L1 emulator for LLP displaced jet trigger (L1 6:1 LUT, L2 jet algo), L1TNtuples addition of jet hwQual #36919

gk199 · 2022-02-09T15:31:43Z

PR description:

This PR implements the L1 emulator of the LLP displaced jet trigger algorithm. Calo L1 takes the 6 fine grain bits from the HCAL uHTR (PRs #34600 and #35599) and uses a 6:1 LUT to set a single bit of tower hwQual based on the logical OR of the timing and depth information. The Calo L2 jet algorithm flags a jet as delayed if there are >= 2 flagged towers within the 9x9 jet region. This is saved in the LSB of jet.hwQual, 1 if jet is flagged as containing an LLP, 0 otherwise. L1 and L2 firmware updates have been completed.

The uGT emulator follows, and is implemented in recent PR #36868. L1 Menu Tools is in the process of being updated to include the displaced jet bit, in PR #74.

The L1 seed JIRA ticket, including slides detailing the testing of the L1 emulator and links to relevant presentations, is here. Slides 2-10 focus on the L1 emulator implementation and validation tests.

The emulator implementation has been presented at L1 Trigger CMS Week and to the HCAL Trigger group.

Additionally, the L1TNtuples are updated to store the jet hwQual information (in L1AnalysisL1Upgrade.cc and L1AnalysisL1UpgradeDataFormat.h). Tested by myself and @elfontan.

A single comment in HcalFinegrainBit.cc is updated to reflect the correct assignment of the 6 fine grain bits from the uHTR.

PR validation:

Passed runTheMatrix.py tests. Tested in CMSSW_12_3_X_2022-02-08-1100.

…dition of jet hwQual, fix comment in HcalFinegrainBit.cc

cmsbuild · 2022-02-09T15:37:51Z

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-36919/28230

This PR adds an extra 40KB to repository

cmsbuild · 2022-02-09T15:38:16Z

A new Pull Request was created by @gk199 (Gillian Kopp) for master.

It involves the following packages:

EventFilter/L1TRawToDigi (l1)
L1Trigger/L1TCaloLayer1 (l1)
L1Trigger/L1TCalorimeter (l1)
L1Trigger/L1TNtuples (l1)
SimCalorimetry/HcalTrigPrimAlgos (l1)

@epalencia, @cmsbuild, @cecilecaillol, @rekovic can you please review it and eventually sign? Thanks.
@kreczko, @missirol, @rovere, @Martin-Grunewald, @bsunanda, @thomreis, @dinyar, @abdoulline this is something you requested to watch as well.
@perrotta, @dpiparo, @qliphy you are the release manager for this.

cms-bot commands are listed here

cecilecaillol · 2022-02-10T10:26:48Z

please test

cmsbuild · 2022-02-10T10:38:32Z

-1

Failed Tests: Build ClangBuild
Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-cc9948/22324/summary.html
COMMIT: b392d89
CMSSW: CMSSW_12_3_X_2022-02-09-2300/slc7_amd64_gcc10
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/36919/22324/install.sh to create a dev area with all the needed externals and cmssw changes.

Build

I found compilation warning when building: See details on the summary page.

Clang Build

I found compilation warning while trying to compile with clang. Command used:

USER_CUDA_FLAGS='--expt-relaxed-constexpr' USER_CXXFLAGS='-Wno-register -fsyntax-only' scram build -k -j 32 COMPILER='llvm compile'

See details on the summary page.

