pism / pism-gris

Scripts and tools to make simulations of the Greenland Ice Sheet (GrIS) with PISM

PISM, a Parallel Ice Sheet Model

_ _ _

The Parallel Ice Sheet Model is an open source, parallel, high-resolution ice sheet model:

• hierarchy of available stress balances
• marine ice sheet physics, dynamic calving fronts
• polythermal, enthalpy-based conservation of energy scheme
• extensible coupling to atmospheric and ocean models
• verification and validation tools
• documentation for users and developers
• uses MPI and PETSc for parallel simulations
• reads and writes CF-compliant NetCDF files

PISM is jointly developed at the University of Alaska, Fairbanks (UAF) and the Potsdam Institute for Climate Impact Research (PIK). UAF developers are based in the Glaciers Group at the Geophysical Institute.

Please see ACKNOWLEDGE.rst and doc/funding.csv for a list of grants supporting PISM development.

Homepage

http://www.pism.io/

Download and Install

See the Installing PISM on pism.io.

Support

Please e-mail [email protected] with questions about PISM.

You can also join the PISM workspace on Slack.

Contributing

Want to contribute? Great! See Contributing to PISM.

pism-gris's Issues

Job scripts should not use relative paths to input files

The title says it all.

Improve reproducibility of stability experiments

Generated run scripts should include a comment with the command used to generate this script and the version of the Python script that generated it (obtained using git describe).

This will make it easier to reproduce issues discovered when running these scripts.

The code generating PBS job headers should take the job name as an argument

Right now all the jobs on Pleiades are called "cfd". Meaningful names would make it easier to interpret qstat output, for example.

See make_batch_header:

pism-gris/resources/resources.py

Line 967 in f87f39c

def make_batch_header(system_name, n_cores, walltime, queue):

Put the config file in the output directory instead of the current directory

@aaschwanden: Right now this script (see the code below) puts the generated config file in the current directory. Do you think it would make more sense to put it in the output directory specified using the command-line option --o_dir? (I think it would.)

pism-gris/stability/lhs_ensemble.py

Lines 149 to 153 in 5d98cd5

	pism_config = 'init_config'
	pism_config_nc = abspath("./" + pism_config + ".nc")
	cmd = "ncgen -o {output} ../config/{config}.cdl".format(output=pism_config_nc, config=pism_config)
	sub.call(shlex.split(cmd))

Make sure that group members can write to output directories used by job scripts

We can automatically run chmod g+rwx when they are created.

Improve naming of files containing captured PISM output

Right now they have names like job_1.3251934.pbspl1.nas.nasa.gov.

Adding a few characters identifying a run to the file name would make it a lot easier to re-submit a job that failed. (Something like job_rcp26_id001_1.3251934.pbspl1.nas.nasa.gov would do fine.)