[Freeciv-Dev] Re: Profile.

[Gaute B Strokkenes <gs234@xxxxxxxxx>]

On Sat, 13 Oct 2001, gs234@xxxxxxxxx wrote:

> Another observation is that the speed gained by inlining
> is_real_tile() in normaliz_map_pos() is only 76.73 - 62.54 = 14.19
> seconds, or 2.5% of the total.

In case it's not clear, if you also count the 84% of the
is_real_tile() calls you're saving then the gain is actually 14.19 +
0.84 * 39.04 = 46.98 seconds, which is a much greater percentage.  But
my point still holds.  (Note that this does not mean that my patch is
only a 6 seconds speedup compared to Raimar's approach; you need to
take into account the pessimisation of is_real_tile() as well as the
optimisation of normalize_map_pos(), but that is much messier to do
since inlining is_real_tile() speeds up the caller as well.)

-- 
Big Gaute                               http://www.srcf.ucam.org/~gs234/
He is the MELBA-BEING...  the ANGEL CAKE...
  XEROX him...  XEROX him --