More Configurable Pixel Tracks and HIon Template #41632
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Contributor
|
-code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-41632/35517
Code check has found code style and quality issues which could be resolved by applying following patch(s)
|
slava77
reviewed
May 11, 2023
| @@ -0,0 +1 @@ | |||
| tests passed, failed | |||
AdrianoDee
force-pushed
the
more_configurable_ca_hion
branch
from
048f9f1 to
03a9bdd
Compare
Contributor
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-41632/35519
|
Contributor
|
A new Pull Request was created by @AdrianoDee (Adriano Di Florio) for master. It involves the following packages:
@civanch, @Dr15Jones, @kskovpen, @pmandrik, @bsunanda, @makortel, @bbilin, @fwyzard, @mdhildreth, @mandrenguyen, @cmsbuild, @AdrianoDee, @srimanob, @clacaputo, @syuvivida, @nothingface0, @sunilUIET, @tjavaid, @micsucmed, @emanueleusai, @rvenditti can you please review it and eventually sign? Thanks. cms-bot commands are listed here |
AdrianoDee
force-pushed
the
more_configurable_ca_hion
branch
from
03a9bdd to
5e079d8
Compare
Contributor
Author
|
test parameters:
|
slava77
reviewed
May 11, 2023
| hiPixelTracksCUDA = _pixelTracksCUDA.clone(pixelRecHitSrc="siPixelRecHitsPreSplittingCUDA", idealConditions = False, | ||
| ptmin = 0.25, hardCurvCut = 0.0756, doPtCut = False, | ||
| onGPU = True, | ||
| phiCuts = cms.vint32([900 for i in range(19)]), #19 pairs |
Contributor
There was a problem hiding this comment.
Suggested change
| phiCuts = cms.vint32([900 for i in range(19)]), #19 pairs | |
| phiCuts = cms.vint32(19*[900]), #19 pairs |
should work as well
Contributor
Author
There was a problem hiding this comment.
And looks better: applied.
| ptmin = 0.25, hardCurvCut = 0.0756, doPtCut = False, | ||
| onGPU = True, | ||
| phiCuts = cms.vint32([900 for i in range(19)]), #19 pairs | ||
| trackQualityCuts = cms.PSet( |
Contributor
There was a problem hiding this comment.
shouldn't this be
Suggested change
| trackQualityCuts = cms.PSet( | |
| trackQualityCuts = dict( |
and then drop all cms.sometype specifications
| module.noFlowDists.append(tmp[ind-1]) | ||
|
|
||
| _defaultSubdirs = ["Tracking/Track/*", "Tracking/TrackTPPtLess09/*", "Tracking/TrackFromPV/*", "Tracking/TrackFromPVAllTP/*", "Tracking/TrackAllTPEffic/*", "Tracking/TrackBuilding/*","Tracking/TrackConversion/*", "Tracking/TrackGsf/*"] | ||
| _defaultSubdirs = ["Tracking/Track/*", "Tracking/TrackTPPtLess09/*", "Tracking/HIPixelTrack/*" , "Tracking/TrackFromPV/*", "Tracking/TrackFromPVAllTP/*", "Tracking/TrackAllTPEffic/*", "Tracking/TrackBuilding/*","Tracking/TrackConversion/*", "Tracking/TrackGsf/*"] |
Contributor
There was a problem hiding this comment.
