| Version 21 (modified by raasch, 4 years ago) (diff) |
|---|
Job chains / Restart runs
Batch systems generally limit the CPU time that is allowed to be requested by a job, e.g. to a maximum of 12 hours or 24 hours. If a simulation needs more time to run, it has to be split into several parts/jobs. The first job is called the initial run or job, the others are called restart runs/jobs. Together they form a so-called job chain. Restart runs require as input the state of all flow variables as they were calculated in the final time step of the previous run. They need to be output by the previous run into a so-called restart-file which is a required input file for the (next) restart run.
palmrun allows you to automatically generate job chains and to handle the restart files. Of course, automatic generation does not work if you run PALM in interactive mode.
A job started by palmrun will be queued by the queuing-system of the local or remote computer into a suitable job class which fulfills the requirements that are set via palmrun options -t, -m, -X, and -T, which define the requested CPU-time, memory, and number of cores. Each job class permits only jobs with certain maximum requirements (e.g. the allowed CPU time or the maximum number of cores that can be used by the job). Some queuing systems automatically sort the jobs into the respective job class, others require an explicit setting of the class for which the job shall be queued. You can set the job class via palmrun option -q. The job classes are important for the scheduling process of the computer. Jobs with small requirements usually come to execution very fast, jobs with larger requirements must wait longer (sometimes several days). Be aware that the available job classes vary a lot among different computer centers.
Before starting a run, you have to estimate how much CPU time your complete simulation will need. The required time in seconds has to be given with palmrun option -t. Due to the fact that the model uses a variable time step by default, the number of time steps to be carried out and consequently the time required to finish the simulation can often only roughly be estimated. So it may happen that more time is needed to finish the simulation than indicated by option -t. That will normally cause the job scheduler to terminate the job as soon as the available CPU time is consumed. In principle, you may avoid this problem by setting a very generously estimated value for -t, but the maximum allowed CPU-time is often limited due to job class restrictions.
Restart runs are the method to circumvent these job class restrictions. During the time stepping, PALM is able to continuously check how much time is left for the execution of the job. If the run can not be completed and finished before expiration of this time, the PALM stops and outputs (nearly) all model variables (especially the 3d-prognostic quantities) in binary format to a file (or folder) with local name BINOUT). After the local output files have been saved, palmrun automatically generates a restart run. For this purpose a new palmrun call is automatically initiated, i.e. palmrun recursively calls itself. The palmrun options of this call correspond to those of the initial call. PALM restarts and this time, at the beginning, it reads the binary data that have been written by the previous run, and continues the run based on this final state of the previous run. If the simulation still cannot be finished, another restart run is generated, etc., until the time to be simulated is reached (this is the one set via parameter end_time). This way a whole set of restart runs may be generated - a so-called job chain.
Restart runs require certain entries in the file-connection file (see .palm.iofiles, and its description) and in the parameter file, which will be described and explained now.
The following entries are important and are already contained in the default file-connection file:
PARIN in:tr d3# $base_data/$run_identifier/INPUT _p3d*
PARIN in:tr d3r $base_data/$run_identifier/INPUT _p3dr*
BININ in:lnpe d3r $restart_data_path/$run_identifier/RESTART _d3d*
#
BINOUT* out:lnpe restart $restart_data_path/$run_identifier/RESTART _d3d
The palmrun call for the initialization run of the job chain reads e.g.:
palmrun -c default -r abcde -t 900 -X 96 -T 48 -a “d3# restart” -b -q queue_name
Giving the activation string restart as argument of option -a is essential. Only for that case the model writes binary data for a restart run to the local file BINOUT (in case of running on more than 1 core, BINOUT is a folder). The local output file is then saved to a permanent file as defined in the file connection statement for BINOUT. The last line of the example above shows that savinf of the file is only be done of the activation string restart has been set.
Only by specifying restart as activation string, PALM is instructed to compute the remaining CPU time after each time step and to stop, if the run is not going to be completed and finished briefly before expiration of this time. Actually the stop takes place when the difference between the available job time (determined by the palmrun option -t) and the time used by the job so far becomes smaller than the time given by the runtime parameter termination_time_needed. The runtime parameter termination_time_needed can be used to inform PALM, how much time is required for copying the binary data for restart runs, as well as for other pre- or post-processing steps that are done within the job. Thus, as soon as the remaining job time is less than termination_time_needed, PALM interrupts the time stepping and outputs the restart data to local file/folder BINOUT. The initialization parameters are also added to that file. In a last step, PALM creates a flag file with local name CONTINUE_RUN?. The presence of this file signals palmrun that a restart run needs to be generated and initiates and starts a respective job.
