Evaluate replacing ECAL CalibratedClusterPtr with CalibratedPFCluster

[fwyzard]

The PFECALSuperClusterAlgo class uses internally a CalibratedPFCluster class which is a wrapper around an edm::Ptr<reco::PFCluster>:

  class CalibratedPFCluster {
  public:
    CalibratedPFCluster(const edm::Ptr<reco::PFCluster>& p) : cluptr(p) {}

    double energy() const { return cluptr->correctedEnergy(); }
    double energy_nocalib() const { return cluptr->energy(); }
    double eta() const { return cluptr->positionREP().eta(); }
    double phi() const { return cluptr->positionREP().phi(); }

    edm::Ptr<reco::PFCluster> the_ptr() const { return cluptr; }

  private:
    edm::Ptr<reco::PFCluster> cluptr;
  };

As well as a shared_ptr<...> and a vector<shared_ptr<...>> types:

  typedef std::shared_ptr<CalibratedPFCluster> CalibratedClusterPtr;
  typedef std::vector<CalibratedClusterPtr> CalibratedClusterPtrVector;

The use of what are essentially a shared_ptr<Ptr<T>> and vector<shared_ptr<Ptr<T>>> seems inefficient, given that edm::Ptr<T> is a lightweight object which should be used by value.

In addition, the use of std::vector<std::shared_ptr<CalibratedPFCluster>> could be replaced by the much more efficient edm::PtrVector<T>.

The use of these non optimal data structures is exacerbated by #37115, which introduces and uses

     typedef std::vector<std::pair<CalibratedClusterPtr, CalibratedClusterPtrVector>> EcalGraphOutput;

vs the more compact

     typedef std::vector<std::pair<edm::Ptr<reco::PFCluster>, edm::PtrVector<reco::PFCluster>>> EcalGraphOutput;

I would suggest to study the impact of replacing CalibratedClusterPtr with just CalibratedPFCluster, and of reworking the interface to support using edm::PtrVector<reco::PFCluster>.

[fwyzard]

assign reconstruction

[cmsbuild]

New categories assigned: reconstruction

@jpata,@slava77,@clacaputo you have been requested to review this Pull request/Issue and eventually sign? Thanks

[cmsbuild]

A new Issue was created by @fwyzard Andrea Bocci.

@Dr15Jones, @perrotta, @dpiparo, @makortel, @smuzaffar, @qliphy can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

[fwyzard]

type performance-improvements

[fwyzard]

assign egamma-pog

[cmsbuild]

New categories assigned: egamma-pog

@lfinco,@swagata87 you have been requested to review this Pull request/Issue and eventually sign? Thanks

[fwyzard]

type egamma

[swagata87]

Thanks a lot Andrea, for the suggestions.
Together with Davide, we will try to address this soon.