Added dRdX to residual #201
Conversation
brownd1978
left a comment
brownd1978
left a comment
There was a problem hiding this comment.
Given the added residual is restricted to a single use case, I wonder if the added functionality should reside in a subclass of Residual, say AlignmentResidual. It appears that could be constructed by the appropriate hit from the existing residual and the closest approach. I'm not sure if the downstream calling sequence supports that.
Detector/Residual.hh
Outdated
| double pvar_; // estimated variance due to parameter uncertainty | ||
| bool active_; // whether this residual is active or not | ||
| DVEC dRdP_; // derivative of this residual WRT the trajectory parameters, evaluated at the reference parameters | ||
| VEC3 dRdX_; // derivative of this residual WRT the trajectory POCA, evaluated at the reference parameters |
There was a problem hiding this comment.
Is there a reason this is returned VEC3 and not SVEC3? SVEC3 supports algebraic manipulation while VEC3 supports geometric manipulation, even if the payload is identical. Nominally a derivative is an algebraic concept.
Detector/ResidualHit.hh
Outdated
| double pvar = ROOT::Math::Similarity(resid.dRdP(),params.covariance()); | ||
| if(pvar<0) throw std::runtime_error("Covariance projection inconsistency"); | ||
| return Residual(uresid,resid.variance(),pvar,resid.active(),resid.dRdP()); | ||
| return Residual(uresid,resid.variance(),pvar,resid.active(),resid.dRdP(),resid.dRdX()); |
There was a problem hiding this comment.
Do you have an estimate of the increased to processing time from adding this derivative?
2 participants
Store derivatives wrt to position in residuals for use in alignment analyses