Hi Niels,
indeed the explanation is exactly as Laurent says. The effect of the bug in the n-body topological trigger lines is basically dropping prompt events with more than one PV (IPCHI2<16), which becomes worse as a function of n. If we wouldn't have the OR in the topo lines, the effect would be indeed catastrophic, but also probably detected already during 2015. For 2018 data-taking, the functor will be removed from the trigger lines. We'd probably do the same for the Stripping if the users agree.
Cheers
On 10/04/18 17:21, Laurent Dufour wrote:
Hi Niels,
Indeed. 10% of the events are not saved.
The effect would be much worse if the IP-OR was not there. For a calibration line for example, I do not have this OR in there. My difference is approximately 250% (I’m sensitive to combinatorics), for example.
Cheers, Laurent
On 10 Apr 2018, at 17:10, Niels Tuning h71@nikhef.nl wrote:
Hi Carlos,
Thanks for the info!! What a nightmare...
I don't understand one thing from Olli's slides: he says that about 10% of the 3-body topo events are affected? What does he mean? That 10% of the events are not selected in the trigger?
But you told me that the OR with a displacement requirement probably saves many events? (AALLSAMEBPV(-1, -1, -1) | (AMINCHILD(MIPCHI2DV(PRIMARY)) > 16)
How can I reconcile these two statements?
Cheers & thanks! Niels
On Tue, 10 Apr 2018, Carlos Vázquez Sierra wrote:
Hi all,
as you probably know there is a bug in one of the LoKi functors (only affecting Run-2 data & MC), namely AALLSAMEBPV. Olli has presented today the topic at the Tuesday meeting:
https://indico.cern.ch/event/715191/contributions/2960112/attachments/163017...
In case you have worked in a Run-2 analysis, please check if you have used this functor explicitly in your selection. If yes, your results might be affected.
Cheers
-- Carlos Vázquez Sierra LHCb Collaboration NIKHEF - National Institute for Subatomic Physics https://phonebook.cern.ch/phonebook/#id=PE704221 + carlos.vazquez@cern.ch
Bfys-group mailing list Bfys-group@nikhef.nl https://mailman.nikhef.nl/mailman/listinfo/bfys-group
Bfys-group mailing list Bfys-group@nikhef.nl https://mailman.nikhef.nl/mailman/listinfo/bfys-group