…t for LLP trigger logic)

cmsbuild · 2022-02-10T14:07:32Z

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-36919/28248

This PR adds an extra 40KB to repository

cmsbuild · 2022-02-10T14:07:55Z

Pull request #36919 was updated. @epalencia, @cmsbuild, @cecilecaillol, @rekovic can you please check and sign again.

cecilecaillol · 2022-02-14T08:10:19Z

please test

gk199 · 2022-02-16T15:54:35Z

L1T DQM histogram comparison comments:

Changes are seen in the jet QUAL bit plot (CenJetsQual), and are expected. The Stage2Layer2JetAlgorithmFirmwareImp1.cc sets the least significant bit of jet hwQual (to 1) if the jet is found to be delayed based on the HCAL feature bits. This bit is set infrequently, as seen in the central jet quality histogram.
In the DQM histograms, I also see the HCAL Feature Bit occupancy histograms at Layer 1. This is also expected since the feature bit is now being set based a 6:1 LUT from the 6 fine grain bits from the HCAL uHTR (update to Layer-1 logic, as previously the feature bit just relied on the first uHTR fine grain bit).

cmsbuild · 2022-02-16T19:10:21Z

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-cc9948/22463/summary.html
COMMIT: e757f7b
CMSSW: CMSSW_12_3_X_2022-02-16-1100/slc7_amd64_gcc10
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week0/cms-sw/cmssw/36919/22463/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

@slava77 comparisons for the following workflows were not done due to missing matrix map:

/data/cmsbld/jenkins/workspace/compare-root-files-short-matrix/data/PR-cc9948/138.4_PromptCollisions+RunMinimumBias2021+ALCARECOPROMPTR3+HARVESTDPROMPTR3
/data/cmsbld/jenkins/workspace/compare-root-files-short-matrix/data/PR-cc9948/138.5_ExpressCollisions+RunMinimumBias2021+TIER0EXPRUN3+ALCARECOEXPR3+HARVESTDEXPR3
/data/cmsbld/jenkins/workspace/compare-root-files-short-matrix/data/PR-cc9948/139.001_RunMinimumBias2021+RunMinimumBias2021+HLTDR3_2021+RECODR3_MinBiasOffline+HARVESTD2021MB

Summary:

No significant changes to the logs found
Reco comparison results: 0 differences found in the comparisons
DQMHistoTests: Total files compared: 49
DQMHistoTests: Total histograms compared: 3965143
DQMHistoTests: Total failures: 8
DQMHistoTests: Total nulls: 0
DQMHistoTests: Total successes: 3965113
DQMHistoTests: Total skipped: 22
DQMHistoTests: Total Missing objects: 0
DQMHistoSizes: Histogram memory added: 0.0 KiB( 48 files compared)
Checked 204 log files, 45 edm output root files, 49 DQM output files
TriggerResults: no differences found

perrotta · 2022-02-17T15:58:07Z

+1

dan131riley · 2022-02-21T14:51:13Z

We're getting ASAN heap-buffer-overflow errors, likely related to this PR, see #37012

gk199 · 2022-02-22T16:22:58Z

EventFilter/L1TRawToDigi/plugins/implementations_stage2/CaloLayer1Packer.cc

@@ -107,9 +107,14 @@ namespace l1t {

            HcalTrigTowerDetId id(cEta, cPhi);
            const auto tp = hcalTPGs->find(id);
+
+            int fg_bits = 0;


uint32_t fg_bits = 0

gk199 · 2022-02-22T16:23:05Z

EventFilter/L1TRawToDigi/plugins/implementations_stage2/CaloLayer1Packer.cc

+
+            int fg_bits = 0;
+            for (int index = 0; index < 6; index++)
+              fg_bits += tp->SOI_fineGrain(index) << index;


fg_bits |= tp->SOI_fineGrain(index) << index;

This loop needs to be inside this conditional:

if (tp != hcalTPGs->end()) {

gk199 · 2022-02-22T16:27:19Z

We're getting ASAN heap-buffer-overflow errors, likely related to this PR, see #37012

I added some suggestions above in CaloLayer1Packer.cc, please let me know if there are any other thoughts to resolve this (or input from someone from L1, @cecilecaillol?). Is there a way to test the ASAN build locally, with a command to run to reproduce the error?

makortel · 2022-03-01T21:16:21Z

Is there a way to test the ASAN build locally, with a command to run to reproduce the error?

