Improvements to the LST workflows and small general LST fixes

Configuration/PyReleaseValidation/README.md

@@ -66,7 +66,7 @@ The offsets currently in use are:
 * 0.701: DisplacedRegionalStep tracking iteration for Run-3
 * 0.702: trackingMkFit modifier for Phase-2 (initialStep only)
 * 0.703: LST tracking (Phase-2 only), initialStep+HighPtTripletStep only, on CPU
-* 0.704: LST tracking (Phase-2 only), initialStep+HighPtTripletStep only, on GPU
+* 0.704: LST tracking (Phase-2 only), initialStep+HighPtTripletStep only, on GPU (if available)
 * 0.75: HLT phase-2 timing menu
 * 0.751: HLT phase-2 timing menu Alpaka variant
 * 0.752: HLT phase-2 timing menu ticl_v5 variant

Configuration/PyReleaseValidation/python/relval_Run4.py

@@ -37,16 +37,14 @@
 numWFIB.extend([32034.0]) #Run4D115
 numWFIB.extend([32434.0]) #Run4D116
 
-# Temporary placement for LST workflow to workaround PR conflicts - to be formatted and placed in an upcoming PR
-numWFIB.extend([24834.703]) #Run4D98 LST tracking (initialStep+HighPtTripletStep only)
-
 #Additional sample for short matrix and IB
 #Default Phase-2 Det NoPU
 numWFIB.extend([prefixDet+34.911]) #DD4hep XML
 numWFIB.extend([prefixDet+34.702]) #mkFit tracking (initialStep)
 numWFIB.extend([prefixDet+34.5])   #pixelTrackingOnly
 numWFIB.extend([prefixDet+34.9])   #vector hits
 numWFIB.extend([prefixDet+34.402]) #Alpaka local reconstruction offloaded on device (GPU if available)
+numWFIB.extend([prefixDet+34.703]) #LST tracking on CPU (initialStep+HighPtTripletStep only)
 numWFIB.extend([prefixDet+34.21])  #prodlike
 numWFIB.extend([prefixDet+96.0])   #CloseByPGun CE_E_Front_120um
 numWFIB.extend([prefixDet+100.0])  #CloseByPGun CE_H_Coarse_Scint
@@ -60,6 +58,7 @@
 numWFIB.extend([prefixDet+234.21])   #prodlike PU
 numWFIB.extend([prefixDet+234.9921]) #prodlike premix stage1+stage2
 numWFIB.extend([prefixDet+234.114])  #PU, with 10% OT inefficiency
+numWFIB.extend([prefixDet+234.703])  #LST tracking on CPU (initialStep+HighPtTripletStep only)
 #
 numWFIB.extend([24834.911]) #D98 XML, to monitor instability of DD4hep
 for numWF in numWFIB:

Configuration/PyReleaseValidation/python/relval_gpu.py

@@ -70,13 +70,11 @@
            13061.402,
 
            # Run4, Alpaka-based noPU
-           29634.402, 29634.403, 29634.404, 29634.406,
+           29634.402, 29634.403, 29634.404, 29634.406, 29634.704,
            29661.402,
-           # Run4, Alpaka-based noPU GPU LST tracking D98
-           24834.704,
 
            # Run4, Alpaka-based PU
-           29834.402, 29834.403, 29834.404
+           29834.402, 29834.403, 29834.404, 29834.704
         ]
 
 for numWF in numWFIB:

Configuration/PyReleaseValidation/python/upgradeWorkflowComponents.py

@@ -498,8 +498,9 @@ def setup__(self, step, stepName, stepDict, k, properties):
         if 'Reco' in step: stepDict[stepName][k] = merge([self.step3, stepDict[step][k]])
         elif 'HARVEST' in step: stepDict[stepName][k] = merge([{'-s': 'HARVESTING:@trackingOnlyValidation+@trackingOnlyDQM'}, stepDict[step][k]])
         elif 'ALCA' in step: stepDict[stepName][k] = None
-    def condition_(self, fragment, stepList, key, hasHarvest):
-        return ('Run4' in key)
+    def condition(self, fragment, stepList, key, hasHarvest):
+        result = (fragment=="TTbar_14TeV") and hasHarvest and ('Run4' in key)
+        return result
 upgradeWFs['lstOnCPUIters01TrackingOnly'] = UpgradeWorkflow_lstOnCPUIters01TrackingOnly(
     steps = [
         'RecoGlobal',
@@ -508,7 +509,10 @@ def condition_(self, fragment, stepList, key, hasHarvest):
         'ALCA',
         'ALCAPhase2'
     ],
-    PU = [],
+    PU = [
+        'RecoGlobal',
+        'HARVESTGlobal',
+    ],
     suffix = '_lstOnCPUIters01TrackingOnly',
     offset = 0.703,
 )
@@ -517,14 +521,15 @@ def condition_(self, fragment, stepList, key, hasHarvest):
     '--accelerators' : 'cpu'
 }
 
