Tracking developments for review and merging #338

fwyzard · 2019-04-30T10:20:11Z

Split #324, part 7 of 8.

fwyzard · 2019-04-30T15:22:19Z

SimTracker/TrackerHitAssociation/plugins/BuildFile.xml

@@ -11,4 +11,5 @@
  <use name="cuda"/>
  <use name="cuda-api-wrappers"/>
  <flags EDM_PLUGIN="1"/>
+  <flags CUDA_FLAGS="-g -DGPU_DEBUG"/>


I think we should not use these flags for a production plugin.

fwyzard · 2019-04-30T15:25:11Z

SimTracker/TrackerHitAssociation/plugins/trackerHitAssociationHeterogeneousProduct.h

@@ -21,8 +21,10 @@ namespace trackerHitAssociationHeterogeneousProduct {

  struct ClusterSLGPU {

+   using Clus2TP = std::array<uint32_t,6>;


Please use a more meaningful or at least less abbreviated name, for example 'ClusterToTP' (assuming that is the meaning).

What is the meaning of 6 ?

Validation/RecoTrack/python/TrackingParticleSelectionsForEfficiency_cff.py

fwyzard · 2019-05-02T00:24:36Z

Thanks to @rovere for cleaning up the commit history !

fwyzard · 2019-05-02T07:18:34Z

Filter spurious commits from upstream pull request cms-sw#26074 .

fwyzard · 2019-05-02T16:40:53Z

Clean up spurious commits that were already included in CMSSW_10_6_X_Patatrack

fwyzard · 2019-05-02T21:07:25Z

CalibTracker/SiPixelESProducers/interface/SiPixelGainCalibrationForHLTGPU.h

@@ -17,6 +17,8 @@ class SiPixelGainCalibrationForHLTGPU {
  ~SiPixelGainCalibrationForHLTGPU();

  const SiPixelGainForHLTonGPU *getGPUProductAsync(cuda::stream_t<>& cudaStream) const;
+  const SiPixelGainForHLTonGPU *getCPUProduct() const { return gainForHLTonHost_;}
+  const SiPixelGainCalibrationForHLT *getOriginalProduct() { return gains_;}


Are these new methods used anywhere ?
Otherwise, do we (still) want them ?

fwyzard · 2019-05-02T21:07:43Z

CalibTracker/SiPixelESProducers/src/SiPixelGainCalibrationForHLTGPU.cc

@@ -26,7 +26,7 @@ SiPixelGainCalibrationForHLTGPU::SiPixelGainCalibrationForHLTGPU(const SiPixelGa
  */

  cudaCheck(cudaMallocHost((void**) & gainForHLTonHost_, sizeof(SiPixelGainForHLTonGPU)));
-  //gainForHLTonHost_->v_pedestals = gainDataOnGPU_; // how to do this?
+  gainForHLTonHost_->v_pedestals = (SiPixelGainForHLTonGPU_DecodingStructure*)this->gains_->data().data();  // so it can be used on CPU as well...


what does this do ?

fwyzard · 2019-05-03T17:08:31Z

current commit map

Add two alternative (faster) track clusterizers: one based on DBSCAN, and one "by density"; use the latter by default. Allow all pixel triplets, but protect the vertex from triplets. Use a larger-then-needed nearest neightbours array to allow for possible duplicate pixels, as a pixel can appear more than once in the same event. Use a separate workspace for temporary data.

Improve pixel doublets and CA: - add pixel cluster size and shape cuts in doublets; - add triplet cleaner; - improved cluster size studies - implement layer-dependent cuts in the CA. Add counters in GPU code and possibility to test full doublet combinatorics. Update python notebook and include z0 resolution.

…amework (#338) Use cleaned hits. Use pixel layer and ladders geometry, and use pixel triplets in the gaps. Optimise GPU memory usage: - reduce the number of memory allocations - fix the size of the cub workspace - allocate memory per event via the caching allocator - use constant memory for geometry and parameters - use shared memory where the content is the same for every thread Optimise kernel launches, and add a protection for empty events and overflows.

Add two alternative (faster) track clusterizers: one based on DBSCAN, and one "by density"; use the latter by default. Allow all pixel triplets, but protect the vertex from triplets. Use a larger-then-needed nearest neightbours array to allow for possible duplicate pixels, as a pixel can appear more than once in the same event. Use a separate workspace for temporary data.

Improve pixel doublets and CA: - add pixel cluster size and shape cuts in doublets; - add triplet cleaner; - improved cluster size studies - implement layer-dependent cuts in the CA. Add counters in GPU code and possibility to test full doublet combinatorics. Update python notebook and include z0 resolution.

…amework (#338) Use cleaned hits. Use pixel layer and ladders geometry, and use pixel triplets in the gaps. Optimise GPU memory usage: - reduce the number of memory allocations - fix the size of the cub workspace - allocate memory per event via the caching allocator - use constant memory for geometry and parameters - use shared memory where the content is the same for every thread Optimise kernel launches, and add a protection for empty events and overflows.

Merge the Riemann and broken line fits into single configurable pixel n-tuplet fitter, and extend it to work with up to 5 hits. Mmake the broken line fit the default algorithm. Try both triplets and quadruplets in the pixel "hole". Limit pT used to compute the multple scattering. Use the inline Cholesky decomposition. Generic clean up and improvements.