Yes, you can check the latest ASAN build from the IB dashboard and run a failing workflow there, e.g.

cmsrel CMSSW_12_3_ASAN_X_2022-02-28-1100
cd CMSSW_12_3_ASAN_X_2022-02-28-1100/src
cmsenv
runTheMatrix.py -l 23234.0

cecilecaillol · 2022-03-14T07:56:44Z

@gk199 Could you check if your suggested fixes above remove the errors?

gk199 · 2022-03-14T14:39:41Z

@gk199 Could you check if your suggested fixes above remove the errors?

I tested my suggested fix, and still have the heap-buffer-overflow issue. I also tested a number of other modifications, such as doing the 6:1 logic in CaloLayer1Packer.cc prior to passing through setFB instead of doing the 6:1 logic in UCTCTP7RawData.h.

Any other suggestions or guidance will be much appreciated as I work to resolve this.

dan131riley · 2022-03-14T16:33:35Z

This loop:

cmssw/EventFilter/L1TRawToDigi/plugins/implementations_stage2/CaloLayer1Packer.cc

Lines 111 to 113 in 62d26a2

    
           int fg_bits = 0; 
        
           for (int index = 0; index < 6; index++) 
        
             fg_bits += tp->SOI_fineGrain(index) << index;

needs to be inside this conditional:

cmssw/EventFilter/L1TRawToDigi/plugins/implementations_stage2/CaloLayer1Packer.cc

Lines 115 to 118 in 62d26a2

    
           if (tp != hcalTPGs->end()) { 
        
             ctp7Data.setET(cType, negativeEta, iEta, iPhi, tp->SOI_compressedEt()); 
        
             ctp7Data.setFB(cType, negativeEta, iEta, iPhi, fg_bits); 
        
           }

As is, when id isn't found it's indexing off the end of the vector.

gk199 · 2022-03-14T20:41:22Z

Many thanks @dan131riley for the fix and the explanation! I confirm this resolves the step2 heap-buffer-overflow from CaloLayer1Packer::makeHCalTPGs when running runTheMatrix.py -l 39834.0 in CMSSW_12_4_ASAN_X_2022-03-14-1100.

However, I see an error in step3, which seems unrelated to any code in this PR, as it is part of HGVHistoProducerAlgo.

READ of size 4 at 0x60d008945b0c thread T0
    #0 0x7fd8a67a0b47 in HGVHistoProducerAlgo::tracksters_to_SimTracksters(HGVHistoProducerAlgoHistograms const&, int, std::vector<ticl::Trackster, std::allocator<ticl::Trackster> > const&, std::vector<reco::CaloCluster, std::allocator<reco::CaloCluster> > const&, std::vector<ticl::Trackster, std::allocator<ticl::Trackster> > const&, HGVHistoProducerAlgo::validationType, std::vector<ticl::Trackster, std::allocator<ticl::Trackster> > const&, std::map<unsigned int, std::vector<unsigned int, std::allocator<unsigned int> >, std::less<unsigned int>, std::allocator<std::pair<unsigned int const, std::vector<unsigned int, std::allocator<unsigned int> > > > > const&, std::vector<SimCluster, std::allocator<SimCluster> > const&, edm::ProductID const&, std::vector<CaloParticle, std::allocator<CaloParticle> > const&, std::vector<unsigned long, std::allocator<unsigned long> > const&, std::vector<unsigned long, std::allocator<unsigned long> > const&, std::unordered_map<DetId, HGCRecHit const*, std::hash<DetId>, std::equal_to<DetId>, std::allocator<std::pair<DetId const, HGCRecHit const*> > > const&, unsigned int) const (/cvmfs/cms-ib.cern.ch/week0/slc7_amd64_gcc11/cms/cmssw/CMSSW_12_4_ASAN_X_2022-03-14-1100/biglib/slc7_amd64_gcc11/pluginSimulation.so+0x17eeb47)
    #1 0x7fd8a67a713b in HGVHistoProducerAlgo::fill_trackster_histos(HGVHistoProducerAlgoHistograms const&, int, std::vector<ticl::Trackster, std::allocator<ticl::Trackster> > const&, std::vector<reco::CaloCluster, std::allocator<reco::CaloCluster> > const&, std::vector<ticl::Trackster, std::allocator<ticl::Trackster> > const&, std::vector<ticl::Trackster, std::allocator<ticl::Trackster> > const&, std::map<unsigned int, std::vector<unsigned int, std::allocator<unsigned int> >, std::less<unsigned int>, std::allocator<std::pair<unsigned int const, std::vector<unsigned int, std::allocator<unsigned int> > > > > const&, std::vector<SimCluster, std::allocator<SimCluster> > const&, edm::ProductID const&, std::vector<CaloParticle, std::allocator<CaloParticle> > const&, std::vector<unsigned long, std::allocator<unsigned long> > const&, std::vector<unsigned long, std::allocator<unsigned long> > const&, std::unordered_map<DetId, HGCRecHit const*, std::hash<DetId>, std::equal_to<DetId>, std::allocator<std::pair<DetId const, HGCRecHit const*> > > const&, unsigned int) const (/cvmfs/cms-ib.cern.ch/week0/slc7_amd64_gcc11/cms/cmssw/CMSSW_12_4_ASAN_X_2022-03-14-1100/biglib/slc7_amd64_gcc11/pluginSimulation.so+0x17f513b)

