Superdrops Observer

Header file: <libs/observers/superdrops_observer.hpp> [source]

template<typename Store, typename T, typename FunctorFunc>
CollectDataForDataset<Store> auto CollectSuperdropVariable(const Dataset<Store> &dataset, const FunctorFunc ffunc, const std::string_view name, const std::string_view units, const double scale_factor, const size_t maxchunk)

Constructs type sastifying the CollectDataForDataset concept for a given Store (using an instance of the GenericCollectData class) which writes a superdroplet variable to a ragged Xarray in a dataset.

Function return type writes a superdroplet varaible “name” to a ragged Xarray for a data type by collecting data according to the given FunctorFunc from within a Kokkos::parallel_for loop over superdroplets with a range policy.

Template Parameters:

  • Store – The type of the dataset store.

  • T – The type of the superdroplet variable data.

  • FunctorFunc – The type of the functor function.

Parameters:

  • dataset – The dataset to collect data from.

  • ffunc – The functor function to apply to collect data.

  • name – The name of the variable.

  • units – The units of the variable.

  • scale_factor – The scale factor of the variable.

  • maxchunk – The maximum chunk size.

Returns:

CollectDataForDataset<Store> An instance of CollectDataForDataset for collecting superdroplet variable data.

struct SdgbxindexFunc

Functor operator to perform a copy of the sdgbxindex of each superdroplet to d_data within Kokkos::parallel_for loop over superdroplets with a range policy.

Signature of operator such that type can be used by GenericCollectData struct for FunctorFunc.

Note: Conversion of sdgbxindex from unsigned long int (8 bytes) to unsigned int (uint32_t = 4 bytes)

Param kk:

The index of the superdrop.

Param d_gbxs:

The view of gridboxes on device.

Param totsupers:

The view of superdroplets on device.

Param d_data:

The mirror view buffer for the superdroplets’ sdgbxindex.

Public Functions

inline void operator()(const size_t kk, viewd_constgbx d_gbxs, const viewd_constsupers totsupers, Buffer<uint32_t>::mirrorviewd_buffer d_data) const
template<typename Store>
CollectDataForDataset<Store> auto CollectSdgbxindex(const Dataset<Store> &dataset, const size_t maxchunk)

Constructs a type satisyfing the CollectDataForDataset concept for each superdroplets’ gridbox index data.

This function constructs a type satisfying the CollectDataForDataset to collect each superdroplet’s gridbox index “sdgbxindex” as a 4 byte unsigned integer and write it to a ragged array in a dataset.

Template Parameters:

Store – The type of the dataset store.

Parameters:

  • dataset – The dataset to collect data from.

  • maxchunk – The maximum chunk size (number of elements).

Returns:

CollectDataForDataset<Store> An instance of CollectDataForDataset for collecting sdgbxindex data.

struct SdIdFunc

Functor operator to perform a copy of the identity of each superdroplet to d_data within Kokkos::parallel_for loop over superdroplets with a range policy.

Signature of operator such that type can be used by GenericCollectData struct for FunctorFunc.

Note: Conversion of sdid from unsigned long int (8 bytes) to unsigned int (uint32_t = 4 bytes)

Param kk:

The index of the superdrop.

Param d_gbxs:

The view of gridboxes on device.

Param totsupers:

The view of superdroplets on device.

Param d_data:

The mirror view buffer for the superdroplets’ identity.

Public Functions

inline void operator()(const size_t kk, viewd_constgbx d_gbxs, const viewd_constsupers totsupers, Buffer<uint32_t>::mirrorviewd_buffer d_data) const
template<typename Store>
CollectDataForDataset<Store> auto CollectSdId(const Dataset<Store> &dataset, const size_t maxchunk)

Constructs a type satisyfing the CollectDataForDataset conceptfor each superdroplets’ identity.

This function constructs a type satisfying the CollectDataForDataset to collect each superdroplet’s identity “sdId” as a 4 byte unsigned integer and write it to a ragged array in a dataset.

Template Parameters:

Store – The type of the dataset store.

Parameters:

  • dataset – The dataset to collect data from.

  • maxchunk – The maximum chunk size (number of elements).

Returns:

CollectDataForDataset<Store> An instance of CollectDataForDataset for collecting sdId data.

struct XiFunc

Functor operator to perform a copy of the multiplicity of each superdroplet to d_data within Kokkos::parallel_for loop over superdroplets with a range policy.

Signature of operator such that type can be used by GenericCollectData struct for FunctorFunc.

Note: Conversion of xi from size_t (architecture dependent usually 8 bytes) to 8 byte, long unsigned int (unit64_t).

Param kk:

The index of the superdrop.

Param d_gbxs:

The view of gridboxes on device.

Param totsupers:

The view of superdroplets on device.

Param d_data:

The mirror view buffer for the superdroplets’ multiplicity.

