Skip navigation links
(CGNS Documentation Home Page) (Steering Committee Charter) (Overview and Entry-Level Document) (A User's Guide to CGNS) (Mid-Level Library) (Standard Interface Data Structures) (SIDS File Mapping Manual) (CGIO User's Guide) (Parallel CGNS User's Guide) (ADF Implementation) (HDF5 Implementation) (Python Implementation) (CGNS Tools and Utilities)

(General Remarks) (The PCGNS Software Library) (Example Programs)

General Remarks

The Parallel CGNS library, or PCGNS, is built on top of the collective I/O support in HDF5 using MPI-IO. PCGNS provides specialized routines to open and close a file for parallel I/O and read and write grid coordinates, element data, solution data, and general arrays in a collective fashion. The PCGNS functions are prefixed with cgp_, and documented in The PCGNS Software Library.

All additional accesses to the data use the standard Mid-Level Library (MLL) functions. A typical use of the PCGNS routines in a parallel I/O program would be:

cgp_open              --- open file for parallel I/O
cg_base_write         --- create base and zone using MLL
cgp_coord_write       --- create coordinate dataset for parallel I/O
cgp_coord_write_data  --- write the coordinates collectively
cgp_close             --- close the file

An example Fortran and C program are given in Example Programs.

A detailed discussion on the implementation may be found in the PDF paper "An Efficient and Flexible Parallel I/O implementation for the CFD General Notation System" by Horne, Bensen, and Hauser.