1Blurb:: Write evaluation data to HDF5 for all models that belong directly to methods
2
3Description::
4
5When this option is selection, evaluation data for all models that belong
6directly to methods will be written to HDF5. Models that belong to other
7models (e.g. wrapped models in a recast relationship) are not stored.
8
9By default, when HDF5 output is enabled, Dakota writes evaluation
10data only for the model that belongs to the top-level method. This
11keyword group is used to override the default.
12
13HDF5 output is an experimental feature, and the contents and organization of
14the output file is subject to change. The current organization and a
15brief explanation of HDF5 is provided in the \ref hdf5_output section
16of this manual.
17
18Topics:: dakota_output
19
20Examples::
21
22For the example input in the parent keyword, the following model groups would be
23written:
24
25- \c /models/recast/RECAST_surr_RECAST_1/
26- \c /models/simulation/truth_m/
27
28The recast model, in this case, is automatically generated by Dakota to handle summing the
29pair of objective functions. It is the model that belongs directly to the method 'opt'. See
30\ref hdf5_evaluations-hdf5_eval_sources for further explanation.
31
32The following links would be added to methods and models \c sources groups:
33
34- \c /methods/opt/sources/RECAST_surr_RECAST_1 \f$ \rightarrow \f$ \c /models/recast/RECAST_surr_RECAST_1/
35- \c /methods/training/sources/truth_m \f$ \rightarrow \f$ \c /models/simulation/truth_m/
36
37Depending on \ref environment-results_output-hdf5-interface_selection, the following links may also
38be added to model \c sources groups:
39
40- \c /models/simulation/truth_m/sources/truth \f$ \rightarrow \f$ \c /interfaces/truth/truth_m/
41
42
43Theory::
44Faq::
45See_Also::
46