Changes between Version 10 and Version 11 of doc/app/runs


Ignore:
Timestamp:
Jul 4, 2017 9:20:48 AM (8 years ago)
Author:
raasch
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • doc/app/runs

    v10 v11  
    4343= Handling of large binary restart- or output-files =
    4444
    45 In case of very large files, the copy of data from and to PALM's temporary working directory may need a long time. The CPU cores requested for the job run idle during that time and may consume significant amount of the job time without doing anything. The time required for copying can be spared by using a file link instead of copying the data.
     45In case of very large files, the copy of data from and to '''mrun's''' temporary working directory may need a long time. The CPU cores requested for the job run idle during that time and may consume significant amount of the job time without doing anything. The time required for copying can be spared by using a file link instead of copying the data.
    4646{{{
    4747   cp large_local_file  large_permanent file                                 # may take long time
    4848   ln existing_large_local_TARGET_file  LINK_NAME_to_large_local_file        # is done immediately, i.e. requires almost no time
    4949}}}
    50 You can tell '''mrun''' to use {{{ln}}} instead of {{{cp}}} by giving the file attribute {{{ln}}} in the respective file connection statement:
     50You can tell '''mrun''' to use {{{ln}}} instead of {{{cp}}} by giving the file attribute {{{ln}}} in the respective file connection statement, e.g.:
    5151{{{
    52 bla
     52BININ   in:loc:lnpe  d3f       $base_data/$fname/RESTART  _d3d
     53BINOUT  out:loc:lnpe restart   $base_data/$fname/RESTART  _d3d
    5354}}}
    54 However, performing a link requires that the link to a TARGET file with the name LINK_NAME must be located on the same physical file system as the TARGET file ...
     55However, performing a link requires that the link to a TARGET file with the name LINK_NAME must be located on the same physical file system as the TARGET file. If TARGET file and LINK_NAME are on different file systems, the TARGET file will be copied instead (and the advantage of using the {{{ln}}} command is lost).
     56
     57Most computing centers provide a file systems for fast I/O and this should be used as '''mrun's''' temporary working directory, which can be set in the configuration file by environment variable {{{tmp_user_catalog}}}. Since the LINK_NAME should be on the same file system, the user should provide a directory on that file system for storing the large files. Respective settings in the configuration file could be (example for Cray-XC40 at HLRN):
     58{{{
     59#
     60# folder in which mrun's temporary working catalog is created (will be deleted after end of job)
     61%tmp_user_catalog    /gfs2/work/<replace by username>     lccrayh parallel
     62#
     63# folder in which large binary files shall be stored
     64%tmp_data_catalog    /gfs2/work/<replace by username>     lccrayh parallel
     65#
     66# file connection statements for restart files
     67BININ   in:loc:lnpe  d3f       $tmp_data_catalog/$fname/RESTART  _d3d
     68BINOUT  out:loc:lnpe restart   $tmp_data_catalog/$fname/RESTART  _d3d
     69}}}
     70Such fast file systems are generally not allowed to store files for a longer time, so the user has to take care for archiving himself.
     71
     72'''Attention:'''\\
     73The {{{ln}}} file attribute and the above described method for storing large binary files has been introduced with revision number r2262. {{{mrun}}} does not create empty files/directories for restart files in the folder given with the file connection statement any more, as it was done with feature "fl" before.