Hep MC Proposal Presentation of the proposal for

Hep. MC Proposal Presentation of the proposal for the new version of Hep. MC Lynn Garren July 25, 2006

background CMS requested Math. Core Gen. Vector n Others want no change to Lorentz vector n Concern about dependencies n Compromise proposed last week n A couple more unrelated requests n

The Proposal Remove all dependencies n Replace Hep. Lorentz. Vector with minimal simple vector n Prefer struct of four doubles n Need a bit more n n Continue to use existing code? n n Template constructor, but not class http: //lcgapp. cern. ch/project/simu/Hep. MC/download/

Simple. Vector. h n class Four. Vector n n class Three. Vector n n 4 -momentum or position Position Independent n Four. Vector does not have a Three. Vector

Four. Vector constructors Four. Vector( double x, double y, double z, double t=0) n Four. Vector(double t) n Four. Vector() n template <class T > inline Four. Vector( const T& v) n § n Uses x(), y(), z(), t() Side effect: problem with Four. Vector(0) n Seen, for instance, with Gen. Particle(0, …)

Four. Vector methods n n n n px() py() pz() e() m 2() perp 2() mag() x() y() z() theta() phi() rho() = == != pseudo. Rapidity() eta() set(, , , ) set. X() set. Y() set. Z() set. T() set. Px() set. Py() set. Pz() set. E()

Three. Vector methods x() y() z() n phi() theta() r() n mag() perp 2() perp() n = == != n set(, , ) set. X() set. Y() set. Z() n set. Phi() set. Theta() n

Rationale Methods used by Hep. MC n Similar methods that seem likely to be used n Don’t want a full vector class n

Pdf. Info n CMS requests Pdf. Info class n Hep. MC 1. 28? n Gen. Event has pointer (undefined by default) n n double x 1; // fraction of beam momentum carried by first parton ("beam side") double x 2; // fraction of beam momentum carried by second parton ("target side") int kf 1; // flavour code of first parton int kf 2; // flavour code of second parton double Qscale. PDF; // Q-scale used in evaluation of PDF's (Ge. V) double pdf 1; // PDF (kf 1, x 1, Q) double pdf 2; // PDF (kf 2, x 2, Q) n Needs to be discussed by LCG simulation n n

Generated Mass n n Expecting request from Monte Carlo authors Store generated mass n n Additional double Certainly there in HEPEVT common block Precision problem Might perhaps store px, py, pz, m n n Code breaks… Doesn’t make sense to calculate mass to store it • Must generated mass directly from generator

Conclusion n n Without templated constructor, not backwards compatible With templated constructor, partially backwards compatible n n n Templated constructor overrides default conversion (int to double) Possible obscure errors Four. Vector is not a full physics vector implementation Pere promised something to help when reading Is this proposal acceptable?