-# LST on GPU, initialStep+highPtTripletStep-only tracking-only
+# LST on GPU (if available), initialStep+highPtTripletStep-only tracking-only
 class UpgradeWorkflow_lstOnGPUIters01TrackingOnly(UpgradeWorkflowTracking):
     def setup__(self, step, stepName, stepDict, k, properties):
         if 'Reco' in step: stepDict[stepName][k] = merge([self.step3, stepDict[step][k]])
         elif 'HARVEST' in step: stepDict[stepName][k] = merge([{'-s': 'HARVESTING:@trackingOnlyValidation+@trackingOnlyDQM'}, stepDict[step][k]])
         elif 'ALCA' in step: stepDict[stepName][k] = None
-    def condition_(self, fragment, stepList, key, hasHarvest):
-        return ('Run4' in key)
+    def condition(self, fragment, stepList, key, hasHarvest):
+        result = (fragment=="TTbar_14TeV") and hasHarvest and ('Run4' in key)
+        return result
 upgradeWFs['lstOnGPUIters01TrackingOnly'] = UpgradeWorkflow_lstOnGPUIters01TrackingOnly(
     steps = [
         'RecoGlobal',
@@ -533,13 +538,15 @@ def condition_(self, fragment, stepList, key, hasHarvest):
         'ALCA',
         'ALCAPhase2'
     ],
-    PU = [],
+    PU = [
+        'RecoGlobal',
+        'HARVESTGlobal',
+    ],
     suffix = '_lstOnGPUIters01TrackingOnly',
     offset = 0.704,
 )
 upgradeWFs['lstOnGPUIters01TrackingOnly'].step3 = upgradeWFs['trackingOnly'].step3 | {
     '--procModifiers': 'trackingIters01,trackingLST',
-    '--accelerators' : 'gpu-*'
 }
 
 #DeepCore seeding for JetCore iteration workflow

DQM/TrackingMonitorSource/python/TrackingSourceConfig_Tier0_cff.py

@@ -395,6 +395,11 @@ def _copyIfExists(mod, pset, name):
 from Configuration.ProcessModifiers.seedingDeepCore_cff import seedingDeepCore
 seedingDeepCore.toReplaceWith(TrackSeedMonSequence,_seedingDeepCore_TrackSeedMonSequence)
 
+from Configuration.ProcessModifiers.trackingLST_cff import trackingLST
+trackingLST.toModify(locals()["TrackSeedMonhighPtTripletStep"],
+    SeedProducer = "lstPixelSeedInputProducer"
+)
+
 TrackingDQMSourceTier0 += TrackSeedMonSequence
 
 from DQM.TrackingMonitorSource.shortTrackResolution_cff import *

RecoTracker/LST/plugins/LSTOutputConverter.cc

@@ -259,6 +259,7 @@ void LSTOutputConverter::produce(edm::Event& iEvent, const edm::EventSetup& iSet
 
   LogDebug("LSTOutputConverter") << "done with conversion: Track candidate output size = " << outputpTC.size()
                                  << " (p* objects) + " << outputT5TC.size() << " (T5 objects)";
+  std::vector<SeedStopInfo> outputSeedStopInfo(pixelSeeds.size());
   iEvent.emplace(trajectorySeedPutToken_, std::move(outputTS));
   iEvent.emplace(trajectorySeedpLSPutToken_, std::move(outputpLSTS));
   iEvent.emplace(trackCandidatePutToken_, std::move(outputTC));
@@ -267,7 +268,7 @@ void LSTOutputConverter::produce(edm::Event& iEvent, const edm::EventSetup& iSet
   iEvent.emplace(trackCandidateNopLSTCPutToken_, std::move(outputNopLSTC));
   iEvent.emplace(trackCandidatepTTCPutToken_, std::move(outputpTTC));
   iEvent.emplace(trackCandidatepLSTCPutToken_, std::move(outputpLSTC));
-  iEvent.emplace(seedStopInfoPutToken_, 0U);  //dummy stop info
+  iEvent.emplace(seedStopInfoPutToken_, std::move(outputSeedStopInfo));  //dummy stop info
 }
 
 DEFINE_FWK_MODULE(LSTOutputConverter);

