#include <teca_space_time_executive.h>

Inheritance diagram for teca_space_time_executive:

Collaboration diagram for teca_space_time_executive:

Public Member Functions
std::shared_ptr< teca_space_time_executive >	shared_from_this ()

std::shared_ptr< teca_space_time_executive const >	shared_from_this () const

int	initialize (MPI_Comm comm, const teca_metadata &md) override

teca_metadata	get_next_request () override

void	set_number_of_spatial_partitions (long n_partitions)

void	set_number_of_temporal_partitions (long n_partitions)

void	set_temporal_partition_size (long n_steps)

void	set_time_step (long s)
	Set the time step to process. More...

void	set_first_step (long s)
	Set the first index in the series to process.The default is 0. More...

void	set_last_step (long s)

void	set_extent (unsigned long *ext)

void	set_extent (const std::vector< unsigned long > &ext)

void	set_bounds (double *bounds)

void	set_bounds (const std::vector< double > &bounds)

void	set_arrays (const std::vector< std::string > &arrays)

void	enable_index_executive_compatibility ()

void	set_partition_x (int val)
	A flag that enables/disables partitioning in the x-direction. More...

void	set_partition_y (int val)
	A flag that enables/disables partitioning in the y-direction. More...

void	set_partition_z (int val)
	A flag that enables/disables partitioning in the z-direction. More...

void	set_minimum_block_size_x (unsigned long val)
	Sets the minimum block size in the x-direction. More...

void	set_minimum_block_size_y (unsigned long val)
	Sets the minimum block size in the y-direction. More...

void	set_minimum_block_size_z (unsigned long val)
	Sets the minimum block size in the z-direction. More...

Public Member Functions inherited from teca_algorithm_executive
void	set_verbose (int a_verbose)

int	get_verbose () const

Static Public Member Functions
static p_teca_space_time_executive	New ()

Static Public Member Functions inherited from teca_algorithm_executive
static p_teca_algorithm_executive	New ()

Additional Inherited Members
Protected Member Functions inherited from teca_algorithm_executive
	teca_algorithm_executive (const teca_algorithm_executive &)=default

	teca_algorithm_executive (teca_algorithm_executive &&)=default

teca_algorithm_executive &	operator= (const teca_algorithm_executive &)=default

teca_algorithm_executive &	operator= (teca_algorithm_executive &&)=default

Detailed Description

An executive that generates requests such that the upstream or user defined index space is partitoned using a fixed partition size and spatially across MPI ranks. Set the partition size to 0 to use only the spatial partitioning.

This executive uses the following meta data keys:

| Key  | Description |
| ---- | ----------- |
| whole_extent | The spatial extent to partiton. Partitionaing is |
| ^            | applied such that each MPI rank has a unique subset to |
| ^            | process |
| index_initializer_key -| holds the name of the key that tells how |
| ^                      | many indices are available. the named key |
| ^                      | must also be present and should conatin the |
| ^                      | number of indices available |
| index_request_key | holds the name of the key used to request |
| ^                 | a specific inclusive range of indices [i0, i1]. |
| ^                 | Requests are generated with this name set to |
| ^                 | a specific inclusive range of indices to be |
| ^                 | processed. Some upstream algorithm is |
| ^                 | expected to produce the data associated |
| ^                 | with the requested index extent. |

Member Function Documentation

◆ enable_index_executive_compatibility()

void teca_space_time_executive::enable_index_executive_compatibility ( )

inline

Condigures the executive to make one request per time step using the index_request_key as the teca_index_executive would. This could be used parallelize exsiting algorithms over space and time.

◆ New()

static p_teca_space_time_executive teca_space_time_executive::New ( )

inlinestatic

Allocate a new teca_space_time_executive

◆ set_arrays()

void teca_space_time_executive::set_arrays ( const std::vector< std::string > & arrays )

Set a list of arrays to include in upstream requests. This provides a means of pulling data through the pipeline when the upstream algorithms do not do this automatically or to pull additional arrays.

◆ set_bounds()

void teca_space_time_executive::set_bounds ( double * bounds )

Set the spatial bounds to partition and process. The default is to partition and process the entire domain.

◆ set_extent()

void teca_space_time_executive::set_extent ( unsigned long * ext )

Set the extent to partition and process. The default is to partition and process the whole_extent.

◆ set_first_step()

void teca_space_time_executive::set_first_step ( long s )

Set the first index in the series to process.The default is 0.

◆ set_last_step()

void teca_space_time_executive::set_last_step ( long s )

set the last index in the series to process. default is -1. negative number results in the last available time step being used.

◆ set_minimum_block_size_x()

void teca_space_time_executive::set_minimum_block_size_x ( unsigned long val )

inline

Sets the minimum block size in the x-direction.

◆ set_minimum_block_size_y()

void teca_space_time_executive::set_minimum_block_size_y ( unsigned long val )

inline

Sets the minimum block size in the y-direction.

◆ set_minimum_block_size_z()

void teca_space_time_executive::set_minimum_block_size_z ( unsigned long val )

inline

Sets the minimum block size in the z-direction.

◆ set_number_of_spatial_partitions()

void teca_space_time_executive::set_number_of_spatial_partitions ( long n_partitions )

Set the number of partitions to partition the spatial extent into. If set to a value less than one, the number of MPI ranks is used. If set to a number greater or equal to one and less than the number of MPI ranks, a warning is issued but processing will continue. The default is zero, resulting in one partition per MPI rank.

◆ set_number_of_temporal_partitions()

void teca_space_time_executive::set_number_of_temporal_partitions ( long n_partitions )

Set the number of partitions to partition the temporal extent into. Use set_time_partition_size to control the partition size rather then the number.

◆ set_partition_x()

void teca_space_time_executive::set_partition_x ( int val )

inline

A flag that enables/disables partitioning in the x-direction.

◆ set_partition_y()

void teca_space_time_executive::set_partition_y ( int val )

inline

A flag that enables/disables partitioning in the y-direction.

◆ set_partition_z()

void teca_space_time_executive::set_partition_z ( int val )

inline

A flag that enables/disables partitioning in the z-direction.

◆ set_temporal_partition_size()

void teca_space_time_executive::set_temporal_partition_size ( long n_steps )

Set the partition size to use when partitioning the temporal extent. Setting to less than one results in a spatial only partitioning. When greater or equal to one the value will be used to determine the number of temporal partitions.

When the total number of time steps is not evenly devided by the temporal partition size, the final request will specify the remainder of steps.

Parameters

[in] n_steps the number of indices to request per request.

◆ set_time_step()

void teca_space_time_executive::set_time_step ( long s )

Set the time step to process.

The documentation for this class was generated from the following file:

develop/alg/teca_space_time_executive.h

Public Member Functions

Static Public Member Functions

Additional Inherited Members

Detailed Description

Member Function Documentation

◆ enable_index_executive_compatibility()

◆ New()

◆ set_arrays()

◆ set_bounds()

◆ set_extent()

◆ set_first_step()

◆ set_last_step()

◆ set_minimum_block_size_x()

◆ set_minimum_block_size_y()

◆ set_minimum_block_size_z()

◆ set_number_of_spatial_partitions()

◆ set_number_of_temporal_partitions()

◆ set_partition_x()

◆ set_partition_y()

◆ set_partition_z()

◆ set_temporal_partition_size()

◆ set_time_step()