>Which can be caused by having other processes loaded. I'm not sure >what you mean by "paging to the pagefile begins". Are you referring >to the swap file managed by Windows, or are you referring to something >special managed by P95? If the latter, then the circumstance I describe >should not cause that.
Paging is the process by which XP working sets are managed; the virtual memory subsystem generates pagefaults when a page needed by a process is not in memory, and the requisite page is returned from the pagefile, which in the case of XP is called pagefile.sys. Swapping is a term originally used to describe a process being completely swapped to disk, as opposed to have just some pages paged to the pagefile. > priority processes have been paged-out, performance is always going to > be terrible. The solution is of course more physical memory. > If I understand you correctly, not so. If the higher priority processes >are in virtual, but are not being scheduled, then performance will be >normal until one of the idle processes starts getting scheduled. Then >there will be a delay while enough of it gets loaded for it to respond >to whatever event woke it up. If you are in a situation where high priority processes are scheduled that don't have enough pages in memory to be able to execute, then pagefaults occur, which is OK within limits. Things get out of hand when the system is thrashing pages in and out of the pagefile in order to schedule different processes... This thread is getting a bit off-topic, shall we stop?
_______________________________________________ Prime mailing list [email protected] http://hogranch.com/mailman/listinfo/prime
