Skip to content

Very high failure rate of OWNPVLTIME functor for V0

If this is a duplication, please follow up with a link, such that we can close this one.

I got notified that there may be a serious issue with an very large failure rate (NaN) of the OWNPVLTIME functor for (LL and DD) V0:

https://gitlab.cern.ch/-/project/465/uploads/703b83bcfbc71da5955b6e9e1a2d4ad1/SMaccolini_210325_KShLLDDHlt2cuts.pdf

The LTIME functor can only legitimately return NAN if there is no PV. We need to figure out if this is

  • a problem with the PV assignment, or
  • a problem with the LTIME functor, or
  • a problem with the V0 vertexing

If it is a problem with the LTIME functor (it is possible that it diverged for very large IPCHI2), then a possible fix is to return the value of the 'initialization' ( which is a simple projection that cannot fail). A divergence failure is easy to spot from the algorithm itself. AFAIK the C++ code is here:

https://gitlab.cern.ch/lhcb/Rec/-/blob/master/Phys/SelTools/include/SelTools/DistanceCalculator.h?ref_type=heads#L195

I am on vacation, so I cannot do much, but it would be nice if we can fix this before the June 19 deadline. I do not read e-mail until July 7th, but I do read mattermost.

If there is more information, then it would also be nice to follow up here.

@mmulder, @msaur, @jadevrie, @mveghel, @dovombru, @mstahl, @tevans, @wmorren, @smaccoli

Edited by Mick Mulder