Reduce the rates of double-charm lines in B&Q wg
As mentioned in #410, the rates of double-charm lines in B&Q are found to be very high. The rate is reduced to about 10% followed the selection in Charm wg. 50k MB MC samples are used to test the rate of the previous and updated double-charm line.
updated double-charm line:
Hlt2BandQ_DoubleCharmOppositeSign_LineDecisionWithOutput #=50000 Sum=48 Eff=|(0.09600000 +- 0.0138498)%|
Hlt2BandQ_DoubleCharmSameSign_LineDecisionWithOutput #=50000 Sum=56 Eff=|(0.1120000 +- 0.0149582)%|
previous double-charm line:
Hlt2BandQ_DoubleCharmOppositeSign_LineDecisionWithOutput #=50000 Sum=533 Eff=|( 1.066000 +- 0.0459268)%|
Hlt2BandQ_DoubleCharmSameSign_LineDecisionWithOutput #=50000 Sum=569 Eff=|( 1.138000 +- 0.0474352)%|
Merge request reports
Activity
assigned to @hmu
added RTA label
requested review from @ngrieser
- Resolved by Sebastien Ponce
/ci-test
added ci-test-triggered label
- [2022-06-09 08:21] Validation started with lhcb-master-mr#4693
added hlt2-throughput-decreased label
requested review from @mengzhen
- Resolved by Hongjie Mu
- Resolved by Hongjie Mu
- Resolved by Sebastien Ponce
In the first version: https://gitlab.cern.ch/lhcb/Moore/-/blob/master/Hlt/Hlt2Conf/python/Hlt2Conf/lines/bandq/builders/c_to_hadrons.py#L234 , we had Xicplus involved in the charm builder, but it does not contribute to any descriptors in https://gitlab.cern.ch/lhcb/Moore/-/blob/master/Hlt/Hlt2Conf/python/Hlt2Conf/lines/bandq/builders/doublecharm.py#L56 .
I notice that hongjie remove the Xicplus in the Merger for all the charm states. @gcavalle do you remember if we intended to contain the Xicplus in one of or some of the double-charm combiners ?
mentioned in issue #444 (closed)
- Resolved by Sebastien Ponce
Is this now ready for test ?
- Resolved by Sebastien Ponce
- Resolved by Sebastien Ponce
- Resolved by Sebastien Ponce
- Resolved by Sebastien Ponce
/ci-test
- [2022-06-14 18:07] Validation started with lhcb-master-mr#4741
removed hlt2-throughput-decreased label
mentioned in commit ab4189f4
mentioned in issue #410
mentioned in merge request !1897 (merged)