As this is an unrelated file to this L1 emulator PR, may I now open a new PR to fix the heap-buffer-overflow with the changes in CaloLayer1Packer::makeHCalTPGs?

dan131riley · 2022-03-15T12:22:55Z

@gk199 please make a new PR with the step2 fix. The fix is clearly correct and needed, the step3 error can be dealt with separately.

L1T ASAN heap-buffer-overflow fix resulting from PR #36919 (12_4_X)

L1T ASAN heap-buffer-overflow fix resulting from PR #36919

L1 emulator (Calo L1 6:1 LUT, L2 jet algo with hwQual), L1TNtuples ad…

b392d89

…dition of jet hwQual, fix comment in HcalFinegrainBit.cc

cmsbuild added this to the CMSSW_12_3_X milestone Feb 9, 2022

cmsbuild added code-checks-pending l1-pending orp-pending pending-signatures tests-pending labels Feb 9, 2022

cmsbuild added code-checks-approved and removed code-checks-pending labels Feb 9, 2022

elfontan mentioned this pull request Feb 9, 2022

Inclusion of the displaced jet bit cms-l1-dpg/L1MenuTools#74

Merged

cmsbuild added tests-started and removed tests-pending labels Feb 10, 2022

cmsbuild added tests-rejected and removed tests-started labels Feb 10, 2022

fixing unused variables, commenting on finegrain bit 5 and 6 (MIP, no…

3ef9bf0

…t for LLP trigger logic)

cmsbuild added code-checks-pending tests-pending and removed tests-rejected code-checks-approved labels Feb 10, 2022

gk199 closed this Feb 10, 2022

gk199 reopened this Feb 10, 2022

cmsbuild added code-checks-approved and removed code-checks-pending labels Feb 10, 2022

cmsbuild mentioned this pull request Feb 16, 2022

A few possible bug fixes in the L1T code #36985

Merged

cmsbuild added tests-approved and removed tests-started labels Feb 16, 2022

cmsbuild added orp-approved and removed orp-pending labels Feb 17, 2022

cmsbuild merged commit 988e293 into cms-sw:master Feb 17, 2022

cmsbuild mentioned this pull request Feb 18, 2022

Implementation of beam halo tagger algorithm for the endcap photons #36901

Merged

dan131riley mentioned this pull request Feb 21, 2022

heap-buffer-overflow in l1t::stage2::CaloLayer1Packer::makeHCalTPGs() #37012

Closed

gk199 commented Feb 22, 2022

View reviewed changes

This was referenced Mar 15, 2022

L1T ASAN heap-buffer-overflow fix resulting from PR #36919 #37253

Merged

L1T ASAN heap-buffer-overflow fix resulting from PR #36919 (12_4_X) #37257

Merged

cmsbuild added a commit that referenced this pull request Mar 18, 2022

Merge pull request #37257 from gk199/L1emulator_ASANfix_12_4_X

895c99a

L1T ASAN heap-buffer-overflow fix resulting from PR #36919 (12_4_X)

