The structure contains information about each bucket. More...

#include <bucket.h>

Data Fields
int	stride
	Bucket stride. More...

int	min_size
	Minimum number of global indices across all buckets. More...

int	min_size_stride
	Minimum number of global indices across all buckets for a single stride. More...

int	size
	Number of global indices in the bucket. More...

int	size_stride
	Number of global indices in the bucket for a single stride. More...

int	max_size
	Maximum number of global indices across all buckets. More...

int	max_size_stride
	Maximum number of global indices across all buckets for a single stride. More...

int *	idxlist
	Array of indices inside the bucket sorted in ascending order of the MPI processes owning them. Size is sizes. More...

int *	ranks
	Array of MPI ranks owning the indices in the idxlist array Size is size. More...

int	count_recv
	Number of MPI processes that the bucket receives data from. More...

int *	src_recv
	Array of MPI processes sending data to the bucket. Size is count_recv. More...

int *	msg_size_recv
	Array of number of indices that each MPI ranks in src_recv send to the bucket. Size is count_recv. More...

int *	size_ranks
	Array of number of indices that the MPI process send to each bucket. Size is the number of MPI processes in the communicator. More...

int *	rank_exch
	Array of MPI ranks to send / recv data to / from. This is the result of the distributed directory algorithm. Each element of the array is associated with an element of the idxlist in the same order that the list was given to new_idxlist function. Size is the same of the idxlist->count. More...

Detailed Description

The structure contains information about each bucket.

Field Documentation

◆ count_recv

int t_bucket::count_recv

Number of MPI processes that the bucket receives data from.

◆ idxlist

int* t_bucket::idxlist

Array of indices inside the bucket sorted in ascending order of the MPI processes owning them. Size is sizes.

◆ max_size

int t_bucket::max_size

Maximum number of global indices across all buckets.

◆ max_size_stride

int t_bucket::max_size_stride

Maximum number of global indices across all buckets for a single stride.

◆ min_size

int t_bucket::min_size

Minimum number of global indices across all buckets.

◆ min_size_stride

int t_bucket::min_size_stride

Minimum number of global indices across all buckets for a single stride.

◆ msg_size_recv

int* t_bucket::msg_size_recv

Array of number of indices that each MPI ranks in src_recv send to the bucket. Size is count_recv.

◆ rank_exch

int* t_bucket::rank_exch

Array of MPI ranks to send / recv data to / from. This is the result of the distributed directory algorithm. Each element of the array is associated with an element of the idxlist in the same order that the list was given to new_idxlist function. Size is the same of the idxlist->count.

◆ ranks

int* t_bucket::ranks

Array of MPI ranks owning the indices in the idxlist array Size is size.

◆ size

int t_bucket::size

Number of global indices in the bucket.

◆ size_ranks

int* t_bucket::size_ranks

Array of number of indices that the MPI process send to each bucket. Size is the number of MPI processes in the communicator.

◆ size_stride

int t_bucket::size_stride

Number of global indices in the bucket for a single stride.

◆ src_recv

int* t_bucket::src_recv

Array of MPI processes sending data to the bucket. Size is count_recv.

◆ stride

int t_bucket::stride

Bucket stride.

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

/home/runner/work/DistDir/DistDir/src/core/algorithm/bucket.h

Data Fields