ITS: new CPU + GPU seeding vertexer by cima22 · Pull Request #15733 · AliceO2Group/AliceO2

cima22 · 2026-08-27T11:12:06Z

Adds an optional seeding vertexer that runs as a prepended tracker pass (diamond trackleting -> cells -> lines -> parallel seeding), on both the CPU and GPU traits.
Additionally, unsorted clusters are loaded to GPU and then sorted on GPU.

To set it:
ITSCATrackerParam.seedingVertexIteration=1;
ITSVertexerParam.useParallelSeeding=1;

Tuned parameters for Pb--Pb:

ITSCATrackerParam.seedingVertexIteration=1;ITSVertexerParam.useParallelSeeding=1;ITSVertexerParam.useTruthSeeding=0;ITSCATrackerParam.diamondTrackletingPVres=1.5102214480735774;ITSCATrackerParam.diamondTrackletingNSigmaCut=2.976946815605669;ITSCATrackerParam.diamondTrackletingCellDeltaTanLambdaSigma=0.005566295924057422;ITSCATrackerParam.diamondCellTanLambdaNSigma=2.5;ITSCATrackerParam.diamondTrackletingCellDeltaPhiMinPt=0.12;ITSCATrackerParam.cellLineSharedClusterCut=1;ITSVertexerParam.clusterCut=0.07;ITSVertexerParam.pairCut=0.025684202919354766;ITSVertexerParam.phiCut=0.008875191490279988;ITSVertexerParam.clusterContributorsCut=3;ITSVertexerParam.lineMinPt=0.10;ITSVertexerParam.nSigmaCut=0.0016;ITSVertexerParam.goodLineChi2Cut=9.814679066181695;ITSVertexerParam.goodLinePtCut=0.12;ITSVertexerParam.goodContributorsSignificance=0.070;ITSVertexerParam.suppressLowMultDebris=13;ITSVertexerParam.fineZWindow=0.010;ITSVertexerParam.fineMinDensity=8;ITSVertexerParam.fineMaxDrift=0.005;ITSVertexerParam.duplicateZScale=0.7;ITSVertexerParam.duplicateZCut=0.1388966993405415;

Tuned parameters for pp:

ITSCATrackerParam.seedingVertexIteration=1;ITSVertexerParam.useParallelSeeding=1;ITSVertexerParam.useTruthSeeding=0;ITSCATrackerParam.diamondTrackletingPVres=3.74344532796542;ITSCATrackerParam.diamondTrackletingNSigmaCut=5.352263543811316;ITSCATrackerParam.diamondTrackletingCellDeltaTanLambdaSigma=0.002563403956702525;ITSCATrackerParam.diamondCellTanLambdaNSigma=3.257473162361808;ITSCATrackerParam.diamondTrackletingCellDeltaPhiMinPt=0.09407482537938365;ITSCATrackerParam.cellLineSharedClusterCut=3;ITSVertexerParam.clusterCut=0.05557266167329535;ITSVertexerParam.pairCut=0.05176031785438137;ITSVertexerParam.phiCut=0.017639175978851104;ITSVertexerParam.clusterContributorsCut=2;ITSVertexerParam.lineMinPt=0.1;ITSVertexerParam.nSigmaCut=0.0016;ITSVertexerParam.goodLineChi2Cut=9.814679066181695;ITSVertexerParam.goodLinePtCut=0.12;ITSVertexerParam.goodContributorsSignificance=0.07;ITSVertexerParam.suppressLowMultDebris=13;ITSVertexerParam.fineZWindow=0.013014502308220061;ITSVertexerParam.fineMinDensity=4;ITSVertexerParam.fineMaxDrift=0.004003953018135234;ITSVertexerParam.duplicateZScale=0.1230148343342053;ITSVertexerParam.duplicateZCut=0.06571512005272727;

shahor02 · 2026-08-27T12:14:39Z

Hi @cima22 , are there some slides or plots on the new algo., particularly its performance? Perhaps you could present it next Wed., either at the ITS meeting at 11 (if it is planned, @fprino ?) or at sim/rec WP12/13 at 15:30?

f3sch

Hi @cima22, thanks! Some preliminary comments although I have to admit to not understand the logic in its entirety yet.

f3sch · 2026-08-27T12:26:02Z

Do we need to duplicate the already existing ClusterLines math, can we not have only one of these classes?

