked around in firmware (and causing more write cycles on already-marginal
by userbinator 2y ago
That sounds like NAND degradation (retention failures) which can only be partially worked around in firmware (and causing more write cycles on already-marginal QLC). Unfortunately the real solution is "use better NAND", which is unlikely to happen unless enough people demand it.
- ciupicri 2y agoKingston KC3000 supposedly uses Micron 176L TLC memory [1]. The Seagate Firecuda 530 datasheet clearly says "Built with a Seagate-validated E18 controller and the latest 3D TLC NAND". A review is more precise: "Phison PS5018-E18" & "Micron B47R 176-layer 3D TLC NAND" [2]. [1]: https://www.tomshardware.com/reviews/kingston-kc3000-m2-ssd-review https://www.tomshardware.com/reviews/kingston-kc3000-m2-ssd-... [2]: https://www.kitguru.net/components/ssd-drives/simon-crisp/seagate-firecuda-530-2tb-ssd-review/all/1/ https://www.kitguru.net/components/ssd-drives/simon-crisp/se...
- userbinator 2y agoB47R is indeed TLC, rated for only 1000 cycles (and 35k in SLC mode, at 1/3 the capacity.) There's also the question of whether this is "true" Micron NAND, or SpecTek which is basically Micron's rejects (and rated for even fewer cycles; only 300 in the case of their B16A.)