-
I have some large transient simulations with 25k steps and result files >100 GB. When a simulation is finished, there are a lot of files (eg. file.rst and additional 4 files file0.rst ... for each process, 25k step-files, ...). As far as I understand, not all of these files are needed for post-processing. Why does pymapdl not clean up after a simulation (or can I force it to do so)? Which files are neccessary for postprocessing? Also - can I somehow change the path for the temporary files when using |
Beta Was this translation helpful? Give feedback.
Replies: 6 comments 2 replies
-
Hi @natter1 The MAPDL default is to keep everything just in case they are needed later. For example at one time it was necessary to have all the "core" files (those whose jobname is appended with the _N where N is 0 to the number of CPU cores used minus 1) for a multi-frame restart analysis. Then the multi-frame restart procedure changed such that we can restart a distributed parallel solution with the set of combined files. So the question becomes what to do with the files given that MAPDL needs the combined result file to post-process the whole FEM domain. There is no answer to suite everyone...so see the APDL command /FCLEAN. /FCLEAN is mean to be used to 'clean up' the core files after a distributed solve. Now PyMAPDL has not implemented /FCLEAN. Before I recommend its usage in a mapdl.run style I'd like to ask @akaszynski and/or @germa89 to weigh in first. I did test and it seems to delete the files as expected. But I'm not a developer and there could be a reason to not use this. Or maybe it's better to use the Python OS module remove to delete a file? If /FCLEAN is ok to use, be sure to issue it after leaving the solution processor for the last time. The distributed core files are auto-combined upon finishing out of the solution processor. So something like: mapdl.solve()
mapdl.finish()
mapdl.run("/fclean") Regarding the question on open_gui(), please see the "inplace" option that allows us to start MAPDL interactively in the same working directory as the PyMAPDL MAPDL instance. Mike |
Beta Was this translation helpful? Give feedback.
-
Hi @natter1 mapdl.solve()
mapdl.finish()
mapdl.slashdelete('file','rst',1) The 1 in the third field of the slashdelete command results in the core rst files being deleted but the combined rst file is left. You could repeat for the other non-needed files like esav etc. Mike |
Beta Was this translation helpful? Give feedback.
-
I'd recommending using It's kept |
Beta Was this translation helpful? Give feedback.
-
I found that one too, and its better than nothing. But in that case you use th original - and I would like to work on a copy in some cases to not mess up the original.
Hmm, as far a I understand, this will delete the whole simulation result afterwards. I want to keep the result file - but I would like to only save a 100GB data instead of 200GB, if thats enough.
|
Beta Was this translation helpful? Give feedback.
-
Thanks @mikerife |
Beta Was this translation helpful? Give feedback.
-
@mikerife Got an additional question regarding this topic. Are the step-files ( |
Beta Was this translation helpful? Give feedback.
Hi @natter1 The MAPDL default is to keep everything just in case they are needed later. For example at one time it was necessary to have all the "core" files (those whose jobname is appended with the _N where N is 0 to the number of CPU cores used minus 1) for a multi-frame restart analysis. Then the multi-frame restart procedure changed such that we can restart a distributed parallel solution with the set of combined files. So the question becomes what to do with the files given that MAPDL needs the combined result file to post-process the whole FEM domain. There is no answer to suite everyone...so see the APDL command /FCLEAN. /FCLEAN is mean to be used to 'clean up' the core files afte…