in reply to Re^3: What could cause excessive page faults?
in thread What could cause excessive page faults?

I am actually surprised that on a large 64 bit machine running some kind of *nix, that there are any page faults at all.

Are we sure there's actually even a problem? A page fault is not the same as paging (to swap). You'll take page faults any time you use memory, the first time you touch a page after you allocate it (on most *nix systems anyway; terms could mean totally different things on Windows, but I don't think they do).

  • Comment on Re^4: What could cause excessive page faults?

Replies are listed 'Best First'.
Re^5: What could cause excessive page faults?
by ikegami (Pope) on Mar 11, 2010 at 15:42 UTC

    Are we sure there's actually even a problem

    Yes. The OP's third snippet should take about 0.21 + (0.21-0.06) = 0.56s, but it's taking 31.48s.

Re^5: What could cause excessive page faults?
by Marshall (Canon) on Mar 11, 2010 at 08:58 UTC
    I ran another benchmark (above) with 16M numbers. I think that it is very clear that memory paging is happening on my Win 32 bit XP Pro system.