cmsbuild added a commit that referenced this pull request Mar 20, 2022

Merge pull request #37253 from gk199/L1emulator_ASANfix

4145e52

L1T ASAN heap-buffer-overflow fix resulting from PR #36919

This was referenced Mar 22, 2022

HCAL: fixing fine grain bits to be set based on unpacked TDC #37302

Merged

HCAL: fixing fine grain bits to be set based on unpacked TDC [12_3_X backport] #37307

Merged

This was referenced Apr 2, 2022

HCAL finegrain bit fix to account for TDC packing, and SOI alignment #37438

Merged

[Backport of] HCAL finegrain bit fix to account for TDC packing, and SOI alignment #37439

Merged

gk199 mentioned this pull request Aug 30, 2022

Calo L1 emulator for LLP trigger: masking the 6:1 LUT for HE #39252

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

L1 emulator for LLP displaced jet trigger (L1 6:1 LUT, L2 jet algo), L1TNtuples addition of jet hwQual #36919

L1 emulator for LLP displaced jet trigger (L1 6:1 LUT, L2 jet algo), L1TNtuples addition of jet hwQual #36919

gk199 commented Feb 9, 2022

cmsbuild commented Feb 9, 2022

cmsbuild commented Feb 9, 2022

cecilecaillol commented Feb 10, 2022

cmsbuild commented Feb 10, 2022

cmsbuild commented Feb 10, 2022

cmsbuild commented Feb 10, 2022

cecilecaillol commented Feb 14, 2022

gk199 commented Feb 16, 2022

cmsbuild commented Feb 16, 2022

perrotta commented Feb 17, 2022

dan131riley commented Feb 21, 2022

gk199 Feb 22, 2022

gk199 Feb 22, 2022

dan131riley Mar 14, 2022

gk199 commented Feb 22, 2022 •

edited

Loading

makortel commented Mar 1, 2022

cecilecaillol commented Mar 14, 2022

gk199 commented Mar 14, 2022

dan131riley commented Mar 14, 2022

gk199 commented Mar 14, 2022

dan131riley commented Mar 15, 2022

L1 emulator for LLP displaced jet trigger (L1 6:1 LUT, L2 jet algo), L1TNtuples addition of jet hwQual #36919

L1 emulator for LLP displaced jet trigger (L1 6:1 LUT, L2 jet algo), L1TNtuples addition of jet hwQual #36919

Conversation

gk199 commented Feb 9, 2022

PR description:

PR validation:

cmsbuild commented Feb 9, 2022

cmsbuild commented Feb 9, 2022

cecilecaillol commented Feb 10, 2022

cmsbuild commented Feb 10, 2022

Build

Clang Build

cmsbuild commented Feb 10, 2022

cmsbuild commented Feb 10, 2022

cecilecaillol commented Feb 14, 2022

gk199 commented Feb 16, 2022

cmsbuild commented Feb 16, 2022

Comparison Summary

perrotta commented Feb 17, 2022

dan131riley commented Feb 21, 2022

gk199 Feb 22, 2022

Choose a reason for hiding this comment

gk199 Feb 22, 2022

Choose a reason for hiding this comment

dan131riley Mar 14, 2022

Choose a reason for hiding this comment

gk199 commented Feb 22, 2022 • edited Loading

makortel commented Mar 1, 2022

cecilecaillol commented Mar 14, 2022

gk199 commented Mar 14, 2022

dan131riley commented Mar 14, 2022

gk199 commented Mar 14, 2022

dan131riley commented Mar 15, 2022

gk199 commented Feb 22, 2022 •

edited

Loading