EIC code displayed by LXR master/​include/​celeritas/​optical/​gen/​WlsGeneratorAction.hh	Source navigation Diff markup Identifier search General search
	Version: master test Revert   Change

File indexing completed on 2026-11-07 10:11:38

 //------------------------------- -*- C++ -*- -------------------------------//
 // Copyright Celeritas contributors: see top-level COPYRIGHT file for details
 // SPDX-License-Identifier: (Apache-2.0 OR MIT)
 //---------------------------------------------------------------------------//
 //! \file celeritas/optical/gen/WlsGeneratorAction.hh
 //---------------------------------------------------------------------------//
 #pragma once
 
 #include <memory>
 
 #include "corecel/Macros.hh"
 #include "corecel/cont/Span.hh"
 #include "corecel/data/AuxInterface.hh"
 #include "corecel/data/AuxStateVec.hh"
 #include "celeritas/optical/action/ActionInterface.hh"
 #include "celeritas/phys/GeneratorInterface.hh"
 
 #include "GeneratorBase.hh"
 #include "GeneratorData.hh"
 #include "OffloadData.hh"
 
 namespace celeritas
 {
 namespace optical
 {
 class CoreParams;
 class CoreStateBase;
 class WavelengthShiftModel;
 
 //---------------------------------------------------------------------------//
 /*!
  * Generate photons from optical wavelength shifting distribution data.
  *
  * This samples and initializes optical photons directly in a track slot in a
  * reproducible way. At each step of the optical loop:
  * - The distribution buffer is compacted to remove invalid distributions
  * - An inclusive scan over \c num_photons is computed. All distributions up to
  *   the buffer size are guaranteed to have a positive photon count.
  * - Each thread is assigned to exactly one photon via the cumulative sum \c
  *   offsets array and generates one WLS secondary in a vacant track slot.
  * - The photon count in the distribution is decremented atomically after
  *   generation. Because the thread to distribution mapping is determined from
  *   the pre-generation cumulative sum, \c num_photons can only reach zero
  *   after all threads assigned to a distribution have generated from it. The
  *   updated counts are used solely to identify exhausted distributions during
  *   the subsequent compaction.
  * - The buffer is compacted again to free space for new distributions that may
  *   be added during the step.
  *
  * This differs from the Cherenkov/scintillation generation in a few ways:
  * - New WLS distributions can be added during the optical stepping loop, so
  *   the buffer must be compacted each step to reclaim space.
  * - Because the buffer is compacted before generation, there are no
  *   zero-photon distributions in the active portion of the buffer; the
  *   executor does not need to skip over them by finding the \c buffer_start
  *   offset.
  * - The number of photons generated in a WLS interaction is small (O(1))
  *   compared to scintillation, which can be large (O(1000)), so there should
  *   be very few collisions in the atomic update and it should have no
  *   performance impact.
  */
 class WlsGeneratorAction final : public GeneratorBase
 {
   public:
     //!@{
     //! \name Type aliases
     using SPConstWavelengthShiftModel
         = std::shared_ptr<WavelengthShiftModel const>;
     //!@}
 
   public:
     struct Input
     {
         ActionId action_id;
         AuxId aux_id;
         GeneratorId gen_id;
         SPConstWavelengthShiftModel wls;
         SPConstWavelengthShiftModel wls2;
         size_type capacity{};
     };
 
     // Construct with IDs, WLS model, and buffer capacity
     explicit WlsGeneratorAction(Input&&);
 
     //!@{
     //! \name Aux interface
 
     // Build state data for a stream
     UPState create_state(MemSpace, StreamId, size_type) const final;
     //!@}
 
     //!@{
     //! \name StepAction interface
 
     // Launch kernel with host data
     void step(CoreParams const&, CoreStateHost&) const final;
     // Launch kernel with device data
     void step(CoreParams const&, CoreStateDevice&) const final;
     //!@}
 
   private:
     //// DATA ////
 
     SPConstWavelengthShiftModel wls_;
     SPConstWavelengthShiftModel wls2_;
     size_type capacity_;
 
     //// HELPER FUNCTIONS ////
 
     template<MemSpace M>
     void step_impl(CoreParams const&, CoreState<M>&) const;
 
     void generate(CoreParams const&, CoreStateHost&) const;
     void generate(CoreParams const&, CoreStateDevice&) const;
 };
 
 //---------------------------------------------------------------------------//
 }  // namespace optical
 }  // namespace celeritas

This page was automatically generated by the 2.3.7 LXR engine. The LXR team