EIC code displayed by LXR master/​include/​G4HepEm/​G4HepEmElectronData.hh	Source navigation Diff markup Identifier search General search
	Version: master test Revert   Change

File indexing completed on 2026-07-13 08:21:03

 #ifndef G4HepEmElectronData_HH
 #define G4HepEmElectronData_HH
 
 /**
  * @file    G4HepEmElectronData.hh
  * @struct  G4HepEmElectronData
  * @author  M. Novak
  * @date    2020
  *
  * @brief All energy loss process related data used for \f$e^-/e+\f$ simulations by `G4HepEm`.
  *
  * Energy loss processes are the *Ionisation* and *Bremsstrahlung* interactions
  * described by using the **condensed history** approach. It means, that *sub-threshold*
  * *interactions* are modelled as **continous energy losses** along the particle steps
  * while *super-threshold interactions*, i.e. **generation of secondary** \f$e^-/\gamma\f$
  * *particles* in case of ionisation/bremsstrahlung with intitial energy above the
  * secondary production threshold, are modelled explicitly as point like, **discrete**
  * **interactions**. This data structure contains all the data required to account these
  * interactions at run-time.
  *
  * The **continous energy loss** is characterised by the **restricted stopping power**, which
  * is the mean value of the energy losses due to the sub-threshold interactions
  * along a unit step lenght, and other related quantities such as the corresponding
  * **restricted range** or **restricted inverse range** values. The
  * *restricted* (ionisation - electronic)/(bremsstrahlung - radiative) *stopping power*
  * depend on the primary particle type, kinetic energy, target material and secondary
  * \f$e^-\f$(electronic)/\f$\gamma\f$(radiative) production threshold values. Therefore,
  * such tables are built separately \f$e^-\f$ and \f$e^+\f$, over a wide enough range of
  * primary particle kinetic energy (100 [eV] - 100 [TeV] by default) for all
  * material and secondary prodcution threshold pairs, or according to the Geant4 terminology,
  * for all material - cuts couples. The tables contain the *sum of the electronic and radiative*
  * contributions. These tables, stored in this data structure,
  * are **used to determine the continous** part of the **step limit** and **to compute the**
  * sub-threshold realted, **continous energy losses** at each step done by \f$e^-/e^+\f$.
  *
  * The rate of the **dicrete** super-threshold ionisation and bremsstrahlung
  * interactions are characterised by the corresponding **restricted macroscopic cross sections**.
  * These also depend on the primary particle type, kinetic energy, target material -cuts
  * couple. Moreover, the minimum value of the kinetic energy grid is determined by the
  * secondary \f$e^-\f$(ionisation)\f$/\gamma\f$(bremsstrahlung) production energy thresholds
  * that (since Geant4 rquires the user to specify these in lenght) is diffrent in case of
  * each material-cuts couple whenever wither the material or the cut value (in length)
  * is different. Therefore, these *restricted macroscopic cross sction* tables are
  * *built* separately for \f$e^-/e^+\f$ primary particles, separately for *ionisation*
  * and *bremsstrahlung* for all different material - cuts couples with individual
  * kinetic energy grids. These tables are **used to determine the discrete** part of the
  * **step limit**, i.e. the length that the primary \f$e^-/e^+\f$ travels till the next
  * ionisation/bremsstrahlung interaction in the given material - cuts couple,
  * resulting in secondary \f$e^-/\gamma\f$ particle production with initial energy
  * above the secondary \f$e^-/\gamma\f$ production cut in the given material - cuts
  * couple.
  *
  * The *macroscopic cross section* determines the (mean) path length, the primary
  * particles travels, till the next *discrete* interaction in the given *material*
  * (actually material - cuts couple in case of ioni. and brem. since we use the
  * condensed history approach).
  * The *discrete* bremsstrahlung interaction takes place in the vicinity of one
  * of the *elements of the material*. A so-called **target atom selector** is
  * constructed for each model (also for ionisation though its not used), based
  * on the partial contribution of the individual elements of a given
  * material - cuts to the corresponding macroscopic cross section. These data
  * are **used to select the target atom for discrete interaction** at run-time.
  *
  *
  *
  * @note
  * Other interactions, beyond ionisation and bremsstrahlung, are also active
  * in case of \f$e^-/e^+\f$ like Coulomb scattering or annihilation into two
  * \f$\gamma\f$-s in case of \f$e^+\f$. However, their descriptions are rather
  * different compared to these two energy loss processes:
  *    - Coulomb scattering is described by a so-called *multiple scattering model*
  *    - while \f$e^+\f$ annihilation is a discrete process, so similar to the discrete
  *      ionisation and bremsstrahlung interactions, the corresponding cross section
  *      doesn't depend neither the secondary production thresholds nor the element
  *      composition (directly). Unlike ionisation or even more bremsstrahlung, the
  *      cross section for annihilation can be easily computed on-the-fly so there
  *      is no need to pre-compute and store values in tables.
  *
  */
 