RecoTracker/LSTCore/BuildFile.xml

@@ -1,6 +1,7 @@
 <use name="alpaka"/>
 <use name="boost_header"/>
 <use name="root"/>
+<use name="FWCore/MessageLogger"/>
 <use name="HeterogeneousCore/AlpakaInterface"/>
 <use name="RecoTracker/LSTCore" for="alpaka"/>
 <flags ALPAKA_BACKENDS="1"/>

RecoTracker/LSTCore/interface/Common.h

@@ -1,8 +1,8 @@
 #ifndef RecoTracker_LSTCore_interface_Common_h
 #define RecoTracker_LSTCore_interface_Common_h
 
-#include "HeterogeneousCore/AlpakaInterface/interface/config.h"
 #include "DataFormats/Common/interface/StdArray.h"
+#include "HeterogeneousCore/AlpakaInterface/interface/config.h"
 
 #if defined(FP16_Base)
 #if defined ALPAKA_ACC_GPU_CUDA_ENABLED

RecoTracker/LSTCore/interface/QuintupletsSoA.h

@@ -24,8 +24,8 @@ namespace lst {
                       SOA_COLUMN(bool, tightCutFlag),  // tight pass to be a TC
                       SOA_COLUMN(bool, partOfPT5),
                       SOA_COLUMN(float, regressionRadius),
-                      SOA_COLUMN(float, regressionG),
-                      SOA_COLUMN(float, regressionF),
+                      SOA_COLUMN(float, regressionCenterX),
+                      SOA_COLUMN(float, regressionCenterY),
                       SOA_COLUMN(float, rzChiSquared),  // r-z only chi2
                       SOA_COLUMN(float, chiSquared),
                       SOA_COLUMN(float, nonAnchorChiSquared),

RecoTracker/LSTCore/interface/alpaka/Common.h

@@ -3,6 +3,7 @@
 
 #include <numbers>
 
+#include "FWCore/MessageLogger/interface/MessageLogger.h"
 #include "RecoTracker/LSTCore/interface/Common.h"
 
 namespace ALPAKA_ACCELERATOR_NAMESPACE::lst {
@@ -11,6 +12,11 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst {
 
   Vec3D constexpr elementsPerThread(Vec3D::all(static_cast<Idx>(1)));
 
+  ALPAKA_FN_HOST ALPAKA_FN_INLINE void lstWarning(std::string warning) {
+    edm::LogWarning("LST") << warning;
+    return;
+  }
+
   // Adjust grid and block sizes based on backend configuration
   template <typename Vec, typename TAcc = Acc<typename Vec::Dim>>
   ALPAKA_FN_HOST ALPAKA_FN_INLINE WorkDiv<typename Vec::Dim> createWorkDiv(const Vec& blocksPerGrid,

RecoTracker/LSTCore/src/ModuleMethods.h

@@ -334,9 +334,8 @@ namespace lst {
         host_moduleType[index] = (m_t == 25 ? lst::TwoS : lst::PS);
         host_moduleLayerType[index] = (m_t == 23 ? lst::Pixel : lst::Strip);
 
-        if (host_moduleType[index] == lst::PS and host_moduleLayerType[index] == lst::Pixel) {
-          host_isAnchor[index] = true;
-        } else if (host_moduleType[index] == lst::TwoS and host_isLower[index]) {
+        if ((host_moduleType[index] == lst::PS and host_moduleLayerType[index] == lst::Pixel) ||
+            (host_moduleType[index] == lst::TwoS and host_isLower[index])) {
           host_isAnchor[index] = true;
         } else {
           host_isAnchor[index] = false;

RecoTracker/LSTCore/src/alpaka/Hit.h

@@ -56,27 +56,6 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst {
     return dPhi;
   }
 
-  ALPAKA_FN_HOST_ACC ALPAKA_FN_INLINE int binary_search(const unsigned int* data,  // Array that we are searching over
-                                                        unsigned int search_val,  // Value we want to find in data array
-                                                        unsigned int ndata)       // Number of elements in data array
-  {
-    unsigned int low = 0;
-    unsigned int high = ndata - 1;
-
-    while (low <= high) {
-      unsigned int mid = (low + high) / 2;
-      unsigned int test_val = data[mid];
-      if (test_val == search_val)
-        return mid;
-      else if (test_val > search_val)
-        high = mid - 1;
-      else
-        low = mid + 1;
-    }
-    // Couldn't find search value in array.
-    return -1;
-  }
-
   struct ModuleRangesKernel {
     template <typename TAcc>
     ALPAKA_FN_ACC void operator()(TAcc const& acc,
@@ -129,13 +108,19 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst {
             ((ihit_z > 0) - (ihit_z < 0)) *
             alpaka::math::acosh(
                 acc, alpaka::math::sqrt(acc, ihit_x * ihit_x + ihit_y * ihit_y + ihit_z * ihit_z) / hits.rts()[ihit]);
-        int found_index = binary_search(modules.mapdetId(), iDetId, nModules);
+        auto found_pointer = std::lower_bound(modules.mapdetId(), modules.mapdetId() + nModules, iDetId);
+        int found_index = std::distance(modules.mapdetId(), found_pointer);
+        if (found_pointer == modules.mapdetId() + nModules)
+          found_index = -1;
         uint16_t lastModuleIndex = modules.mapIdx()[found_index];
 
