You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardexpand all lines: docs/source/backends/adios2.rst
+5
Original file line number
Diff line number
Diff line change
@@ -48,6 +48,10 @@ Exceptions to this are the BP3 and SST engines which require their endings ``.bp
48
48
49
49
For file engines, we currently leverage the default ADIOS2 transport parameters, i.e. ``POSIX`` on Unix systems and ``FStream`` on Windows.
50
50
51
+
.. tip::
52
+
53
+
Use the ``adios2.engine.treat_unsupported_engine_as`` :ref:`JSON/TOML parameter <backendconfig-adios2>` for experimentally interacting with an unsupported ADIOS2 engine.
``OPENPMD_ADIOS2_PRETEND_ENGINE`` *empty* Pretend that an (unknown) ADIOS2 engine is in fact another one (also see the ``adios2.pretend_engine`` :ref:`parameter <backendconfig-adios2>`).
84
89
``OPENPMD2_ADIOS2_USE_GROUP_TABLE`` ``0`` Use group table (see below)
85
90
``OPENPMD_ADIOS2_STATS_LEVEL`` ``0`` whether to generate statistics for variables in ADIOS2. (``1``: yes, ``0``: no).
86
91
``OPENPMD_ADIOS2_ASYNC_WRITE`` ``0`` ADIOS2 BP5 engine: 1 means setting "AsyncWrite" in ADIOS2 to "on". Flushes will go to the buffer by default (see ``preferred_flush_target``).
Copy file name to clipboardexpand all lines: docs/source/backends/hdf5.rst
+14
Original file line number
Diff line number
Diff line change
@@ -15,6 +15,15 @@ I/O Method
15
15
16
16
HDF5 internally either writes serially, via ``POSIX`` on Unix systems, or parallel to a single logical file via MPI-I/O.
17
17
18
+
Virtual File Drivers
19
+
********************
20
+
21
+
Rudimentary support for HDF5 VFDs (`virtual file driver <https://www.hdfgroup.org/wp-content/uploads/2021/10/HDF5-VFD-Plugins-HUG.pdf>`_) is available (currently only the *subfiling* VFD).
22
+
Note that the subfiling VFD needs to be enabled explicitly when configuring HDF5 and threaded MPI must be used.
23
+
24
+
Virtual file drivers are configured via JSON/TOML.
25
+
Refer to the page on :ref:`JSON/TOML configuration <backendconfig-hdf5>` for further details.
26
+
18
27
19
28
Backend-Specific Controls
20
29
-------------------------
@@ -45,6 +54,10 @@ Although we choose the default to be non-collective (independent) for ease of us
45
54
For independent parallel I/O, potentially prefer using a modern version of the MPICH implementation (especially, use ROMIO instead of OpenMPI's ompio implementation).
46
55
Please refer to the `HDF5 manual, function H5Pset_dxpl_mpio <https://support.hdfgroup.org/HDF5/doc/RM/H5P/H5Pset_dxpl_mpio.htm>`_ for more details.
47
56
57
+
.. tip::
58
+
59
+
Instead of using an environment variable, independent/collective data transfer can also be configured at the API level via :ref:`JSON/TOML <backendconfig-hdf5>`.
60
+
48
61
``OPENPMD_HDF5_ALIGNMENT``: this sets the alignment in Bytes for writes via the ``H5Pset_alignment`` function.
49
62
According to the `HDF5 documentation <https://support.hdfgroup.org/HDF5/doc/RM/H5P/H5Pset_alignment.htm>`_:
50
63
*For MPI IO and other parallel systems, choose an alignment which is a multiple of the disk block size.*
@@ -56,6 +69,7 @@ Any file object greater than or equal in size to threshold bytes will be aligned
56
69
57
70
``OPENPMD_HDF5_CHUNKS``: this sets defaults for data chunking via `H5Pset_chunk <https://support.hdfgroup.org/HDF5/doc/RM/H5P/H5Pset_chunk.htm>`__.
58
71
Chunking generally improves performance and only needs to be disabled in corner-cases, e.g. when heavily relying on independent, parallel I/O that non-collectively declares data records.
72
+
The chunk size can alternatively (or additionally) be specified explicitly per dataset, by specifying a dataset-specific chunk size in the JSON/TOML configuration of ``resetDataset()``/``reset_dataset()``.
59
73
60
74
``OPENPMD_HDF5_COLLECTIVE_METADATA``: this is an option to enable collective MPI calls for HDF5 metadata operations via `H5Pset_all_coll_metadata_ops <https://support.hdfgroup.org/HDF5/doc/RM/RM_H5P.html#Property-SetAllCollMetadataOps>`__ and `H5Pset_coll_metadata_write <https://support.hdfgroup.org/HDF5/doc/RM/RM_H5P.html#Property-SetCollMetadataWrite>`__.
61
75
By default, this optimization is enabled as it has proven to provide performance improvements.
0 commit comments