 struct G4HepEmElectronData {
   /** Number of G4HepEm material - cuts: number of G4HepEmMCCData structures stored in the G4HepEmMatCutData::fMatCutData array. */
   int        fNumMatCuts   = 0;
   /** Number of G4HepEm material : number of G4HepEmMatData structures stored in the G4HepEmMaterialData::fMaterialData array. */
   int        fNumMaterials = 0;
 
 
 //// === ENERGY LOSS DATA
   /**
    * @name Energy loss related data members:
    * These members are used to store all continuous energy loss related data (there
    * is a single primary kinetic energy grid for all material - cuts equally spaced
    * in log-scale).
    */
 ///@{
   /** Number of discrete kinetic energy values in the grid (\f$N\f$). */
   int        fELossEnergyGridSize = 0;
   /** Logarithm of the minimum kinetic energy value of the grid (\f$\ln(E_0)\f$)*/
   double     fELossLogMinEkin = 0.0;     // log of the E_0
   /** Inverse of the log-scale delta value (\f$ 1/[log(E_{N-1}/E_0)/(N-1)]\f$). */
   double     fELossEILDelta = 0.0;
   /** The grid of the discrete kinetic energy values (\f$E_0, E_1,\ldots, E_{N-1}\f$).*/
   double*    fELossEnergyGrid = nullptr; // [fELossEnergyGridSize]
   /** The energy loss data: **restricted dE/dx, range and inverse range** data.
     *
     * The restricted dE/dx, range (and corresponding inverse range) data values,
     * over the above kinetic energy grid for all material - cuts couples, are
     * stored continuously in this G4HepEmElectronData::fELossData single array.
     * The second derivative values, required for the run-time spline interpolation,
     * are also stored together with the data.
     *
     * The data are stored in the following format for each of the G4HepEmMCCData material - cuts couples,
     * stored in the G4HepEmMatCutData::fMatCutData array:
     *   - for each material - cuts couple, there are \f$N := \f$ G4HepEmElectronData::fELossEnergyGridSize **range** values associated to the primary
     *     \f$e^-/e^+\f$ kinetic energy values stored in G4HepEmElectronData::fELossEnergyGrid. These
     *     \f$R_i, i=0,\ldots,N-1\f$ **range values are stored**
     *     with the corresponding \f$R_i^{''}, i=0,\ldots,N-1\f$
     *     **second derivatives** in the form of \f$R_0,R_0^{''},R_{1},R_{1}^{''},\ldots,R_{N-1},R_{N-1}^{''}\f$
     *     in order **to resonate the best with the run-time access pattern**.
     *   - **then the corresponding** \f$N\f$, \f$dE/dx\f$ **values are stored** in
     *     a similar way  \f$dE/dx_0,dE/dx_0^{''},dE/dx_{1},dE/dx_{1}^{''},\ldots,dE/dx_{N-1},dE/dx_{N-1}^{''}\f$
     *   - since both the range, and the kinetic energy values
     *     are already stored (see above), **only** the corresponding \f$N\f$
     *     **second derivative values associated to the inverse range are stored then** as
     *     \f$S_0,S_1,\ldots,S_{N-1}\f$
     *   - it means, that **there are** \f$5\times N\f$ **energy loss realted data values stored continuously in
     *     the** G4HepEmElectronData::fELossData **array for each material - cuts couples**. Therefore, in the case
     *     of a G4HepEmMCCData material - cuts couple data with the index of \f$\texttt{imc}\f$ (i.e. in the case of
     *     the G4HepEmMCCData, stored at G4HepEmMatCutData::fMatCutData[\f$\texttt{imc}\f$]), the start indices of
     *     the corresponding energy loss related data in the G4HepEmElectronData::fELossData array:
     *     - **range data** starts at the index of \f$\texttt{imc}\times(5\times N\f$)
     *     - **dE/dx data** starts at the index of \f$\texttt{imc}\times(5\times N\f$) + \f$2\times N\f$
     *     - **inverse range data** starts at the index of \f$\texttt{imc}\times(5\times N\f$) + \f$4\times N\f$
     *
     * The total number of data stored in the G4HepEmElectronData::fELossData array is
     * G4HepEmElectronData::fNumMatCuts\f$\times5\times\f$G4HepEmElectronData::fELossEnergyGridSize
     *
     * At run-time, for a given \f$E\f$ primary kinetic energy G4HepEmElectronData::fELossLogMinEkin
     * and G4HepEmElectronData::fELossEILDelta are used to compute the energy bin index \f$i\f$ such that
     * \f$ E_i \leq E < E_{i+1}, i=0,\ldots,N-1\f$. Then for the given material - cuts couple index,
     * the above starts indices can be used to access the corresponding energy loss data and second derivatives
     * associated to the primary kinetic energies of \f$ E_i, E_{i+1}\f$ needed to perform the spline interpolation.
     *
     *
     * @note
     * There is a spline interpolation function in G4HepEmRunUtils, specialised
     * for the above pattern used to store the dE/dx and range data. Using this
     * function **ensures optimal** data **cache utilisation at run-time**.
     * A separate, more traditional spline interpolation function is used for
     * run-time inverse range data interpolation. These are utilised in the
     * G4HepEmElectronManager to ensure the optimal run-time performance (both in
     * terms of memory consumption and speed) when accessing the restricted energy loss
     * related, i.e. stopping power, range and inverse range data in the \f$e^-/e^+\f$ stepping.
     */
   double*    fELossData = nullptr; // [5xfELossEnergyGridSize x fNumMatCuts]
 /// @} */ // end: eloss
   //
 
