How do boost::numeric::ublas::vector
and std::vector
compare in runtime efficiency?
Is it safe to assume that I can convert an entire program from using std::vector
to use boost::numeric::ublas::vector
just by writing:
#include <boost/numeric/ublas/vector.hpp>
using namespace boost::numeric::ublas;
instead of #include<vector>
? Can I just use boost vectors as if they were ST开发者_Python百科L vectors in all aspects?
Do functions from <algorithm>
work with boost vectors? Do they use the same iterators?
Do they work in C++0x? Do they work for range based loops?
These are completely orthogonal data types: the former represents the algebraic definition of 'vector' (a one-dimensional matrix), while the latter represents the computer science definition of 'vector' (a one-dimensional array).
They don't compare.
You should only use ublas::vector if you want to do linear algebra operations, such as matrix vector multiplication etc. They do not provide the same functionality nor the same interface as std::vector. In terms of run-time efficiency, there is nothing, that I know of, that beats std::vector.
精彩评论