FunTuple: prepare call for every candidate
When looking at Analysis!871 (merged), I noticed that the function get_thor_output looks like:
auto get_thor_output( const LHCb::Particle& cand, const unsigned int& nfield, const unsigned int& nfunctor ) const {
auto prep = m_fun_thor[nfield][nfunctor].prepare();
return m_fun_thor[nfield][nfunctor]( cand );
}
My understanding is that prepare()
only needs to be called once per event. I wonder if this has significant impact on the speed of FunTuple.