 //// === MACROSCOPIC CROSS SECTION DATA
   /**
    * @name Restricted macroscopic cross section related data members:
    * These members are used to store all restricted macroscopic cross section related data both for
    * **ionisation** and **bremsstrahlung** for all material - cuts couples.
    */
 ///@{
   /** Total number of restricted macroscopic cross sections realted data stored in the single G4HepEmElectronData::fResMacXSecData array.*/
   int        fResMacXSecNumData = 0;
   /** Start index of the macroscopic cross section data, for the material - cuts couple with the given index, in the G4HepEmElectronData::fResMacXSecData array.*/
   int*       fResMacXSecStartIndexPerMatCut = nullptr;  // [fNumMatCuts]
   /** The restricted macroscopic cross section data for **ionisation** and **bremsstrahlung** for all material - cuts couples.
    *
    * All the restricted macroscopic cross section data are stored continuously in this G4HepEmElectronData::fResMacXSecData single array.
    * The *restricted macroscopic cross sections* go to *zero at primary kinetic energies lower than or equal to the secondary
    * production threshold*: at the secondary \f$e^-\f$ production threshold in the case of \f$e^+\f$ and \f$2\times\f$ of it in the case
    * of \f$e^-\f$ ionisation, while the secondary \f$\gamma\f$ production threshold energy in case of bremsstrahlung. It means that,
    * the minimum value of *the primary kinetic energy grid depends on the type of the interaction* (ioni. or brem.) as well as
    * *the production cut values*. Therefore, an individual primary kinetic energy grid is generated, and stored together with the
    * corresponding restricted macroscopic cross section values, for each individual material - cuts couples, separately for
    * ionisation and for bremsstrahlung in each cases.
    *
    * The data are stored in the following format for each of the G4HepEmMCCData material - cuts couples (stored in the
    * G4HepEmMatCutData::fMatCutData array):
    *   - for a G4HepEmMCCData material - cuts couple data with the index of \f$\texttt{imc}\f$ (i.e.
    *     stored at G4HepEmMatCutData::fMatCutData[\f$\texttt{imc}\f$]), the macroscopic scross section realted
    *     data starts at G4HepEmElectronData::fResMacXSecData[\f$\texttt{ioniStarts}\f$], where \f$\texttt{ioniStarts}=\f$G4HepEmElectronData::fResMacXSecStartIndexPerMatCut [\f$\texttt{imc}\f$]
    *   - then relative to this \f$\texttt{ioniStarts}\f$ start index, **first** the restricted macroscopic cross section data for **ionisation**:
    *       - ``[0]``: \f$M:=M^{\text{(ioni)}\texttt{-imc}}\f$: **number of** \f$E_i, i=0,\ldots,M-1\f$ **primary kinetic energy points** over
    *         which the \f$\Sigma:=\Sigma^{\text{(ioni)}\texttt{-imc}}(E_i)\f$ restricted macroscopic cross section **for ionisation**
    *         is computed and stored **for this material - cuts couple** with the index of \f$\texttt{imc}\f$.
    *       - ``[1]``: \f$\texttt{argmax}\{\Sigma(E_i)\}, i=0,\ldots,M-1\f$
    *       - ``[2]``: \f$\texttt{max}\{\Sigma(E_i)\}, i=0,\ldots,M-1\f$
    *       - ``[3]``: \f$\log(E_0)\f$
    *       - ``[4]``: \f$1/[log(E_{M-1}/E_0)/(M-1)]\f$
    *       - ``[5 : 5 + 3xM-1]``: \f$E_0,\Sigma(E_0),\Sigma(E_0)^{''},E_1,\Sigma(E_1),\Sigma(E_1)^{''},\ldots,E_{M-1},\Sigma(E_{M-1}), \Sigma(E_{M-1})^{''}\f$
    *         where \f$^{''}\f$ denotes the second derivatives.
    *   - then continuously from the \f$\texttt{bremStarts} = \texttt{ioniStarts} + 3\times M+5 \f$ index,
