I've written a program for a class and after we've written it we're suppose to do some data analysis on it. A couple of the several things we're suppose to report given different conditions are the number of page faults and the number of page ins/page outs. Are there times when a page fault occurs that that you don't have a page in? What all is the difference? Thanks!

Hi -

Q: What is the definition of a "page out"? (that's easy )

Q: What is the definition of a "page in"?
Under what circumstances might it occur?

Q: Let's say you write a program that might allocate 2GB.
Do you believe the entire 2GB is allocated all at once, the moment the program starts? If not, why not?

Q: Back to the question of "page in": what might happen if I have a memory-mapped file (and plenty of RAM, and 0 swap being used)?

Q: Another question about "page in": where does a program's static data and and constants come from? Do you think they're allocated all at once, up front?

Q: What about shared libraries a program might use?

Thoughts for your consideration

Here's a great link (one I'm sure you've already encountered):
http://tldp.org/LDP/tlk/mm/memory.html

'Hope that helps .. PSM

Yes - Linux distinguishes major and minor faults. The latter do not require a page-in.

As for the difference, I presume that is what your teacher wants you to investigate. You now have some terminology to google.