         hits.moduleIndices()[ihit] = lastModuleIndex;
 
         if (modules.subdets()[lastModuleIndex] == Endcap && modules.moduleType()[lastModuleIndex] == TwoS) {
-          found_index = binary_search(geoMapDetId, iDetId, nEndCapMap);
+          found_pointer = std::lower_bound(geoMapDetId, geoMapDetId + nEndCapMap, iDetId);
+          found_index = std::distance(geoMapDetId, found_pointer);
+          if (found_pointer == geoMapDetId + nEndCapMap)
+            found_index = -1;
           float phi = geoMapPhi[found_index];
           float cos_phi = alpaka::math::cos(acc, phi);
           hits.highEdgeXs()[ihit] = ihit_x + 2.5f * cos_phi;

RecoTracker/LSTCore/src/alpaka/LSTEvent.dev.cc

@@ -173,11 +173,12 @@ void LSTEvent::addPixelSegmentToEvent(std::vector<unsigned int> const& hitIndice
   unsigned int size = ptIn.size();
 
   if (size > n_max_pixel_segments_per_module) {
-    printf(
-        "*********************************************************\n"
-        "* Warning: Pixel line segments will be truncated.       *\n"
-        "* You need to increase n_max_pixel_segments_per_module. *\n"
-        "*********************************************************\n");
+    lstWarning(
+        "\
+        *********************************************************\n\
+        * Warning: Pixel line segments will be truncated.       *\n\
+        * You need to increase n_max_pixel_segments_per_module. *\n\
+        *********************************************************");
     size = n_max_pixel_segments_per_module;
   }
 
@@ -714,12 +715,13 @@ void LSTEvent::createTrackCandidates(bool no_pls_dupclean, bool tc_pls_triplets)
   auto nTrackCandidatesT5 = *nTrackCanT5Host_buf.data();
   if ((nTrackCandidatespT5 + nTrackCandidatespT3 + nTrackCandidatespLS == n_max_pixel_track_candidates) ||
       (nTrackCandidatesT5 == n_max_nonpixel_track_candidates)) {
-    printf(
-        "****************************************************************************************************\n"
-        "* Warning: Track candidates were possibly truncated.                                               *\n"
-        "* You may need to increase either n_max_pixel_track_candidates or n_max_nonpixel_track_candidates. *\n"
-        "* Run the code with the WARNINGS flag activated for more details.                                  *\n"
-        "****************************************************************************************************\n");
+    lstWarning(
+        "\
+        ****************************************************************************************************\n\
+        * Track candidates were possibly truncated.                                                        *\n\
+        * You may need to increase either n_max_pixel_track_candidates or n_max_nonpixel_track_candidates. *\n\
+        * Run the code with the WARNINGS flag activated for more details.                                  *\n\
+        ****************************************************************************************************");
   }
 }
 

RecoTracker/LSTCore/src/alpaka/PixelQuintuplet.h

@@ -597,8 +597,8 @@ namespace ALPAKA_ACCELERATOR_NAMESPACE::lst {
     centerX = segmentsPixel.circleCenterX()[pixelSegmentArrayIndex];
     centerY = segmentsPixel.circleCenterY()[pixelSegmentArrayIndex];
 
-    float T5CenterX = quintuplets.regressionG()[quintupletIndex];
-    float T5CenterY = quintuplets.regressionF()[quintupletIndex];
+    float T5CenterX = quintuplets.regressionCenterX()[quintupletIndex];
+    float T5CenterY = quintuplets.regressionCenterY()[quintupletIndex];
     quintupletRadius = quintuplets.regressionRadius()[quintupletIndex];
 
     rPhiChiSquared = computePT5RPhiChiSquared(acc, modules, lowerModuleIndices, centerX, centerY, pixelRadius, xs, ys);