Min heap with std::priority_queue is my bottleneck

Question

Alternative title:

Implement min heap with something faster than std::priority_queue.

---

grpof gave me:

time seconds seconds calls s/call s/call name
84.12 105.54 105.54 320000 0.00 0.00 _ZN3RKDI24Division_Euclidean_spaceIfEE2nnEjRKSt6vectorIfSaIfEERKfRS3_ISt4pairIfiESaISB_EERiiPjRSt14priority_queueISt5tupleIJfiiEES3_ISJ_SaISJ_EESt7greaterISJ_EES9_RKjS7_S7_i

which I believe is my only std::priority_queue used in the project. The 'Division_Euclidean_space' part confuses me, since it's a class/file not used any more in my project.

Here is what I use exactly:

/**
 * Min_heap is actually a std::priority_queue,
 * with std::greater as a parameter.
 */
typedef std::priority_queue<std::tuple<float, int, int>,
    std::vector<std::tuple<float, int, int> >,
    std::greater<std::tuple<float, int, int> > > Min_heap;

I use the first element as the key for comparison.

and as one can see in my repo, I create only one Min_heap and I use it in two parts:

if(...) {
  branch.push(std::make_tuple(new_dist, other_child_i, tree_i));
}

and

while (branch.size()) {
  std::tie(new_mindist, node_i, tree_i) = branch.top();
  branch.pop();
  ...
}

I feel that if I replace this data structure with something else, my project may run a bit faster (super duper great). Any ideas?

I push items in the heap for a while, then I pop one and I will probably push other items do and so on. Most of the times I stop with another condition, not when the heap gets empty.

Answer 1

http://demangler.com translated that function into: (indented by me)

RKD<Division_Euclidean_space<float> >::nn(
  unsigned int,
  std::vector<float, std::allocator<float> > const&,
  float const&,
  std::vector<std::pair<float, int>, std::allocator<std::pair<float, int> > >&,
  int&,
  int,
  unsigned int*,
  std::priority_queue<std::tuple<float, int, int>,
                      std::vector<std::tuple<float, int, int>,
                                  std::allocator<std::tuple<float, int, int> > >,
                      std::greater<std::tuple<float, int, int> > >&,
  float const&,
  unsigned int const&,
  std::vector<float, std::allocator<float> > const&,
  std::vector<float, std::allocator<float> > const&,
  int)

I don't know what nn does, but I suppose it does a lot more than priority queue operations.