...
Note: The page may be changed until the competition stats, maybe sure to follow up until the opening ceremony.
RegCM Presentation to the teams:
Widget Connector | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|
|
Presentation file:
View file | ||
---|---|---|
|
Building and Running example
...
The created configure script contains various options to conditionally activate some part of the code, provide different I/O options or optimize for some HPC platforms the ICTP has had access to in the past years.
Not all the legacy platforms/compilers combination are anymore relevant, supported or even tested if the ICTP has not access to them.
Set the path to point your NetCDF installation to avoid from using another NetCDF version.
Code Block |
---|
export PATH=<path to netcdf>/bin:$PATH |
To configure the build system using the GNU make program should be enough on most systems to just type in:
...
Code Block |
---|
make version make install |
Sample build script on PSC:
Code Block |
---|
#!/bin/bash
source /jet/packages/oneapi/v2023.2.0/compiler/2023.2.1/env/vars.sh
source /jet/packages/oneapi/v2023.2.0//mpi/2021.10.0/env/vars.sh
export MPIFC=mpiifort
export CC=mpiicc
export FC=$MPIFC
export HDF5_DIR=<path>
export NETCDF_DIR=<path>
export PATH=$NETCDF_DIR/bin:$PATH
autoreconf -f -i
./configure --prefix=$BASE/build-$MPI
make version
make install |
For the ISC24 competition, the build system is configured to make the RCEMIP model. This setting will initialize a prescribed initial vertical profile simulation with periodic boundary conditions, that will be used to evaluate the model in the ensemble Radiative-Convective Equilbrium Model Intercomparison Project. In this framework, the model is run in an idealized setting using a prescribed convection permitting configuration over an ocean box domain. The selected configuration it is very much stressing the dynamical core, the radiation package, the model micro-physics and the turbulence parameterization of the boundary layer over the ocean.
...
possibly on a convenient I/O system partition with high throughput. The user must then link/copy some files in the run directory:
Code Block |
---|
ln -sf $REGCM_ROOT/bin/regcmMPIRCEMIP .
cp $REGCM_ROOT/Testing/isc24.in .
cp $REGCM_ROOT/Testing/rcemip.in profile.in
mkdir output |
...
To run the model, the user needs to invoke the executable using mpirun providing as argument the name of the namelist file used for the competition:
Code Block |
---|
mpirun ./-np <# procs> regcmMPIRCEMIP isc24.in |
While running, the model will produce diagnostic text on the standard output. The below is an example relative to a model test run on an ICTP Intel Core i7-8700 CPU @ 3.20GHz desktop computer with 6 cores:
...
Build and Run the application with the given input isc24.inc in using 4 CPU nodes, submit the best performance to the team’s folder.
Run IPM or another MPI profiler and show the most 3 MPI calls being used, show your work in the teams presentation.
Visualize the input output netCDF files (in the output directory) and create video or set of photos based on that using any tool. show your work in the teams presentation. The netCDF output data should contain all the metadata required to use most of the visualization packages.
...