    *     the restricted macroscopic cross section data for **bremsstrahlung**:
    *       - ``[0]``: \f$N:=N^{\text{(brem)}\texttt{-imc}}\f$: **number of** \f$E_i, i=0,\ldots,N-1\f$ **primary kinetic energy points** over
    *         which the \f$\Sigma:=\Sigma^{\text{(brem)}\texttt{-imc}}(E_i)\f$ restricted macroscopic cross section **for bremsstrahlung**
    *         is computed and stored **for this material - cuts couple** with the index of \f$\texttt{imc}\f$.
    *       - ``[1]``: \f$\texttt{argmax}\{\Sigma(E_i)\}, i=0,\ldots,N-1\f$
    *       - ``[2]``: \f$\texttt{max}\{\Sigma(E_i)\}, i=0,\ldots,N-1\f$
    *       - ``[3]``: \f$\log(E_0)\f$
    *       - ``[4]``: \f$1/[log(E_{N-1}/E_0)/(N-1)]\f$
    *       - ``[5 : 5 + 3xN-1]``: \f$E_0,\Sigma(E_0),\Sigma(E_0)^{''},E_1,\Sigma(E_1),\Sigma(E_1)^{''},\ldots,E_{N-1},\Sigma(E_{N-1}), \Sigma(E_{N-1})^{''}\f$
    *         where \f$^{''}\f$ denotes again the second derivatives.
    *
    * The total number of data, i.e. the length of the G4HepEmElectronData::fResMacXSecData array,
    * is stored in G4HepEmElectronData::fResMacXSecData.
    *
    * At run-time, for a given \f$E\f$ primary kinetic energy and material - cuts couple with the index of \f$\texttt{imc}\f$,
    *  - the start index of the **macroscopic cross section data for ionisation** is given by
    *    \f$\texttt{ioniStarts}\f$=G4HepEmElectronData::fResMacXSecStartIndexPerMatCut[\f$\texttt{imc}\f$]
    *  - then G4HepEmElectronData::fResMacXSecData[\f$\texttt{ioniStarts}\f$+3] and
    *    G4HepEmElectronData::fResMacXSecData[\f$\texttt{ioniStarts}\f$+4] can be used to compute the
    *    kinetic energy bin index \f$i\f$ such that \f$ E_i \leq E < E_{i+1}, i=0,\ldots,\f$G4HepEmElectronData::fResMacXSecData[\f$\texttt{ioniStarts}\f$]\f$-1\f$.
    *  - then the kinetic energies, macroscopic cross sections and their second derivatives,
    *    associated to the primary kinetic energies of \f$ E_i, E_{i+1}\f$ are used to perform the spline interpolation
    *  - the start index of the corresponding **macroscopic cross section data for bremsstrahlung** is given by
    *    \f$\texttt{bremStarts} = \texttt{ioniStarts} + 5 + 3\times\f$G4HepEmElectronData::fResMacXSecData[\f$\texttt{ioniStarts}\f$]
    *  - then the same procedure can be applied as above to compute the kinetic energy bin index and perform the interpolation,
    *    but now relative to \f$\texttt{bremStarts}\f$ instead of the above \f$\texttt{ioniStarts}\f$
    *
    * @note
    * Note, that all the 6 data, that are needed for the run-time interpolation of the restricted macroscopic scross sections
    * are stored next to each other in the memory for both interactions (ionisation and bremsstrahlung). Moreover, for a given
    * material - cuts couple, the data for the two interactions are stored one after the other. Together with the corresponding
    * special spline interpolation function of G4HepEmRunUtils, that ensures a maximal profit of this memory layout, it makes
    * **optimal utilisation of the** data **cache at run-time**. This special spline interpolation is utilised in the
    * G4HepEmElectronManager to ensure the optimal run-time performance (both in terms of memory consumption and speed) when
    * accessing the restricted macroscopic cross section data in the \f$e^-/e^+\f$ stepping.
    *
    */
   double*    fResMacXSecData = nullptr; // [fResMacXSecNumData]
 /// @} */ // end: restricted macroscopic cross section
 