Could we promote the GPU class to be the common one? The CPU one uses ROOT, I don't think it is available for GPUs?

sure, you can adapt the CPU class however you like to make it GPU compatible.

now there should be only one class 👍

cima22 · 2026-08-27T15:33:40Z

@shahor02 This algorithm should be tested, but I have some preliminary results.

On 50 simulated PbPb TFs at 50kHz:

On a PbPb TF with embedded pp collisions:

PbPb:

pp:

pp track efficiency:

shahor02 · 2026-08-27T15:43:41Z

Thanks, looks good! Note that there are conflicts in the PR.
So, will you be available to present it next Wed.?

cima22 · 2026-08-27T15:45:43Z

Thanks, looks good! Note that there are conflicts in the PR. So, will you be available to present it next Wed.?

Yes absolutely

fprino · 2026-08-27T17:36:59Z

Thanks, looks good! Note that there are conflicts in the PR. So, will you be available to present it next Wed.?

Yes absolutely

Sorry for replying late: next week we are considering to move the ITS WP2 meeting from Wednesday 2 september to Thursday 3 September, starting at 10 am.
Would this work for you (Gabriele, Ruben, Felix)?

f3sch · 2026-08-27T17:38:16Z

@fprino I will be on vacation from Thursday onwards.

f3sch · 2026-08-27T17:44:19Z

Some things from an initial test:
Timing:

[PROD][fschlepp@epn000 pvtest]$ python3 time.py gabriele/reco.log
Processing:
Original count: 925
Trimmed count:  649
Discarded:      138 low + 138 high
Total:          336405.15 ms (336.41 s)
Average:        518.34 ms
Std deviation:  194.22 ms

Vertex seeding:
Original count: 925
Trimmed count:  649
Discarded:      138 low + 138 high
Total:          60625.25 ms (60.63 s)
Average:        93.41 ms
Std deviation:  13.94 ms

Tracking:
Original count: 919
Trimmed count:  645
Discarded:      137 low + 137 high
Total:          231575.46 ms (231.58 s)
Average:        359.03 ms
Std deviation:  86.99 ms
[PROD][fschlepp@epn000 pvtest]$ python3 time.py dev/reco.log

Processing:
Original count: 925
Trimmed count:  649
Discarded:      138 low + 138 high
Total:          511572.46 ms (511.57 s)
Average:        788.25 ms
Std deviation:  239.93 ms

Vertex seeding:
Original count: 925
Trimmed count:  649
Discarded:      138 low + 138 high
Total:          222618.46 ms (222.62 s)
Average:        343.02 ms
Std deviation:  89.72 ms

Tracking:
Original count: 924
Trimmed count:  648
Discarded:      138 low + 138 high
Total:          240011.05 ms (240.01 s)
Average:        370.39 ms
Std deviation:  70.58 ms

note this timing from MI50 EPNs. the new vertexing seems to be 3.7 times faster which in total means a speedup of 1.5 in total processing time (-34.2%).

23 kHz Pb-Pb.
red:dev, blue:this PR, green:this PR but relaxing ITSVertexerParam.clusterContributorsCut from 3 to 2:
pv_cont.pdf
There seems be recovery of PVs below 80 contributors and a slight degradation for higher mult vertices.

black:dev, red:this PR, blue:this PR but relaxing ITSVertexerParam.clusterContributorsCut from 3 to 2:
K0sFitOverlay.pdf
K0sFitVsPt.pdf
K0s yield decreases by 0.2% although this is dominated by the 1-1.5 GeV/c bin.

Note that there is currently no second iteration for UPC flagging.

shahor02 · 2026-08-27T18:05:39Z

Then let's have it presented at the wp12/13 meeting next Wed. 15:30.
@cima22 Was suppressLowMultDebris=13 added to pp settings intentionally? Usually, we don't remove debris there.
@f3sch thanks, looks good. I assume for the UPC you left default settings?

f3sch · 2026-08-27T18:09:29Z

@f3sch thanks, looks good. I assume for the UPC you left default settings?

