STARS-H
Software for Testing Accuracy, Reliability and Scalability of Hierarchical computations
starsh_cluster Struct Reference

Info about clusterization of physical data. More...

#include <starsh.h>

Data Fields

void * data
 Pointer to structure, holding physical data.
 
STARSH_int ndata
 Number of discrete elements, corresponding to physical data. More...
 
STARSH_intpivot
 Pivoting of discrete elements for clusterization. More...
 
STARSH_int nblocks
 Total number of subclusters (blocks) of discrete elements.
 
STARSH_int nlevels
 Number of levels of hierarchy. More...
 
STARSH_intlevel
 Index of first cluster for each level of hierarchy. More...
 
STARSH_intstart
 Index of first pivoted discrete element of a cluster. More...
 
STARSH_intsize
 Number of discrete elements in each cluster.
 
STARSH_intparent
 Parent cluster for each subcluster. More...
 
STARSH_intchild_start
 Start index of child for each cluster. More...
 
STARSH_intchild
 Children clusters of each cluster. More...
 
enum STARSH_CLUSTER_TYPE type
 Type of cluster (tiled or hierarchical).
 

Detailed Description

Info about clusterization of physical data.

Field Documentation

◆ child

STARSH_int* starsh_cluster::child

Children clusters of each cluster.

Clusters with indexes from child[child_start[i]] to child[child_start[i+1]]-1inclusively are children for a cluster i. In case of tiled clusterization child is NULL.

◆ child_start

STARSH_int* starsh_cluster::child_start

Start index of child for each cluster.

Clusters with indexes from child[child_start[i]] to child[child_start[i+1]]-1inclusively are children for a cluster i. In case of tiled clusterization child_start is NULL.

◆ level

STARSH_int* starsh_cluster::level

Index of first cluster for each level of hierarchy.

All subclusters, corresponding to given level of hierarchy are stored in compressed format: subclusters with indexes from level[i] to level[i+1]-1 inclusively correspond to i-th level of hierarchy. Field level is an array with nlevels+1 elements in hierarchical case and is NULL in tiled case.

◆ ndata

STARSH_int starsh_cluster::ndata

Number of discrete elements, corresponding to physical data.

Discrete element can be anything, i.e. particles, grid nodes, edges or mesh elements.

◆ nlevels

STARSH_int starsh_cluster::nlevels

Number of levels of hierarchy.

0 in case of tiled clusterization.

◆ parent

STARSH_int* starsh_cluster::parent

Parent cluster for each subcluster.

Root node is 0 node and it has no parent, so parent[0] = -1.

◆ pivot

STARSH_int* starsh_cluster::pivot

Pivoting of discrete elements for clusterization.

After pivoting discrete elements of a single cluster are one after another. Used by field start.

◆ start

STARSH_int* starsh_cluster::start

Index of first pivoted discrete element of a cluster.

Indexes of discrete elements, corresponding to a single cluster, are located one after another in array pivot.


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