   //
   // Electron - and positron - nuclear cross sections per material:
   // --- Grid: 127 bins form 100 MeV - 100 TeV
   const int     fENucEnergyGridSize = 128;
   double        fENucLogMinEkin = 0.0;    // = 4.605170185988092;  // log(100.0)
   double        fENucEILDelta = 0.0;      // = 9.192566533618830;  // 1./[log(emax/emin)/127]
   double*       fENucEnergyGrid = nullptr;    // [fENucEnergyGrid]
 
   double*       fENucMacXsecData = nullptr;   // [#materials*2*fENucEnergyGridSize]
 
 
   /**
    * @name Macroscopic first transport corss section related data members:
    * These members are used to store all macroscopic first transport cross section related data
    * for all materials. The discrete energy grid, above which that disceret cross section
    * values are computed and stored, is the same as used for the energy loss data.
    */
 ///@{
   /** The macroscopic first transport cross section.
    *
    * The data are stored in the following format for each of the G4HepEmMatData material
    * (stored in the G4HepEmMaterialData::fMaterialData array):
    *   - there are \f$N := \f$ G4HepEmElectronData::fELossEnergyGridSize **(macroscopic) first transport cross section
    *     values** associated to the primary \f$e^-/e^+\f$ kinetic energy values stored
    *     in G4HepEmElectronData::fELossEnergyGrid. These \f$TR1_i, i=0,\ldots,N-1\f$ **values** are stored
    *     with the corresponding \f$TR1_i^{''}, i=0,\ldots,N-1\f$
    *     **second derivatives** in the form of \f$TR1_0,TR1_0^{''},TR1_{1},TR1_{1}^{''},\ldots,TR1_{N-1},TR1_{N-1}^{''}\f$
    *     in order **to resonate the best with the run-time access pattern**.
    *   - for a G4HepEmMatData material data with the index of \f$\texttt{im}\f$ (i.e.
    *     stored at G4HepEmMaterialData::fMaterialData[\f$\texttt{im}\f$]), the macroscopic first
    *     transport cross section realted data starts at \f$\texttt{iStart} = 2\times\f$G4HepEmElectronData::fELossEnergyGridSize\f$\times\texttt{im}\f$
    *     in the G4HepEmElectronData::fTr1MacXSecData array
    *   - the total number of data in the array is \f$2\times\f$G4HepEmElectronData::fELossEnergyGridSize\f$\times\f$G4HepEmElectronData::fNumMaterials
    *
    * @note
    * There is a spline interpolation function in G4HepEmRunUtils, specialised
    * for the above pattern used to store the dE/dx and range data. Using this
    * function **ensures optimal** data **cache utilisation at run-time**.
    */
   double*    fTr1MacXSecData = nullptr; // [2xfELossEnergyGridSize x fNumMaterials]
 /// @} */ // end: macroscopic first transport cross section
 
