(NoUT) Tracking performance in dd4hep builds

Thanks to a number of recent fixes, we now have a tracking + RICH reco sequence running in the dd4hep builds (via the RICH refractive index test in Panoptes).

This is good, but means we now need to focus on performance and not just getting it running.

Compare the logs for the dd4hep build

https://lhcb-nightlies.web.cern.ch/nightly/lhcb-dd4hep/1175/Panoptes/x86_64_v2-centos7-gcc11%2Bdd4hep-dbg/tests#Panoptes.rich_ref_index_calib_noUT

to the DetDesc one

https://lhcb-nightlies.web.cern.ch/nightly/lhcb-dd4hep/1175/Panoptes/x86_64_v2-centos7-gcc11-opt/tests#Panoptes.rich_ref_index_calib_noUT

I am not a tracking expert, but there are significant differences, and from the RICH output I can tell the dd4hep tracking is far from right.

Note the above tests are using the same data file as in the RICH unit test

https://lhcb-nightlies.web.cern.ch/nightly/lhcb-dd4hep/1175/Rec/x86_64_v2-centos7-gcc11%2Bdd4hep-dbg/tests#RichFutureRecSys.pmts-v3.reco-from-dst

So I am confident the RICH performance in the above tests is OK.

@bcouturi @decianm @cmarinbe @gligorov @sponce FYI Could you take a look and see what you can deduce from the two log files ?

b.t.w. I would also suggest we add to Moore some noUT+noCALO+noMuon tests. I know the intention (hope) is Muon and/or CALO will be ready in a few months, and that is fine, but we need these tests now not in a few months to commission the sequences....