@shahor02 there is only one iteration done in this PR, if you mean the one from dev then yes.

shahor02 · 2026-08-27T18:29:55Z

OK, I saw https://github.com/AliceO2Group/AliceO2/pull/15733/changes#diff-27fc8aaeb7769a496f838182d30a759ba63b3163bc7f458dd9bbbc895d3c7211R258-R260, but then the ITSVertexerParam are irrelevant in that mode, and there is no ITS trackfinding UPC iteration?
If the answer is long, don't waste time, will learn at the presentation.

f3sch · 2026-08-27T19:09:24Z

OK, I saw https://github.com/AliceO2Group/AliceO2/pull/15733/changes#diff-27fc8aaeb7769a496f838182d30a759ba63b3163bc7f458dd9bbbc895d3c7211R258-R260, but then the ITSVertexerParam are irrelevant in that mode, and there is no ITS trackfinding UPC iteration?

There is no track finding since no ROFs are flagged by a vertex carrying the UPC flag in the UPC mask, all ROFs are automatically masked in UPC iteration

cima22 · 2026-08-28T07:13:15Z

Then let's have it presented at the wp12/13 meeting next Wed. 15:30.

@cima22 Was suppressLowMultDebris=13 added to pp settings intentionally? Usually, we don't remove debris there.

@f3sch thanks, looks good. I assume for the UPC you left default settings?

Hi, sorry today I am travelling so I won't be very responsive. I didn't put yet the UPC iteration in the PR, I will test one in the next days.

suppressLowMultDebris=13 slipped when copying the parameters, but in any case it is not used by this algorithm by default, thanks for noticing.

cima22 · 2026-09-04T11:03:42Z

Fixed comments by @f3sch and added UPC iteration. For PbPb:

ITSCATrackerParam.diamondTrackletingCellDeltaPhiMinPt=0.12;ITSCATrackerParam.diamondCellTanLambdaNSigma=2.5;ITSCATrackerParam.seedingVertexIteration=1;ITSCATrackerParam.doUPCIteration=1;ITSCATrackerParam.cellLineSharedClusterCut=1;ITSCATrackerParam.diamondTrackletingPVres=1.5102214480735774;ITSCATrackerParam.diamondTrackletingNSigmaCut=2.976946815605669;ITSCATrackerParam.diamondTrackletingCellDeltaTanLambdaSigma=0.005566295924057422;ITSVertexerParam.suppressLowMultDebris=13;ITSVertexerParam.goodLineChi2Cut=9.814679066181695;ITSVertexerParam.goodLinePtCut=0.12;ITSVertexerParam.clusterContributorsCut=3;ITSVertexerParam.clusterCut=0.07;ITSVertexerParam.duplicateZCut=0.1388966993405415;ITSVertexerParam.pairCut=0.025684202919354766;ITSVertexerParam.nSigmaCut=0.0016;ITSVertexerParam.phiCut=0.008875191490279988;ITSVertexerParam.goodContributorsSignificance=0.070;ITSVertexerParam.fineZWindow=0.010;ITSVertexerParam.fineMinDensity=8;ITSVertexerParam.fineMaxDrift=0.005;ITSVertexerParam.lineMinPt=0.10;ITSVertexerParam.duplicateZScale=0.7;

For PbPb UPC iteration, clusterContributorsCut is set to 2, but it may need other tuning.

alibuild · 2026-09-04T21:44:26Z

Error while checking build/O2/fullCI_slc9 for 6c20ad5 at 2026-09-05 15:28:

## sw/BUILD/O2Physics-latest/log
c++: fatal error: Killed signal terminated program cc1plus
c++: fatal error: Killed signal terminated program cc1plus


## sw/BUILD/o2checkcode-latest/log
--
========== List of errors found ==========
++ GRERR=0
++ grep -v clang-diagnostic-error error-log.txt
++ grep ' error:'
grep: error-log.txt: binary file matches
++ GRERR=1
++ [[ 1 == 0 ]]
++ mkdir -p /sw/INSTALLROOT/2820ef2fd7971b0b882b12016412662dd544f397/slc9_x86-64/o2checkcode/1.0-local557/etc/modulefiles
++ alibuild-generate-module
+ HAS_BIN=
+ HAS_LIB=
+ HAS_CMAKE=
+ HAS_ROOT=
+ '[' 0 '!=' 0 ']'
+ cat
+ printf 'if ![ is-loaded '\''BASE/1.0'\'' ] {\n module load BASE/1.0\n}'
+ echo 'BUILD_REQUIRES=CMake alibuild-recipe-tools defaults-release'
BUILD_REQUIRES=CMake alibuild-recipe-tools defaults-release
+ FULL_BUILD_REQUIRES='FairCMakeModules safe_int UUID googlebenchmark double-conversion ninja-fortran Python-modules-list defaults-release Alice-GRID-Utils MPFR json-c date bz2 ninja pytorch_cpuinfo O2-customization alibuild-recipe-tools CMake'
+ echo 'FULL_BUILD_REQUIRES=FairCMakeModules safe_int UUID googlebenchmark double-conversion ninja-fortran Python-modules-list defaults-release Alice-GRID-Utils MPFR json-c date bz2 ninja pytorch_cpuinfo O2-customization alibuild-recipe-tools CMake'
FULL_BUILD_REQUIRES=FairCMakeModules safe_int UUID googlebenchmark double-conversion ninja-fortran Python-modules-list defaults-release Alice-GRID-Utils MPFR json-c date bz2 ninja pytorch_cpuinfo O2-customization alibuild-recipe-tools CMake
++ env
++ cut -f1 -d=
++ grep -v '^DEFAULT_'
++ grep -v PKGREVISION
++ grep REVISION
++ grep -v ALIBUILD_RECIPE_TOOLS
++ sed -e s/_REVISION//
+ for x in $(env | cut -f1 -d= | grep -v "^DEFAULT_" | grep -v PKGREVISION | grep -v ALIBUILD_RECIPE_TOOLS | grep REVISION | sed -e 's/_REVISION//')
++ eval 'echo $LIBFFI_REVISION'
+++ echo 14
+ REVISION_VALUE=14
++ eval 'echo $LIBFFI_VERSION'
+++ echo v3.2.1-alice1
+ VERSION_VALUE=v3.2.1-alice1
++ eval 'echo $LIBFFI_ROOT'
+++ echo /sw/slc9_x86-64/libffi/v3.2.1-alice1-14
+ ROOT_PATH_VALUE=/sw/slc9_x86-64/libffi/v3.2.1-alice1-14
+ '[' -z 14 ']'
+ echo 'FairCMakeModules safe_int UUID googlebenchmark double-conversion ninja-fortran Python-modules-list defaults-release Alice-GRID-Utils MPFR json-c date bz2 ninja pytorch_cpuinfo O2-customization alibuild-recipe-tools CMake'
+ tr '[:lower:]' '[:upper:]'
+ tr - _
[0 more errors; see full log]

Full log here.

Adds a seeding vertexer that runs as a prepended tracker pass (diamond trackleting -> cells -> lines -> parallel seeding), on both the CPU and GPU traits, replacing the per-ROF CPU vertexer for the seeding step.

alibuild · 2026-09-07T17:11:56Z

Error while checking build/O2/fullCI_slc9 for 9257bc9 at 2026-09-07 19:11:

## sw/BUILD/O2-full-system-test-latest/log
command alien-token-init had nonzero exit code 2


## sw/BUILD/o2checkcode-latest/log
--
========== List of errors found ==========
++ GRERR=0
++ grep -v clang-diagnostic-error error-log.txt
++ grep ' error:'
grep: error-log.txt: binary file matches
++ GRERR=1
++ [[ 1 == 0 ]]
++ mkdir -p /sw/INSTALLROOT/669a1b23d87054378167507b6800fd1385f5d456/slc9_x86-64/o2checkcode/1.0-local2/etc/modulefiles
++ alibuild-generate-module
+ HAS_BIN=
+ HAS_LIB=
+ HAS_CMAKE=
+ HAS_ROOT=
+ '[' 0 '!=' 0 ']'
+ cat
+ printf 'if ![ is-loaded '\''BASE/1.0'\'' ] {\n module load BASE/1.0\n}'
+ echo 'BUILD_REQUIRES=CMake alibuild-recipe-tools defaults-release'
BUILD_REQUIRES=CMake alibuild-recipe-tools defaults-release
+ FULL_BUILD_REQUIRES='ninja-fortran pytorch_cpuinfo FairCMakeModules UUID json-c safe_int double-conversion ninja MPFR bz2 googlebenchmark defaults-release O2-customization Python-modules-list date CMake alibuild-recipe-tools Alice-GRID-Utils'
+ echo 'FULL_BUILD_REQUIRES=ninja-fortran pytorch_cpuinfo FairCMakeModules UUID json-c safe_int double-conversion ninja MPFR bz2 googlebenchmark defaults-release O2-customization Python-modules-list date CMake alibuild-recipe-tools Alice-GRID-Utils'
FULL_BUILD_REQUIRES=ninja-fortran pytorch_cpuinfo FairCMakeModules UUID json-c safe_int double-conversion ninja MPFR bz2 googlebenchmark defaults-release O2-customization Python-modules-list date CMake alibuild-recipe-tools Alice-GRID-Utils
++ env
++ cut -f1 -d=
++ grep -v '^DEFAULT_'
++ grep -v PKGREVISION
++ grep -v ALIBUILD_RECIPE_TOOLS
++ grep REVISION
++ sed -e s/_REVISION//
+ for x in $(env | cut -f1 -d= | grep -v "^DEFAULT_" | grep -v PKGREVISION | grep -v ALIBUILD_RECIPE_TOOLS | grep REVISION | sed -e 's/_REVISION//')
++ eval 'echo $LIBFFI_REVISION'
+++ echo 14
+ REVISION_VALUE=14
++ eval 'echo $LIBFFI_VERSION'
+++ echo v3.2.1-alice1
+ VERSION_VALUE=v3.2.1-alice1
++ eval 'echo $LIBFFI_ROOT'
+++ echo /sw/slc9_x86-64/libffi/v3.2.1-alice1-14
+ ROOT_PATH_VALUE=/sw/slc9_x86-64/libffi/v3.2.1-alice1-14
+ '[' -z 14 ']'
+ echo 'ninja-fortran pytorch_cpuinfo FairCMakeModules UUID json-c safe_int double-conversion ninja MPFR bz2 googlebenchmark defaults-release O2-customization Python-modules-list date CMake alibuild-recipe-tools Alice-GRID-Utils'
+ tr '[:lower:]' '[:upper:]'
+ tr - _
+ tr ' ' '\n'
[0 more errors; see full log]

Full log here.

f3sch

Hi @cima22, some more nits to be addressed but I think besides the MC label on GPU thingy, we can merge this after a small local test.

f3sch · 2026-09-07T13:10:30Z

  auto trackParams = TrackingMode::getTrackingParameters(mMode);
  auto vertParams = TrackingMode::getVertexingParameters(mMode);
  overrideParameters(trackParams, vertParams);
+  for (auto& p : trackParams) {
+    if (p.PassFlags[IterationStep::SeedingVertexPass]) {
+      p.CreateArtefactLabels = mIsMC;
+    }
+  }


We should not change the tracking parameters outside of the Configuration.cxx, maybe we can just propagate mIsMC into the get*Parameters functions?

f3sch · 2026-09-07T13:15:06Z

+      const float cellDeltaPhiCut =
+        mTrkParams[iteration].PassFlags[IterationStep::SeedingVertexPass]
+          ? math_utils::cellDeltaPhiBound(mBz, mTrkParams[iteration].CellDeltaPhiMinPt,
+                                          mTrkParams[iteration].LayerRadii[firstLink.fromLayer],
+                                          mTrkParams[iteration].LayerRadii[firstLink.toLayer],
+                                          mTrkParams[iteration].LayerRadii[secondLink.toLayer],
+                                          mTimeFrame->getLinkMSAngle(cellTopology.firstLink))
+          : -1.f;


nit: the indentation feels off here should it not be something like:

Suggested change

const float cellDeltaPhiCut =

mTrkParams[iteration].PassFlags[IterationStep::SeedingVertexPass]

? math_utils::cellDeltaPhiBound(mBz, mTrkParams[iteration].CellDeltaPhiMinPt,

mTrkParams[iteration].LayerRadii[firstLink.fromLayer],

mTrkParams[iteration].LayerRadii[firstLink.toLayer],

mTrkParams[iteration].LayerRadii[secondLink.toLayer],

mTimeFrame->getLinkMSAngle(cellTopology.firstLink))

: -1.f;

const float cellDeltaPhiCut = mTrkParams[iteration].PassFlags[IterationStep::SeedingVertexPass]

? math_utils::cellDeltaPhiBound(mBz, mTrkParams[iteration].CellDeltaPhiMinPt,

mTrkParams[iteration].LayerRadii[firstLink.fromLayer],

mTrkParams[iteration].LayerRadii[firstLink.toLayer],

mTrkParams[iteration].LayerRadii[secondLink.toLayer],

mTimeFrame->getLinkMSAngle(cellTopology.firstLink))

: -1.f;

f3sch · 2026-09-07T13:17:31Z

-
-    for (int iLayer{0}; iLayer < NLayers; ++iLayer) {
-      clearResizeBoundedVector(mIndexTables[iLayer], getNrof(iLayer) * ((trkParam.ZBins * trkParam.PhiBins) + 1), getMaybeFrameworkHostResource());
-    }
-    for (int iLayer{0}; iLayer < trkParam.NLayers; ++iLayer) {
-      if (trkParam.SystErrorY2[iLayer] > 0.f || trkParam.SystErrorZ2[iLayer] > 0.f) {
-        for (auto& tfInfo : mTrackingFrameInfo[iLayer]) {
-          /// Account for alignment systematics in the cluster covariance matrix
-          tfInfo.covarianceTrackingFrame[0] += trkParam.SystErrorY2[iLayer];
-          tfInfo.covarianceTrackingFrame[2] += trkParam.SystErrorZ2[iLayer];
+    allocateClusterSortStorage(trkParam, maxLayers);
+    // Applied by the only FirstPass pass that covers every layer (tracking iteration 0).
+    // The 3-layer seeding passes must be excluded: they run first and would otherwise apply this to layers 0..2 only
+    if (trkParam.NLayers == NLayers) {
+      for (int iLayer{0}; iLayer < trkParam.NLayers; ++iLayer) {
+        if (trkParam.SystErrorY2[iLayer] > 0.f || trkParam.SystErrorZ2[iLayer] > 0.f) {
+          for (auto& tfInfo : mTrackingFrameInfo[iLayer]) {
+            tfInfo.covarianceTrackingFrame[0] += trkParam.SystErrorY2[iLayer];
+            tfInfo.covarianceTrackingFrame[2] += trkParam.SystErrorZ2[iLayer];
+          }
        }
      }
    }


maybe I misunderstand, but why apply the additional uncertainty on the cluster position not for the seeding pass?

f3sch · 2026-09-07T13:27:15Z

I am not sure why this is necessary, the MC labels we only compute on the host (and general we are not using the labels in the first place for GPU)?

f3sch · 2026-09-15T12:07:33Z

-#pragma link C++ class o2::its::Line + ;
-#pragma link C++ class std::vector < o2::its::Line> + ;
-
-#pragma link C++ class o2::its::ClusterLines + ;
-#pragma link C++ class std::vector < o2::its::ClusterLines> + ;


not that it is important since we never used it but why do we drop the dictionaries?

f3sch · 2026-09-15T12:17:32Z

+  const bool useDiamond = this->mTrkParams[iteration].UseDiamond;
+  if (useDiamond) {
+    const Vertex diamondVert(this->mTrkParams[iteration].Diamond, this->mTrkParams[iteration].DiamondCov, 1, 1.f);
+    mTimeFrameGPU->createDiamondDevice(diamondVert);


I would probably have increased simply the time error on the diamond to be max, called updateROFLookup once and avoided the extra gating on vtxMode

f3sch · 2026-09-15T12:19:03Z

+      total += evaluateTask(&Tracker::computeBeamFromVertices, StateNames[mCurStep = BeamPositioning], it, logger, it);
+      const float dx = mTimeFrame->getBeamX() - prevBeamX;
+      const float dy = mTimeFrame->getBeamY() - prevBeamY;
+      if (dx * dx + dy * dy < constants::BeamConvergence2) {
+        logger(std::format(" - Beam bootstrap converged after pass {} (beam shift < 50 um)", pass));
+        break;


you have the early return in the computeBeamFromVertices but I would lift it to be here without running any of this.

f3sch · 2026-09-15T12:21:46Z

  ROFVertexLookupTableN::View mROFVertexLookupTableView;
  ROFMaskTableN mMultiplicityCutMask;
  ROFMaskTableN mUPCCutMask;
+  ROFMaskTableN mSeedingUPCMask;


Why is there an extra mask required for the UPC seeding?

f3sch · 2026-09-15T12:22:56Z

+  std::vector<o2::its::Line> mLinesHost;
+  std::vector<int> mLineRofHost;
+  std::vector<int> mLineClustersHost;
+  std::vector<VertexCand> mVertexCandsHost;
+  std::vector<int> mPeakOffsetsHost;
+  PeakMembershipHost mPeakMembershipHost;
+  std::vector<o2::MCCompLabel> mLineLabelFlatHost;


should these not use the bounded_vector?

f3sch · 2026-09-15T12:24:35Z

+GPUdi() int clusterROF(const int* rofArr, const int nRofs, const int clusterIdx)
+{
+  const int key = clusterIdx + 1;
+  int lo = 0, hi = nRofs + 1;
+  while (lo < hi) {
+    const int mid = (lo + hi) >> 1;
+    if (rofArr[mid] < key) {
+      lo = mid + 1;
+    } else {
+      hi = mid;
+    }
+  }
+  return lo - 1;
+}


debris, this should go into the Utils.h

cima22 requested review from f3sch, fprino, mconcas, mcoquet642 and shahor02 as code owners August 27, 2026 11:12

alibuild mentioned this pull request Aug 27, 2026

Please consider the following formatting changes to #15733 cima22/AliceO2#10

Closed

cima22 force-pushed the parallel-vertexer branch from 5c4ab24 to 06b0b02 Compare August 27, 2026 11:18

f3sch reviewed Aug 27, 2026

View reviewed changes

cima22 force-pushed the parallel-vertexer branch from 06b0b02 to d8a8978 Compare September 4, 2026 10:59

cima22 force-pushed the parallel-vertexer branch from d8a8978 to 6c20ad5 Compare September 4, 2026 13:25

cima22 requested a review from a team as a code owner September 4, 2026 13:25

ITS: new CPU + GPU seeding vertexer

9257bc9

Adds a seeding vertexer that runs as a prepended tracker pass (diamond trackleting -> cells -> lines -> parallel seeding), on both the CPU and GPU traits, replacing the per-ROF CPU vertexer for the seeding step.

cima22 force-pushed the parallel-vertexer branch from 6c20ad5 to 9257bc9 Compare September 7, 2026 12:15

f3sch requested changes Sep 15, 2026

View reviewed changes

Uh oh!

Conversation

cima22 commented Aug 27, 2026

Uh oh!

shahor02 commented Aug 27, 2026

Uh oh!

f3sch left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

cima22 commented Aug 27, 2026

On 50 simulated PbPb TFs at 50kHz:

On a PbPb TF with embedded pp collisions:

PbPb:

pp:

pp track efficiency:

Uh oh!

shahor02 commented Aug 27, 2026

Uh oh!

cima22 commented Aug 27, 2026

Uh oh!

fprino commented Aug 27, 2026

Uh oh!

f3sch commented Aug 27, 2026

Uh oh!

f3sch commented Aug 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

shahor02 commented Aug 27, 2026

Uh oh!

f3sch commented Aug 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

shahor02 commented Aug 27, 2026

Uh oh!

f3sch commented Aug 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cima22 commented Aug 28, 2026

Uh oh!

cima22 commented Sep 4, 2026

Uh oh!

alibuild commented Sep 4, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

alibuild commented Sep 7, 2026

Uh oh!

f3sch left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

f3sch commented Aug 27, 2026 •

edited

Loading

f3sch commented Aug 27, 2026 •

edited

Loading

f3sch commented Aug 27, 2026 •

edited

Loading

alibuild commented Sep 4, 2026 •

edited

Loading