 
 //// === TARGET ELEMENT SELECTOR
   /**
    * @name Target element selector related data members:
    * These members store data utilised at run-time for the selection of the target
    * elements (in case of multi element atoms) on which the interaction takes palce.
    * Data are stored for all the interaction models used to describe both **ionisation** and
    * **bremsstrahlung** and for all material - cuts couples.
    *
    * These data are the normalised, element-wise contributions to the corresonding
    * macroscopic cross sections in case of multi element materials. Therefore,
    * similarly to the above restricted macroscopic cross sections, different energy
    * grids are generated for the the different material - cuts couples. The data
    * are also stored in a separate, single continuous array per interaction model.
    * The total number of data, i.e. the size of this single array as well as the start indices
    * of the data, related to the different material - cuts couples, are stored for each
    * of the three interaction models.
    *
    * The data are stored in the following format for each of the individual interaction models,
    * for each the G4HepEmMCCData material - cuts couples (stored in the G4HepEmMatCutData::fMatCutData array):
    *   - for a G4HepEmMCCData material - cuts couple data with the index of \f$\texttt{imc}\f$ (i.e.
    *     stored at G4HepEmMatCutData::fMatCutData[\f$\texttt{imc}\f$]), the element selector
    *     data starts at the index \f$\texttt{iStarts}=\texttt{fElemSelectorXYStartIndexPerMatCut[imc]}\f$ where
    *     \f$XY\f$ is one of the three models, i.e. \f$\{\texttt{Ioni, BremSB, BremRB}\}\f$ for the
    *     *Moller-Bhabha ionisation*, *Seltzer-Berger* or the *relativistic bremsstrahlung models*.
    *   - if the material, associated to this material - cuts couple, is composed of a single element,
    *     \f$\texttt{iStarts}=-1\f$
    *   - the following data are stored otherwise continuously in the appropriate \f$\texttt{fElemSelectorXYData}\f$
    *     array relative to this \f$\texttt{iStarts}\f$ index
    *       - ``[0]``: \f$K:=\f$ *number of discrte* \f$E_i, i=0,\ldots,K-1\f$ *primary particle kinetic energy values* used to compute and store
    *           the \f$P(Z_j,E_i):=\Sigma^{Z_j}(E_i)/\Sigma(E_i)\f$ normalised, element-wise contributions to the macroscopic cross section of the material.
    *       - ``[1]``: \f$Q:=\f$ *number of elements the given material is composed of*. So above, \f$Z_j, j=0,\ldots,Q-1\f$ at each individual
    *           \f$E_i\f$ kinetic energy values. However, since \f$P(Z_{j=Q-1},E_i)=\Sigma^{Z_{Q-1}}(E_i)/\Sigma(E_i) = 1\f$ for all \f$i=0,\ldots,K-1\f$
    *           due to the normalisation, data are computed and stored only for element indices of \f$j=0,\ldots,Q-2\f$.
