Grids and fields I/O

Topics
	ADIOS2 interface
	Interface routine to the massivelly parallel I/O ADIOS2 library.
	ADIOS interface
	Interface routines to the massivelly parallel I/O ADIOS library.
	Disk cache
	Store fields to reduce computations in next runs.
	EnSight Gold writer
	Provide a class to write EnSight Gold files.
	HDF5 interface
	Interface routine to the write Hdf5 files compatible with Pixie and Xdmf formats.
	MOF mesh generator
	Generate a mesh from MOF structure to visualize the MOF reconstruction.
	Read global cell file
	Read a field in a file that contains the global cell field.
	Linked list of fields
	Linked list of fields used for writing data fields.
	XDMF interface
	Interface routine to the write XDMF outputs composed of and HDF5 data file and an XML data description file.

Namespaces
module	variables_output
	Define variables associated to outputs.
module	mod_write_time
	Write Time in a .vtk file.

Functions
subroutine	mod_io_write_grids_and_fields::io_write_grids_and_fields ()
	Write fields specified by the variables_output::is_print_* parameters. More...

Detailed Description

ADIOS

The Adaptable IO System [ADIOS](https://www.olcf.ornl.gov/center-projects/adios] provides a simple, flexible way for scientists to describe the data in their code that may need to be written, read, or processed outside of the running simulation. Files generated can be readen by VisIt.

ADIOS is used by default as output format for visualization and checkpoint / restart files. ".bp" files are generated.

ADIOS2

[ADIOS2](https://adios2.readthedocs.io/en/latest] is the latest implementation of ADIOS. This brand new architecture continues the performance legacy of ADIOS1, and extends its capabilities to address the extreme challenges of scientific data IO.

Coupling with Notus is experimental and done only for 3D files (should be easily adapted to 2D)

At each time iteration, a "_bp4" directory is created that contains the data and an XML description file that can be readen by Paraview and VisIt.

Pixie

Pixie file format is based on parallel [HDF5](https://www.hdfgroup.org/solutions/hdf5] (high-performance data management and storage suite). Files can be readen by VisIt (and Paraview if VisIt plugins are built). It is designed for very large datasets. The advantage compared to ADIOS plugin is that domain decomposition is done by ViSit launched in parallel. Performance is thus increased. Moreover, very large mesh size (above 2 billion cells) can be readen by ViSit if Pixie output is used.

One .h5 file is written at each time iteration. Done for 3D only (should be easily adapted to 2D)

Xdmf

Based on HDF5, Xdmf files can be readen by VisIt (without domain decomposition) or Paraview. Three files are generated at each time iteration: an ".h5" file that contains the data, another ".h5" file that contains the mesh coordinates, and an XML file that describes the data.

Function Documentation

io_write_grids_and_fields()

subroutine mod_io_write_grids_and_fields::io_write_grids_and_fields

This routine write all fields upon each call. The actual output format is specified by the variables_output::output_type variable. Secondary variables, such as q_criterion and strain_rate_magnitude, are computed in this routine.

It is recommended to specialize all the following subroutines for specific output engines or libraries.