On Tue, Sep 2, 2008 at 5:34 AM, Steven Watanabe <watanabesj@gmail.com> wrote:
AMDG
Giovanni Piero Deretta wrote:
Hum, I've played a little with some mockup iterators, and I can definitely see the N^2 stack growth, and with some expression template heavy program, I wouldn't be surprised if the growth would actually become noticeable (even if it weren't enough to blow the stack, it could still produce cache misses and such). I can't seem to come-up with any clever scheme to reduce the growth (short of using pointers to the original range of course).
Enough for today. Maybe tomorrow I'll get some idea.
I've been playing around with this a bit. See attached.
I can't compile it because I lack a recent phoenix. What numbers are you getting? Also, isn't end() still O(N^2) on the stacking depth? -- gpd