Originally posted by Bassat Are you stating that this technology is only activated when FORMATTING; that DELETING does not do the same thing
While they do have wear leveling like SSDs do it isn't as competent or capable. There are 2 types of wear leveling one which actively manages sectors and moves things around as needed on demand and that is used in SSDs and in industrial flash cards. Doing this requires additional power and is slower and makes the assumption that the device will be powered on for an extended period. The other is more passive and is really only triggered during formats and with writes to unused sectors but for writes it is basically a if failed mark the sector as bad and try a new one. This is faster and requires less power but doesn't do as good of a job. The proper wear leveling happens during a format when logical and physical sectors are remapped. Also since the file allocation table remains on the same sectors any write to it overwrites the same sectors until it is formatted and things can be remapped. I suspect that what causes failures for a lot of cards that aren't formatted frequently is that the sectors for the FAT have been overwritten too many times so are just worn out yet the other sectors have plenty of life left. This actually happens to a lot of people running those small system on a chip (SOC) computers like a raspberry Pi that are trying to use a SD card as a hard drive.
Originally posted by Bassat Modern SD cards are generally accepted to be reliable for 100,000 R/W cycles.
This assumes that full wear leveling has taken place. Individual cells and physical sectors have a substantially lower number of cycles usually only a few thousand and as feature sizes have shrunk and the bits per cell have gone up this number has gone down but processes have improved so which it has gone down it hasn't gone down anywhere near as fast as capacity has increased. Add in under provisioning, spare capacity that is only visible to the controller, and we have a case where the reliablity has actually gone up. SSD actually have more spare capacity and the better ones more than others. This is why the cheaper SSDs from a manufacturer will have a slightly higher capacity but a shorter warranty. For example Samsung sells both 500GB and 512GB SSDs, they both use the exact same flash chips and the same number of them but the more expensive 500GB one has 12GB extra spare capacity that is accessible only by the controller. So if one wants to save some money on SSDs you can buy the bigger cheaper drive and actually under provision it some more your self by only have a file system that uses 500GB of the 512GB provided by the drive. You will however not gain the extra warranty offered by the more expensive drive but will experience a similar lifetime to the more expensive one.