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
We currently have 2 types of files that store the state of a stellar model:
model (.mod) files, which store only the basic structure, thermodynamic state, and composition. These are plain text files, and are portable across machines and MESA versions.
photos (e.g. photos/x050 in a work directory), which are non-portable binary files, but have the advantage of preserving the full state of the model and enabling restarts that proceed exactly as if no interruption had occurred. What constitutes the "full state" of a stellar model depends on MESA version, so this functionality cannot be portable across MESA versions.
In principle, HDF5 could enable making photos fully portable across different machine architectures for a given MESA version. We could therefore merge the functions of models and photos into one file type, with the caveat that these won't necessarily be fully portable across different MESA versions. Perhaps it would be worth discussing the best strategy for what to do when loading one of these model/photo files into a newer MESA version where the components of the star structure have changed and would therefore not be fully specified by an older file. One option would be to fall back to just the basic structure quantities that are currently represented in .mod files.
Since .mod files have been the basic MESA idiom for storing stellar models for over a decade, we would of course also need to continue to support loading and running from these.
The text was updated successfully, but these errors were encountered:
We currently have 2 types of files that store the state of a stellar model:
.mod
) files, which store only the basic structure, thermodynamic state, and composition. These are plain text files, and are portable across machines and MESA versions.photos/x050
in a work directory), which are non-portable binary files, but have the advantage of preserving the full state of the model and enabling restarts that proceed exactly as if no interruption had occurred. What constitutes the "full state" of a stellar model depends on MESA version, so this functionality cannot be portable across MESA versions.In principle, HDF5 could enable making photos fully portable across different machine architectures for a given MESA version. We could therefore merge the functions of models and photos into one file type, with the caveat that these won't necessarily be fully portable across different MESA versions. Perhaps it would be worth discussing the best strategy for what to do when loading one of these model/photo files into a newer MESA version where the components of the star structure have changed and would therefore not be fully specified by an older file. One option would be to fall back to just the basic structure quantities that are currently represented in
.mod
files.Since
.mod
files have been the basic MESA idiom for storing stellar models for over a decade, we would of course also need to continue to support loading and running from these.The text was updated successfully, but these errors were encountered: