问题
There is a topic already on this topic but I have doubts still. To calculate the size of a vector, which one is correct:
sizeof(VEC) + sizeof(int) * VEC.capacity()
or
VEC.capacity() * (sizeof(VEC) + sizeof(int))
回答1:
What do you mean by size of the vector? The size of the vector object is just
sizeof(vec);
If you are interested in how much memory the vector has allocated on the heap, you can use
vec.capacity()*sizeof(T)
So, if you add these, you'll get how much memory you've "lost" because of the vector.
vec.capacity()*sizeof(T) + sizeof(vec)
Please note that exactly how much memory is allocated is implementation-dependent. It's just that the formula above will be practically correct (or approximately correct) on most if not all implementations.
回答2:
If you want to know the size of the data contained within the vector
std::vector<int> vec;
...
vec.size() * sizeof(decltype(bufferIn)::value_type))
If the type is changed from int to say long long nothing else has to be changed. I find this solution safer then the one using the type in the sizeof because it one can easily be changed without the other.
std::vector<long long> vec; // lets go from int to int64
...
vec.size() * sizeof(int); // woops problem on the horizon
来源：https://stackoverflow.com/questions/8778091/sizeof-stdvector-c