> Regarding your experience, I can't make an objective analysis however from the fact that you're running Linux and that you wrote to disk more than the nominal

> Regarding your experience, I can't make an objective analysis however from the fact that you're running Linux and that you wrote to disk more than the nominal lifespan, I'm guessing that you probably were doing some IO-intensive things that normal consumers wouldn't do. 1. Low-end SSDs starts at 40 TBW 2. Overwrites are counted for internal writes, not OS-level writes. In my case SMART reported 2-3x higher number of internal writes than OS-level writes (probably due to strained wear-leveling algorithm when the SSD was used mostly full). 3. I would consider my case as low-intensity workload (basic desktop usage), but 10 GB/day seems like too low. Just watching a movie is 2 GB write, if one download it first instead of watching stream. But let say it is 10 GB/day. 40000/(10×2.5×365) = 4.38 years, I think my SSD was 6 years old. But all this is irrelevant, as the original argument about swap on SSD is not about regular usage of SSD, it is about writes triggered by swapping, esp. thrashing during near-OOM situation.