I'm trying to move from using Java for CP to C++. As such I'm missing some knowledge, but usually I'm able to google it. Today I encountered something that doesn't make sense for me and wasn't able to find good answer for that. In this solution for latest Div 3 contest (problem D) I used unordered_map and received TLE. Then I changed it to map and got Accepted.
I know that what are the differences between map and unordered_map, but I would expect unordered_map to be even faster then map. Does hashing ints really creates so many conflicts, or I'm missing something (for example initialization of size of hash table)?