Clean up L1TTrackMatch producers [11_1_X] #30982
Conversation
|
A new Pull Request was created by @kpedro88 (Kevin Pedro) for CMSSW_11_1_X. It involves the following packages: L1Trigger/L1TTrackMatch @cmsbuild, @rekovic, @benkrikler, @kpedro88 can you please review it and eventually sign? Thanks. cms-bot commands are listed here
|
|
please test |
|
The tests are being triggered in jenkins.
|
|
+1 |
|
Comparison job queued. |
|
Comparison is ready Comparison Summary:
|
|
+upgrade |
|
@silviodonato if there will be another 11_1_X patch to include #30979, it would be good to include this one also |
| int match_trk_idx = matches.at(imatch); | ||
| if (match_trk_idx < 0) | ||
| continue; // this muon was not matched to any candidate | ||
|
|
||
| // take properties of the track | ||
| const L1TTTrackType& matchTk = (*l1tksH.product())[match_trk_idx]; | ||
| const auto& p3 = matchTk.momentum(); | ||
| const auto& tkv3 = matchTk.POCA(); | ||
| float p4e = sqrt(mu_mass * mu_mass + p3.mag2()); | ||
| math::XYZTLorentzVector l1tkp4(p3.x(), p3.y(), p3.z(), p4e); | ||
|
|
||
| edm::Ptr<L1TTTrackType> l1tkPtr(l1tksH, match_trk_idx); | ||
| edm::Ref<RegionalMuonCandBxCollection> l1muRef(muonH, imatch); | ||
|
|
||
| float trkisol = -999; | ||
| TkMuon l1tkmu(l1tkp4, l1muRef, l1tkPtr, trkisol); | ||
| l1tkmu.setTrackCurvature(matchTk.rInv()); | ||
| l1tkmu.setTrkzVtx((float)tkv3.z()); | ||
| l1tkmu.setMuonDetector(detector); | ||
| tkMuons.push_back(l1tkmu); | ||
| } | ||
| return; | ||
| } | ||
|
|
||
| void L1TkMuonProducer::build_tkMuons_from_idxs(TkMuonCollection& tkMuons, | ||
| const std::vector<int>& matches, | ||
| const edm::Handle<L1TTTrackCollectionType>& l1tksH, | ||
| int detector) const { | ||
| for (uint imatch = 0; imatch < matches.size(); ++imatch) { | ||
| int match_trk_idx = matches.at(imatch); |
There was a problem hiding this comment.
@kpedro88 could you comment about this deletion? Thanks
There was a problem hiding this comment.
There were two functions that only differed by one line. Since a default-constructed edm::Handle is invalid, I unified the functions by checking if the corresponding argument is valid (a few lines below this). This reduces code duplication.
|
merge |
backport of #30970
This will make TDR production more efficient on multiple threads by eliminating stalls from legacy producers.