hyperslab_by_chunk

This program writes dataset sing chunking. Each process writes exactly one chunk.

Last installing: sudo apt-get install git python3-h5py python3-ipython python3-pint python3-sphinx python3-matplotlib python3-dev python3-numpy build-essential gcc libhdf5-openmpi-dev

Building:

cd hyperslab_by_chunk

sudo cmake .

mpic++ main.cpp -o hyperslab_by_chunk

Usage: mpirun -np [amount of threads+1] hyperslab_by_chunk [OPTIONS]...

Options:

--help (-h ) Print help

--version (-V ) Print program version and exit

--verbose (-v <verbosity) Be more verbose. Can be used multiple times, default 0

--nx (-x ) Dimension of all file by X-axis, default 8

--ny (-y ) Dimension of all file by Y-axis, default 4

--nz (-z ) Dimension of all file by Z-axis, default 4

--ch_nx (-i ) Dimension of one chunk by X-axis, default 4

--ch_ny (-j ) Dimension of one chunk by Y-axis, default 2

--ch_nz (-k ) Dimension of one chunk by Z-axis, default 2

This program writes dataset sing chunking. Each process writes nz/ch_nz chunks in cycle.

Attention!

Amount of threads must be equal nx * ny/(ch_nx * ch_ny), while ch_nx % nx, ch_ny % ny and ch_nz % nz must be zerro for valid work of this application.

Reading of "SDS_chnk.h5" output file: https://git.jinr.ru/abelova/readdataset