============== Idealized Case ============== The following sub-directories can be found in your *practice_exercises/* directory: * **wrf_ideal_serial** : Contains all idealized cases, built for serial computation (i.e., a single processor) * **wrf_ideal_dm** : Contains only the 3D idealized cases, built for parallel computation (multiple processors - using the *dmpar* compile option) | | Set-up an Ideal Case ==================== #. Move to the appropriate directory. .. code-block:: :class: copyable cd /glade/derecho/scratch/$USER/practice_exercises/wrf_ideal_serial/test/em_case_you_desire Or, to run a 3D case (this will take longer than a 2d case): .. code-block:: :class: copyable cd /glade/derecho/scratch/$USER/practice_exercises/wrf_ideal_dm/test/em_case_you_desire |vspace| #. Either use the default *namelist.input* to run the default case, or modify *namelist.input* for integration length, output frequency, domain size, time step, physics options, and other parameters you'd like. | | Initialization ============== Run the ideal.exe to generate initial conditions (*wrfinput_d01*). Because idealized cases are not on a real grid, no lateral boundary file is required. Instead they use other boundary options. `See details `_ |vspace| .. note:: The *em_fire* case requires `additional steps `_ before running ideal.exe. |vspace| The way WRF was built determines the way it is run. * **Serial cases** *(wrf_ideal_serial/test/em_\*)* : follow instructions for :ref:`Serial Initialization` * **Distributed-memory cases** *(wrf_ideal_dm/test/em_\*)* : follow instructions for :ref:`Distributed Memory Initialization` | .. _Serial Initialization: Serial Initialization --------------------- These instructions are specific to cases found in the *wrf_ideal_serial/test/em_\** directories. Issue the following command: .. code-block:: :class: copyable ./ideal.exe | .. _Distributed Memory Initialization: Distributed Memory Initialization --------------------------------- These instructions are specific to cases found in the *wrf_ideal_dm/test/em_\** directories. Issue the following command: .. code-block:: :class: copyable qsub runideal.sh |vspace| .. note:: **Only use a single processor to run ideal.exe**, even for a distributed-memory case. This is applicable on all machines - not just the NCAR HPC. | | Run WRF ======= The way WRF was built determines the way it is run. * **Serial cases** *(wrf_ideal_serial/test/em_\*)* : follow instructions for :ref:`Serial WRF` * **Distributed-memory cases** *(wrf_ideal_dm/test/em_\*)* : follow instructions for :ref:`Distributed Memory WRF` | .. _Serial WRF: Serial WRF ---------- These instructions are specific to cases found in the *wrf_ideal_serial/test/em_\** directories. |vspace| When running WRF serially on NCAR's HPC, use the ``qcmd -A UMMM0017`` prefix to launch wrf.exe on a batch node, preventing login node saturation. This is specific to this tutorial and NCAR's HPC. .. code-block:: :class: copyable qcmd -A UMMM0017 -- ./wrf.exe >& wrf.out | .. _Distributed Memory WRF: Distributed Memory WRF ---------------------- These instructions are specific to cases found in the *wrf_ideal_dm/test/em_\** directories. Issue the following command: .. code-block:: :class: copyable qsub runwrf.sh | | Generate Graphics ================= If interested, `generate graphics <./graphics.html>`_ with one of the available post-processing packages. | | Return to the Practice Exercise home to page to run another exercise. .. rst-class:: horizbuttons-primary-m * `Return to Practice Exercises <./practices.html>`_ .. toctree:: :maxdepth: 4 :hidden: | | | |