    *       - ``[2]``: \f$\log(E_0)\f$
    *       - ``[3]``: \f$1/[log(E_{K-1}/E_0)/(K-1)]\f$
    *       - ``[4 : 4 + QxK-1]``: \f$E_0,P(j=0,E_0),P(j=1,E_0),\ldots,P(j=Q-2,E_0), \ldots,\f$ \f$E_{K-1},P(j=0,E_{K-1}),P(j=1,E_{K-1}),\ldots,P(j=Q-2,E_{K-1})\f$
    *
    * At run-time, when performing an interaction described by model \f$XY \in \{\texttt{Ioni, BremSB, BremRB}\}\f$,
    * with primary particle kinetic energy of \f$E\f$ in the material, related to the material - cuts couple with the index of \f$\texttt{imc}\f$,
    *  - the **start index of the** corresponding element selector **data** is \f$\texttt{iStarts}=\texttt{fElemSelectorXYStartIndexPerMatCut[imc]}\f$
    *  - the the corresonding \f$\texttt{fElemSelectorXYData[iStart+2]}\f$ and \f$\texttt{fElemSelectorXYData[iStart+3]}\f$ values can be used to compute the
    *    kinetic energy bin index \f$i\f$ such that \f$ E_i \leq E < E_{i+1}, i=0,\ldots,\texttt{fElemSelectorXYData[iStart]}-1\f$.
    *  - then the kinetic energies and normalised element-wise partial macroscopic cross sections,
    *    associated to the primary kinetic energies of \f$ E_i, E_{i+1}\f$ are used to perform the linear interpolation (smooth function) and
    *    and to sample the target element index from this discrete distribution.
    *
    * @note
    * Note, that all the data, that are needed for the run-time interpolation and for the target element index sampling are stored next to each other in the memory.
    * The implementations of the individual interaction models, that utilise these data for the run-time target atom selection (if needed), make sure that this
    * memory layout is maximally exploited. These ensure the optimal performance, both in terms of memory consumption and speed, when
    * accessing these data performing the correspondign \f$e^-/e^+\f$ interactions.
    */
 ///@{
   /** Total number of element selector data for the Moller-Bhabha model for e-/e+ ionisation.*/
   int       fElemSelectorIoniNumData = 0;
   /** Indices, at which data starts for a given material - cuts couple.*/
   int*      fElemSelectorIoniStartIndexPerMatCut = nullptr;     // [fNumMatCuts]
   /** Element selector data for all material - cuts couples with multiple element material.*/
   double*   fElemSelectorIoniData = nullptr;                    // [fElemSelectorIoniNumData]
 
   /** Total number of element selector data for the Seltzer-Berger model for e-/e+ bremsstrahlung.*/
   int       fElemSelectorBremSBNumData = 0;
   /** Indices, at which data starts for a given material - cuts couple.*/
   int*      fElemSelectorBremSBStartIndexPerMatCut = nullptr;   // [fNumMatCuts]
   /** Element selector data for all material - cuts couples with multiple element material.*/
   double*   fElemSelectorBremSBData = nullptr;                  // [fElemSelectorBremSBNumData]
 