Public Functions

inline void operator()(const size_t kk, viewd_constgbx d_gbxs, const viewd_constsupers totsupers, Buffer<uint64_t>::mirrorviewd_buffer d_data) const
template<typename Store>
CollectDataForDataset<Store> auto CollectXi(const Dataset<Store> &dataset, const size_t maxchunk)

Constructs a type satisyfing the CollectDataForDataset concept for each superdroplets’ multiplicity.

This function constructs a type satisfying the CollectDataForDataset to collect each superdroplet’s multiplicity “xi” as a 8 byte unsigned integer and write it to a ragged array in a dataset.

Template Parameters:

Store – The type of the dataset store.

Parameters:

  • dataset – The dataset to collect data from.

  • maxchunk – The maximum chunk size (number of elements).

Returns:

CollectDataForDataset<Store> An instance of CollectDataForDataset for collecting xi data.

struct RadiusFunc

Functor operator to perform a copy of the radius of each superdroplet to d_data within Kokkos::parallel_for loop over superdroplets with a range policy.

Signature of operator such that type can be used by GenericCollectData struct for FunctorFunc.

Note: Conversion of radius from double (8 bytes) to float (4 bytes).

Param kk:

The index of the superdrop.

Param d_gbxs:

The view of gridboxes on device.

Param totsupers:

The view of superdroplets on device.

Param d_data:

The mirror view buffer for the superdroplets’ radius.

Public Functions

inline void operator()(const size_t kk, viewd_constgbx d_gbxs, const viewd_constsupers totsupers, Buffer<float>::mirrorviewd_buffer d_data) const
template<typename Store>
CollectDataForDataset<Store> auto CollectRadius(const Dataset<Store> &dataset, const size_t maxchunk)

Constructs a type satisyfing the CollectDataForDataset concept or each superdroplets’ radius.

This function constructs a type satisfying the CollectDataForDataset to collect each superdroplet’s radius as a 4 byte floating point value and write it to a ragged array in a dataset.

Template Parameters:

Store – The type of the dataset store.

Parameters:

  • dataset – The dataset to collect data from.

  • maxchunk – The maximum chunk size (number of elements).

Returns:

CollectDataForDataset<Store> An instance of CollectDataForDataset for collecting radius.

struct MsolFunc

Functor operator to perform a copy of the solute mass of each superdroplet to d_data within Kokkos::parallel_for loop over superdroplets with a range policy.

Signature of operator such that type can be used by GenericCollectData struct for FunctorFunc.

Note: Conversion of msol from double (8 bytes) to float (4 bytes).

Param kk:

The index of the superdrop.

Param d_gbxs:

The view of gridboxes on device.

Param totsupers:

The view of superdroplets on device.

Param d_data:

The mirror view buffer for the superdroplets’ msol.

Public Functions

inline void operator()(const size_t kk, viewd_constgbx d_gbxs, const viewd_constsupers totsupers, Buffer<float>::mirrorviewd_buffer d_data) const
template<typename Store>
CollectDataForDataset<Store> auto CollectMsol(const Dataset<Store> &dataset, const size_t maxchunk)

Constructs a type satisyfing the CollectDataForDataset concept or each superdroplets’ solute mass.

This function constructs a type satisfying the CollectDataForDataset to collect each superdroplet’s solute mass “msol” as a 4 byte floating point value and write it to a ragged array in a dataset.

Template Parameters:

Store – The type of the dataset store.

Parameters:

  • dataset – The dataset to collect data from.

  • maxchunk – The maximum chunk size (number of elements).

Returns:

CollectDataForDataset<Store> An instance of CollectDataForDataset for collecting msol.

struct Coord3Func

Functor operator to perform a copy of the coord3 of each superdroplet to d_data within Kokkos::parallel_for loop over superdroplets with a range policy.

Signature of operator such that type can be used by GenericCollectData struct for FunctorFunc.

Note: Conversion of coord3 from double (8 bytes) to float (4 bytes).

Param kk:

The index of the superdrop.

Param d_gbxs:

The view of gridboxes on device.

Param totsupers:

The view of superdroplets on device.

Param d_data:

The mirror view buffer for the superdroplets’ coord3.

Public Functions

inline void operator()(const size_t kk, viewd_constgbx d_gbxs, const viewd_constsupers totsupers, Buffer<float>::mirrorviewd_buffer d_data) const
template<typename Store>
CollectDataForDataset<Store> auto CollectCoord3(const Dataset<Store> &dataset, const size_t maxchunk)

Constructs a type satisyfing the CollectDataForDataset concept or each superdroplets’ coord3.

This function constructs a type satisfying the CollectDataForDataset to collect each superdroplet’s coord3 as a 4 byte floating point value and write it to a ragged array in a dataset.

Template Parameters:

Store – The type of the dataset store.

Parameters:

  • dataset – The dataset to collect data from.

  • maxchunk – The maximum chunk size (number of elements).