Within PALM, the initial phase of a restart run requires different actions than during during an initial run. In case of a restart, PALM first needs to read the data written by the preceding run and also reads the initialization parameters from the same file. Therefore, these parameters do not need to be provided in the parameter file (local name PARIN). Anyhow, if they are provided and if their value differ from the respective value of the initial run, these settings are ignored. There is exactly one exception to this rule: the initialization parameter initializing_actions determines whether the job is a restart run or an initial run. If initializing_actions = 'read_restart_data', then it is a restart run, otherwise an initial run. The previous explanation make it clear that the model obviously needs two different parameter files (local name PARIN) for the case of job chains. One is required for the initial run and contains all initialization parameters and the other one is needed for restart runs. The last one only contains the initialization parameter initializing_actions (any other initialization parameters may appear in this file, but they will be ignored), which needs to be set to 'read_restart_data'. So you need to provide two different parameter files if you like to carry out restart runs. Since PALM always expects the parameters to be in the local file PARIN, regardless if it is an initial or a restart run, two different file connection statements must be given for that file in the file-connection file. One is active for the initial run only, the other one only for restart runs. The palmrun call for the initial run shown above activates the first of the two specified file connection statements for PARIN, because the activation string d3# with the option -r coincides with the string in the third column of the file connection statement. Obviously the next statement
PARIN in:tr d3r $base_data/$run_identifier/INPUT _p3dr*
must be active for the restart runs. Given that this statement only gets active if the option -r is given the value d3r and that the palmrun call for this restart run is produced automatically (thus not by the user), palmrun obviously has to replace "d3#" of the initial run with "d3r" within the call of this restart run. Actually, with restart runs all "#" characters within the strings given for the options -r , -i and -o are replaced by "f".
For example, for the initial run the permanent file
~/palm/current_version/JOBS/abcde/INPUT/abcde_p3d
and for restart runs the permanent file
~/palm/current_version/JOBS/abcde/INPUT/abcde_p3dr
is used. Only with restart runs the local file BININ is made available as input file, because the appropriate file connection statement also contains the character string "d3r" in the third column. This is logical and necessary since in BININ the binary data, produced by the model of the preceding job of the chain, are expected and the initial run does not need these data. The permanent names of this input file (local name BININ) and the corresponding output file (local name BINOUT) are identical and read
~/palm/current_version/JOBS/abcde/RESTART/abcde_d3d
However, after the file produced by the previous job was read in by the model and after the local file BINOUT was produced at the end of the job, the restart job does not overwrite this permanent file (.../abcde_d3d) with the new data. Instead of that, it is examined whether already a permanent file with the name .../abcde_d3d exists when copying the output file (BINOUT) of palmrun. If this is the case, BINOUT is copied to the file .../abcde_d3d.1. Even if this file is already present, .../abcde_d3d.2 is tried etc. For an input file the highest existing cycle of the respective permanent file is copied. In the example above this means: the initial run creates the permanent file .../abcde_d3d, the first restart run uses this file and creates .../abcde_d3d.1, the second restart run creates .../abcde_d3d.2 etc. After completion of the job chain the user can still access all files created by the jobs. This makes it possible for the user for example to restart the model run of a certain job of the job chain again.
Therefore restart jobs can not only be started automatically through palmrun, but also manually by the user. This is necessary e.g. whenever after the end of a job chain it is decided that the simulation must be continued further, because the phenomenon which should be examined did not reach the desired state yet. In such cases the palmrun options completely correspond to those of the initial call; simply the "#" characters in the arguments of options -r, -i and -o must be replaced by "f".
Handling of large (restart) files
In case of very large files, the copy of data from and to palmrun'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.
cp large_local_file large_permanent file # may take long time ln existing_large_local_TARGET_file LINK_NAME_to_large_local_file # is done immediately, i.e. requires almost no time
You can tell palmrun to use ln instead of cp by giving the file attribute ln in the respective file connection statement, e.g.:
BININ in:loc:lnpe d3r $base_data/$run_identifier/RESTART _d3d BINOUT out:loc:lnpe restart $base_data/$run_identifier/RESTART _d3d
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. 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).
Most computing centers provide a file systems for fast I/O and this should be used as palmrun'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):
# # folder in which palmrun's temporary working catalog is created (will be deleted after end of job) %tmp_user_catalog /gfs2/work/<replace by username> lccrayh parallel # # folder in which large binary files shall be stored %tmp_data_catalog /gfs2/work/<replace by username> lccrayh parallel # # file connection statements for restart files BININ in:loc:lnpe d3r $tmp_data_catalog/$run_identifier/RESTART _d3d BINOUT out:loc:lnpe restart $tmp_data_catalog/$run_identifier/RESTART _d3d
Such fast file systems are generally not allowed to store files for a longer time, so the user has to take care for archiving himself.