should this addition be era-dependent? 1/8 increment of folders is not that large, but still, seeing HI it in pp is not that expected
Contributor
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-41632/35520
|
Contributor
|
+1 |
Contributor
|
@cms-sw/pdmv-l2 your signature is the only thing missing before merging this PR and finally build the already two weeks late CMSSW_13_2_0_pre3. Please either sign or state clearly which is the status of your review and when you plan to conclude it. |
Contributor
|
+pdmv |
cmsbuild
added
fully-signed
pdmv-approved
and removed
pending-signatures
pdmv-pending
labels
Contributor
|
This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @perrotta, @dpiparo, @rappoccio (and backports should be raised in the release meeting by the corresponding L2) |
Contributor
|
+1
|
cmsbuild
added a commit
that referenced
this pull request
Jul 27, 2023
Clean Up for Pixel Tracks - Follow-up to #41632
missirol
added a commit
to missirol/cmssw
that referenced
this pull request
Jul 28, 2023
missirol
added a commit
to missirol/cmssw
that referenced
this pull request
Jul 28, 2023
This was referenced Jul 28, 2023
cmsbuild
added a commit
that referenced
this pull request
Jul 31, 2023
…1632 fix configuration of GPU Pixel unpacker in Run-3 HLT menus post-#41632
cmsbuild
added a commit
that referenced
this pull request
Aug 1, 2023
…1632_132X fix configuration of GPU Pixel unpacker in Run-3 HLT menus post-#41632 [`13_2_X`]
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
15 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR proposes some modifications to the GPU (CUDA based) pixel track reconstruction in order, also, to be run smoothly for HIon events. The modifications include:
HIonPhase1pixel topology;maxNumClustersPerModulestoTrackerTraitsand increasing it to2048for HIon operations (and exampanding theblockPrefixScanfor theclusterChargeCut);gpuCalibPixel(such as the gains and the readout modes) configurable at run-time extendingSiPixelClusterThresholdsstruct. This is not strictly needed but I feel would ease the things in future if these constants need to be changed;SiPixelRawToClusterCUDAin order to include the changes when using an HIon topology (note that the name change is needed since, as far as I have understood, HLT would need to have both module names defined in the same release to do the migration);maxNumberOfDoublets,z0Cut,ptCutandphiCutsconfigurable at run time (this will be needed to run offline pixel tracking/seeding on GPU);*.5) and "Patatrack" (*.501,*.502) together withHydjet_Quenched_MinBias_5020GeV_cfiandHydjet_Quenched_MinBias_5362GeV_cfiin the matrix;.501and.502) for GPU-likehiConformalPixelTracksfor160workflow;doSplittingflag), useless for HIon operations, given also that too much shared memory would be needed for the average number of tracks per vertices (MAXTK) in PbPb conditions;Any comment on this by HIon people is very welcome, especially on the w.f.s introduced.
Solves #40623 #37425.
HI Physics performance plots
HIon pixel tracking performance for:
14950.502_HydjetQMinBias_5362GeV+2023HI_Patatrack_PixelOnlyGPUhere;hiConformalPixelTrackswith w.f.160.502_HydjetQ_MinBias_5362GeV_2023_ppRecohere;Note: these performance may be improved or changed and the sizes of the structures should be better tuned to fit the T4@P5 memory. But I'd leave this to HIon people that know better than me what they need. This PR is here to give the skeleton/template and to may way easier the validation and tuning.
Run2 2018 Data Throughput plots
The pp performance are basically untouched both in physics and throughput.
On
/EphemeralHLTPhysics1/Run2018D-v1/RAWrun=323775lumi=5310 July 2023 Update
Addressed comments (from @fwyzard). Mostly:
CUDADataFormats/SiPixelCluster/interface/gpuClusteringConstants.h;static_cast<T>where needed;gpuClusterChargeCutto run for more than2*maxThreadsclusters per module;~phase2_tracker;List of modules for @Martin-Grunewald .
Under
DQM/SiPixelHeterogeneous/-
SiPixelHIonPhase1CompareRecHitsSoA-
SiPixelHIonPhase1CompareTrackSoA-
SiPixelHIonPhase1MonitorRecHitsSoA-
SiPixelHIonPhase1MonitorTrackSoAUnder
RecoLocalTracker/SiPixelClusterizer/-
SiPixelDigisClustersFromSoAHIonPhase1-
SiPixelRawToClusterCUDAHIonPhase1Under
RecoLocalTracker/SiPixelRecHits/-
PixelCPEFastESProducerHIonPhase1-
SiPixelRecHitCUDAHIonPhase1-
SiPixelRecHitFromCUDAHIonPhase1-
SiPixelRecHitSoAFromCUDAHIonPhase1-
SiPixelRecHitSoAFromLegacyHIonPhase1Under
RecoTracker/PixelSeeding/-
CAHitNtupletCUDAHIonPhase1Under
RecoTracker/PixelTrackFitting/-
PixelTrackDumpCUDAHIonPhase1-
PixelTrackProducerFromSoAHIonPhase1-
PixelTrackSoAFromCUDAHIonPhase1Under
RecoTracker/PixelVertexFinding/-
PixelVertexProducerCUDAHIonPhase112 July 2023 Update
As sanity check, MTV plots by @CesarBernardes for HIon Conformal Pixel Tracks here before and after the comments.
Updated throughput plots on
/EphemeralHLTPhysics1/Run2018D-v1/RAWrun=323775lumi=53:(FYI @denerslemos)