Hacker News new | past | comments | ask | show | jobs | submit login

Couldn't that result in much slower reads when the head is far away from the swap when it needs to read (multiple times)?



Even more interesting idea: pages are oportunistically mirrored between swap partitions and the kernel smartly chooses the closest one whenever needed!


Does the kernel even have the information to know which is closest. I figured that would be absracted away to the disk controller.


The slowest swap was at the end probably because it was farther. The position of the head can be inferred by the geometry and last access.




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: