Skip to content

Commit

Permalink
Fix: reduce objects in track-v0 task (AliceO2Group#5055)
Browse files Browse the repository at this point in the history 
There is an upper limit on objects that can be defined in a workflow and
that limit was hit in the track-v0 task. Removing the old same and mixed
event process function that will not be used in the future resolves the
problem.
  • Loading branch information
@ariedel-cern
ariedel-cern authored Mar 7, 2024
1 parent 5aba8cf commit 4de71cf
Showing 1 changed file with 9 additions and 107 deletions.
116 changes: 9 additions & 107 deletions PWGCF/FemtoDream/Tasks/femtoDreamPairTaskTrackV0.cxx
View file
Original file line number Diff line number Diff line change
Expand Up @@ -251,20 +251,14 @@ struct femtoDreamPairTaskTrackV0 {
}
}
}
if ((doprocessSameEvent && doprocessSameEventMasked) ||
(doprocessMixedEvent && doprocessMixedEventMasked) ||
(doprocessSameEventMC && doprocessSameEventMCMasked) ||
(doprocessMixedEventMC && doprocessMixedEventMCMasked)) {
LOG(fatal) << "Normal and masked processing cannot be activated simultaneously!";
}
}

/// This function processes the same event and takes care of all the histogramming
template <bool isMC, typename PartitionType, typename TableTracks, typename Collision>
void doSameEvent(PartitionType& SliceTrk1, PartitionType& SliceV02, TableTracks& parts, Collision col)
void doSameEvent(PartitionType& SliceTrk1, PartitionType& SliceV02, TableTracks const& parts, Collision const& col)
{
/// Histogramming same event
for (auto& part : SliceTrk1) {
for (auto const& part : SliceTrk1) {
trackHistoPartOne.fillQA<isMC, false>(part, aod::femtodreamparticle::kPt, col.multNtr(), col.multV0M());
}
for (auto& v0 : SliceV02) {
Expand All @@ -287,7 +281,7 @@ struct femtoDreamPairTaskTrackV0 {
}
}
/// Now build particle combinations
for (auto& [p1, p2] : combinations(CombinationsFullIndexPolicy(SliceTrk1, SliceV02))) {
for (auto const& [p1, p2] : combinations(CombinationsFullIndexPolicy(SliceTrk1, SliceV02))) {
const auto& posChild = parts.iteratorAt(p2.index() - 2);
const auto& negChild = parts.iteratorAt(p2.index() - 1);
// cuts on V0 children still need to be applied
Expand All @@ -308,18 +302,6 @@ struct femtoDreamPairTaskTrackV0 {
}
}

void processSameEvent(FilteredCollision const& col, FilteredFDParticles const& parts)
{
eventHisto.fillQA(col);
auto SliceTrk1 = PartitionTrk1->sliceByCached(aod::femtodreamparticle::fdCollisionId, col.globalIndex(), cache);
auto SliceV02 = PartitionV02->sliceByCached(aod::femtodreamparticle::fdCollisionId, col.globalIndex(), cache);
if (SliceTrk1.size() == 0 && SliceV02.size() == 0) {
return;
}
doSameEvent<false>(SliceTrk1, SliceV02, parts, col);
}
PROCESS_SWITCH(femtoDreamPairTaskTrackV0, processSameEvent, "Enable processing same event", true);

void processSameEventMasked(FilteredMaskedCollision const& col, FilteredFDParticles const& parts)
{
if ((col.bitmaskTrackOne() & BitMask) != BitMask && (col.bitmaskTrackTwo() & BitMask) != BitMask) {
Expand All @@ -330,21 +312,9 @@ struct femtoDreamPairTaskTrackV0 {
auto SliceV02 = PartitionV02->sliceByCached(aod::femtodreamparticle::fdCollisionId, col.globalIndex(), cache);
doSameEvent<false>(SliceTrk1, SliceV02, parts, col);
}
PROCESS_SWITCH(femtoDreamPairTaskTrackV0, processSameEventMasked, "Enable processing same event with masks", false);

void processSameEventMC(FilteredCollision& col, FilteredFDMCParts& parts, o2::aod::FDMCParticles&)
{
eventHisto.fillQA(col);
auto SliceMCTrk1 = PartitionMCTrk1->sliceByCached(aod::femtodreamparticle::fdCollisionId, col.globalIndex(), cache);
auto SliceMCV02 = PartitionMCV02->sliceByCached(aod::femtodreamparticle::fdCollisionId, col.globalIndex(), cache);
if (SliceMCTrk1.size() == 0 && SliceMCV02.size() == 0) {
return;
}
doSameEvent<true>(SliceMCTrk1, SliceMCV02, parts, col);
}
PROCESS_SWITCH(femtoDreamPairTaskTrackV0, processSameEventMC, "Enable processing same event MC", false);
PROCESS_SWITCH(femtoDreamPairTaskTrackV0, processSameEventMasked, "Enable processing same event with masks", true);

void processSameEventMCMasked(FilteredMaskedCollision& col, FilteredFDMCParts& parts, o2::aod::FDMCParticles&)
void processSameEventMCMasked(FilteredMaskedCollision const& col, FilteredFDMCParts const& parts, o2::aod::FDMCParticles const&)
{
if ((col.bitmaskTrackOne() & BitMask) != BitMask && (col.bitmaskTrackTwo() & BitMask) != BitMask) {
return;
Expand All @@ -357,39 +327,7 @@ struct femtoDreamPairTaskTrackV0 {
PROCESS_SWITCH(femtoDreamPairTaskTrackV0, processSameEventMCMasked, "Enable processing same event MC with masks", false);

template <bool isMC, typename CollisionType, typename PartType, typename PartitionType, typename BinningType>
void doMixedEvent_NotMasked(CollisionType& cols, PartType& parts, PartitionType& part1, PartitionType& part2, BinningType policy)
{
for (auto const& [collision1, collision2] : soa::selfCombinations(policy, ConfMixingDepth.value, -1, cols, cols)) {
auto SliceTrk1 = part1->sliceByCached(aod::femtodreamparticle::fdCollisionId, collision1.globalIndex(), cache);
auto SliceV02 = part2->sliceByCached(aod::femtodreamparticle::fdCollisionId, collision2.globalIndex(), cache);
if (SliceTrk1.size() == 0 || SliceV02.size() == 0) {
continue;
}
for (auto& [p1, p2] : combinations(CombinationsFullIndexPolicy(SliceTrk1, SliceV02))) {
const auto& posChild = parts.iteratorAt(p2.index() - 2);
const auto& negChild = parts.iteratorAt(p2.index() - 1);
// check cuts on V0 children
if (((posChild.cut() & ConfV02_ChildPos_CutBit) == ConfV02_ChildPos_CutBit) &&
((posChild.pidcut() & ConfV02_ChildPos_TPCBit) == ConfV02_ChildPos_TPCBit) &&
((negChild.cut() & ConfV02_ChildNeg_CutBit) == ConfV02_ChildNeg_CutBit) &&
((negChild.pidcut() & ConfV02_ChildNeg_TPCBit) == ConfV02_ChildNeg_TPCBit)) {
continue;
}
if (ConfOptUseCPR.value) {
if (pairCloseRejection.isClosePair(p1, p2, parts, collision1.magField())) {
continue;
}
}
if (!pairCleaner.isCleanPair(p1, p2, parts)) {
continue;
}
mixedEventCont.setPair<isMC>(p1, p2, collision1.multNtr(), collision1.multV0M(), ConfOptUse4D, ConfOptExtendedPlots, ConfOptsmearingByOrigin);
}
}
}

template <bool isMC, typename CollisionType, typename PartType, typename PartitionType, typename BinningType>
void doMixedEvent_Masked(CollisionType& cols, PartType& parts, PartitionType& part1, PartitionType& part2, BinningType policy)
void doMixedEvent_Masked(CollisionType const& cols, PartType const& parts, PartitionType& part1, PartitionType& part2, BinningType policy)
{
Partition<CollisionType> PartitionMaskedCol1 = (aod::femtodreamcollision::bitmaskTrackOne & BitMask) == BitMask && aod::femtodreamcollision::downsample == true;
Partition<CollisionType> PartitionMaskedCol2 = (aod::femtodreamcollision::bitmaskTrackTwo & BitMask) == BitMask && aod::femtodreamcollision::downsample == true;
Expand Down Expand Up @@ -421,25 +359,7 @@ struct femtoDreamPairTaskTrackV0 {
}
}

void processMixedEvent(FilteredCollisions& cols, FilteredFDParticles& parts)
{
switch (ConfMixingPolicy.value) {
case femtodreamcollision::kMult:
doMixedEvent_NotMasked<false>(cols, parts, PartitionTrk1, PartitionV02, colBinningMult);
break;
case femtodreamcollision::kMultPercentile:
doMixedEvent_NotMasked<false>(cols, parts, PartitionTrk1, PartitionV02, colBinningMultPercentile);
break;
case femtodreamcollision::kMultMultPercentile:
doMixedEvent_NotMasked<false>(cols, parts, PartitionTrk1, PartitionV02, colBinningMultMultPercentile);
break;
default:
LOG(fatal) << "Invalid binning policiy specifed. Breaking...";
}
}
PROCESS_SWITCH(femtoDreamPairTaskTrackV0, processMixedEvent, "Enable processing mixed events", true);

void processMixedEventMasked(FilteredMaskedCollisions& cols, FilteredFDParticles& parts)
void processMixedEventMasked(FilteredMaskedCollisions const& cols, FilteredFDParticles const& parts)
{
switch (ConfMixingPolicy.value) {
case femtodreamcollision::kMult:
Expand All @@ -455,27 +375,9 @@ struct femtoDreamPairTaskTrackV0 {
LOG(fatal) << "Invalid binning policiy specifed. Breaking...";
}
}
PROCESS_SWITCH(femtoDreamPairTaskTrackV0, processMixedEventMasked, "Enable processing mixed events with masks", false);

void processMixedEventMC(FilteredCollisions& cols, FilteredFDMCParts& parts, o2::aod::FDMCParticles&)
{
switch (ConfMixingPolicy.value) {
case femtodreamcollision::kMult:
doMixedEvent_NotMasked<true>(cols, parts, PartitionMCTrk1, PartitionMCV02, colBinningMult);
break;
case femtodreamcollision::kMultPercentile:
doMixedEvent_NotMasked<true>(cols, parts, PartitionMCTrk1, PartitionMCV02, colBinningMultPercentile);
break;
case femtodreamcollision::kMultMultPercentile:
doMixedEvent_NotMasked<true>(cols, parts, PartitionMCTrk1, PartitionMCV02, colBinningMultMultPercentile);
break;
default:
LOG(fatal) << "Invalid binning policiy specifed. Breaking...";
}
}
PROCESS_SWITCH(femtoDreamPairTaskTrackV0, processMixedEventMC, "Enable processing mixed events MC", false);
PROCESS_SWITCH(femtoDreamPairTaskTrackV0, processMixedEventMasked, "Enable processing mixed events with masks", true);

void processMixedEventMCMasked(FilteredMaskedCollisions& cols, FilteredFDMCParts& parts, o2::aod::FDMCParticles&)
void processMixedEventMCMasked(FilteredMaskedCollisions const& cols, FilteredFDMCParts const& parts, o2::aod::FDMCParticles const&)
{
switch (ConfMixingPolicy.value) {
case femtodreamcollision::kMult:
Expand Down

0 comments on commit 4de71cf

Please sign in to comment.