Returns:

CollectDataForDataset<Store> An instance of CollectDataForDataset for collecting coord3.

struct Coord1Func

Functor operator to perform a copy of the coord1 of each superdroplet to d_data within Kokkos::parallel_for loop over superdroplets with a range policy.

Signature of operator such that type can be used by GenericCollectData struct for FunctorFunc.

Note: Conversion of coord1 from double (8 bytes) to float (4 bytes).

Param kk:

The index of the superdrop.

Param d_gbxs:

The view of gridboxes on device.

Param totsupers:

The view of superdroplets on device.

Param d_data:

The mirror view buffer for the superdroplets’ coord1.

Public Functions

inline void operator()(const size_t kk, viewd_constgbx d_gbxs, const viewd_constsupers totsupers, Buffer<float>::mirrorviewd_buffer d_data) const
template<typename Store>
CollectDataForDataset<Store> auto CollectCoord1(const Dataset<Store> &dataset, const size_t maxchunk)

Constructs a type satisyfing the CollectDataForDataset concept or each superdroplets’ coord1.

This function constructs a type satisfying the CollectDataForDataset to collect each superdroplet’s coord1 as a 4 byte floating point value and write it to a ragged array in a dataset.

Template Parameters:

Store – The type of the dataset store.

Parameters:

  • dataset – The dataset to collect data from.

  • maxchunk – The maximum chunk size (number of elements).

Returns:

CollectDataForDataset<Store> An instance of CollectDataForDataset for collecting coord1.

struct Coord2Func

Functor operator to perform a copy of the coord2 of each superdroplet to d_data within Kokkos::parallel_for loop over superdroplets with a range policy.

Signature of operator such that type can be used by GenericCollectData struct for FunctorFunc.

Note: Conversion of coord2 from double (8 bytes) to float (4 bytes).

Param kk:

The index of the superdrop.

Param d_gbxs:

The view of gridboxes on device.

Param totsupers:

The view of superdroplets on device.

Param d_data:

The mirror view buffer for the superdroplets’ coord2.

Public Functions

inline void operator()(const size_t kk, viewd_constgbx d_gbxs, const viewd_constsupers totsupers, Buffer<float>::mirrorviewd_buffer d_data) const
template<typename Store>
CollectDataForDataset<Store> auto CollectCoord2(const Dataset<Store> &dataset, const size_t maxchunk)

Constructs a type satisyfing the CollectDataForDataset concept or each superdroplets’ coord2.

This function constructs a type satisfying the CollectDataForDataset to collect each superdroplet’s coord2 as a 4 byte floating point value and write it to a ragged array in a dataset.

Template Parameters:

Store – The type of the dataset store.

Parameters:

  • dataset – The dataset to collect data from.

  • maxchunk – The maximum chunk size (number of elements).

Returns:

CollectDataForDataset<Store> An instance of CollectDataForDataset for collecting coord2.

template<typename Store>
struct RaggedCount

A struct for collecting ragged count data.

This struct is responsible for collecting ragged count data, which represents the count of the number of super-droplets written during a write of a ragged array of superdroplet data.

Template Parameters:

Store – The type of the dataset store.

Public Functions

inline RaggedCount(const Dataset<Store> &dataset, const size_t maxchunk)

Constructs a RaggedCount object.

Constructs a RaggedCount object with the specified dataset and maximum chunk size.

Parameters:

  • dataset – The dataset to collect data from.

  • maxchunk – The maximum chunk size (number of elements).

inline void write_to_array(const Dataset<Store> &dataset, const viewd_constsupers totsupers) const

Writes the total number of super-droplets to the ragged count array in the dataset.

Note: static conversion from architecture dependent, usually 16 byte unsigned integer (size_t = uint64_t), to 8 byte unsigned integer (uint32_t).

Parameters:

  • dataset – The dataset to write data to.

  • totsupers – The view of total super-droplets.

inline void write_arrayshape(const Dataset<Store> &dataset) const

Writes the shape of the ragged count array to the dataset.

This function writes the shape of the ragged count array to the dataset.

Parameters:

dataset – The dataset to write data to.

Private Members

std::shared_ptr<XarrayZarrArray<Store, uint32_t>> xzarr_ptr

pointer to raggedcount Xarray

template<typename Store>
inline Observer auto SuperdropsObserver(const unsigned int interval, const Dataset<Store> &dataset, const size_t maxchunk, CollectDataForDataset<Store> auto collect_data)

Constructs an observer which writes superdroplet variables (e.g. their attributes) from each superdroplet at start of each observation timestep to a ragged arrays with a constant observation timestep “interval”.

Template Parameters:

Store – Type of store for dataset.

Parameters:

  • interval – Observation timestep.

  • dataset – Dataset to write time data to.

  • maxchunk – Maximum number of elements in a chunk (1-D vector size).

  • collect_data – Object satisfying CollectDataForDataset for given Store to write superdroplet data, such as their attributes, to ragged arrays.

Returns:

Observer An observer instance for writing thermodynamic variables from each gridbox.