Configure JADE

All configuration files are included in the <JADE_root>/cfg directory. This folder contains many files but most of them are of no interest of normal users (i.e. not developers).

Set environmental variables

The first thing that should be done after jade installation is to have a look at the <JADE_root>/cfg/env_vars_cfg.yml file. This file contains the environmental variables that are used by JADE to run the different codes. Here is an example:

It can be seen that in order to submit a job to a cluster, the user needs to provide the path to the batch template file for the code(s) to be run. The batch template file is the job submission script to be utilised on the users chosen system. Several default job submission scripts are provided in the cfg/exe_config folder.

Important

In case of a global_job, in addition to the modules, the bash template should also contain the export command to load the cross sections (different command depending on the code). This is automatically handled by JADE in case of job mode.

These file should contain all the necessary job submission options (e.g. slurm directives) and all the necessary commands to load the modules required by the different transport codes. Some placeholder are defined and will be substituted by JADE at runtime when submitting the job. They are all optional: - INITIAL_DIR: the directory where the job is submitted from (i.e. the simulation folder) - OUT_FILE: the file where the standard output will be written - ERROR_FILE: the file where the standard error will be written - MPI_TASKS: the number of MPI tasks to be used - OMP_THREADS: the number of OpenMP threads to be used - USER: the user who submitted the job

Configure the libraries

Another mandatory configuration is the one of nuclear data libraries. The user should provide the path to libraries he intends to use in the <JADE_root>/cfg/libs_cfg.yml file. Two different kind of libraries are supported for the moment, normal transport libraries and D1S libraries. The following is an example of the settings of a transport library:

FENDL 3.2c:  # this is the name of the library. It will be used in all outputs
    mcnp:
        path: /path/to/xsdir  # path to the xsdir file
        suffix: 32c  # correspondent suffix in the xsdir file
    openmc:
        path: path/to/cross_sections.xml  # path to the library file
    serpent:  # TODO
        path: pathtolib

There is no need to provide paths for the codes that the user does not intend to use.

In some cases, certain cross sections may be missing from a library for given material. This requires an alternative library to be used for the missing nuclides. In the case of the Sphere for OpenMC where the standard material definitions are used, this requires an update to cross_sections.xml file after generation of the inputs. The user may run the Sphere benchmark with the cross_sections.xml for a particular library and after discovering certain runs are not completed due to missing nuclides, update the cross_sections.xml file, appending a library which contains those that are missing. The user may then perform a continue run using the –cnt run (see Configure the run). Otherwise if the user is aware from the start that certain nuclides are missing, they may first generate the inputs using the OnlyInput run option (see Continue executions) and thereafter change the cross_sections.xml file that is being pointed to before running. In benchmarks other than the Sphere, the user may point to a complete cross_sections.xml in their original configuration (noting OpenMC uses cross sections in the order they appear in the file) however this must not be done prior to generation of the inputs for the Sphere as the cross_sections.xml is used to determine which nuclides are run.

Finally, an example of a D1S library settings:

D1SUNED (FENDL 3.1d+EAF2007):
    d1s:
        path: /path/to/xsdir  # path to the xsdir file for the D1SUNED lib
        suffix: 99c  # correspondent suffix in the xsdir file for the D1SUNED lib
        # To run simulations, also a normal transport library is needed as not all
        # isotopes will be set to be activated (i.e. will use the D1S library)
        transport_library_path: /path/to/xsdir  # path to the xsdir file for the transport lib
        transport_suffix: 32c  # correspondent suffix in the xsdir file for the transport lib