To expand a bit on Sid's comment. "Checkpoint" is when an image of the work in progress is written to disk and can be used as a restart point. Most tasks checkpoint every 10 or 20 minutes, but some take an hour or two. All tasks checkpoint at the end of a model, but some checkpoint at stages within a model as well.

Any time your reboot the machine, the task will restart from the last checkpoint (if any, otherwise from the beginning). The leave tasks in memory setting pertains to when a task stops running for some reason other than a reboot. Perhaps BOINC Manager decided to suspend the R@h task and run a task from another project. If you leave the task in memory, the work done since the last checkpoint is retained in the task memory. If you do not leave the task in memory, the work since the last checkpoint is lost because the system task is completely ended. As with any system task, if it is not active, the "memory" used will be swapped to disk by the operating system, if there are other demands for actual memory. So the setting is really talking about virtual memory, not necessarily physical memory.

---

Rosetta Moderator: Mod.Sense