   /** Total number of element selector data for the relativistic (improved Bethe-Heitler) model for e-/e+ bremsstrahlung.*/
   int       fElemSelectorBremRBNumData = 0;
   /** Indices, at which data starts for a given material - cuts couple.*/
   int*      fElemSelectorBremRBStartIndexPerMatCut = nullptr;   // [fNumMatCuts]
   /** Element selector data for all material - cuts couples with multiple element material.*/
   double*   fElemSelectorBremRBData = nullptr;                  // [fElemSelectorBremRBNumData]
 /// @} */ // end: target element selectors
 };
 
 
 /**
   * Allocates and pre-initialises the G4HepEmElectronData structure.
   *
   * This method is invoked from the InitElectronData() function declared
   * in the G4HepEmElectronInit header file. The input argument address of the
   * G4HepEmElectronData structure pointer is the one stored in the G4HepEmData
   * member of the `master` G4HepEmRunManager and the initialisation should be
   * done by the master G4HepEmRunManager by invoking the InitElectronData() function
   * for \f$e^-/e^+\f$ particles.
   *
   * @param theElectronData address of a G4HepEmElectronData structure pointer. At termination,
   *   the correspondig pointer will be set to a memory location with a freshly allocated
   *   G4HepEmElectronData structure with all its pointer members set to nullprt.
   *   If the input pointer was not null at input, the pointed memory, including all
   *   dynamic memory members, is freed before the new allocation.
   */
 void AllocateElectronData (struct G4HepEmElectronData** theElectronData);
 
 /**
  * Initializes a new @ref G4HepEmElectronData structure
  *
  * This function default constructs an instance of G4HepEmElectronData and returns
  * a pointer to the freshly constructed instance. It is the callees responsibility
  * to free the instance using @ref FreeElectronData.
  *
  * @return Pointer to instance of @ref G4HepEmElectronData
  */
 G4HepEmElectronData* MakeElectronData();
 
 /**
   * Frees a G4HepEmElectronData structure.
   *
   * This function deallocates all dynamically allocated memory stored in the
   * input argument related G4HepEmElectronData structure, deallocates the structure
   * itself and sets the input address to store a pointer to null. This makes the
   * corresponding input stucture cleared, freed and ready to be re-initialised.
   * The input argument is supposed to be the address of the corresponding pointer
   * member of the G4HepEmData member of the `master` G4HepEmRunManager.
   *
   * @param theElectronData memory address that stores pointer to a G4HepEmElectronData
   *  structure. The memory is freed and the input address will store a null pointer
   *  at termination.
   */
 void FreeElectronData (struct G4HepEmElectronData** theElectronData);
 
 
 #ifdef G4HepEm_CUDA_BUILD
   /**
     * Allocates memory for and copies the G4HepEmElectronData structure from the
     * host to the device.
     *
     * The input arguments are supposed to be the corresponding members of the
     * G4HepEmData, top level data structure, stored in the `master` G4HepEmRunManager.
     *
     * @param onHOST    pointer to the host side, already initialised G4HepEmElectronData structure.
     * @param onDEVICE  host side address of a pointer to a device side G4HepEmElectronData
     *   structure. The pointed device side memory is cleaned (if not null at input) and
     *   points to the device side memory at termination containing all the copied
     *   G4HepEmElectronData structure members.
     */
   void CopyElectronDataToDevice(struct G4HepEmElectronData* onHOST, struct G4HepEmElectronData** onDEVICE);
 
   /**
     * Frees all memory related to the device side G4HepEmElectronData structure referred
     * by the pointer stored on the host side input argument address.
     *
     * @param onDEVICE host side address of a G4HepEmElectronDataOnDevice structure located on the device side memory.
     *   The correspondig device memory will be freed and the input argument address will be set to null.
     */
   void FreeElectronDataOnDevice(struct G4HepEmElectronData** onDEVICE);
 #endif // DG4HepEm_CUDA_BUILD
 
 #endif // G4HepEmElementData_HH

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