He+CO2, Fermi resonance, CC and CS

Coupled channels calculation for He+CO₂, where CO₂ bend (v2) and symmetric stretch (v1) are included.

Software versions (git commits):

gsci_lib6:   30015e4 - Sun, 7 Apr 2024 12:33:42 +0200
HeCO2_fermi: 95b56c6 - Sun, 7 Apr 2024 13:39:58 +0000
gb_sql:      6251f2d - Wed, 3 Apr 2024 00:19:37 +0200

Quick start

First, pick a NAMESPACE. Here, we use the name of the directory with the jobs as the namespace: HeCO2fm_v16:

git clone lilo.science.ru.nl:/vol/thchem/git/HeCO2_fermi HeCO2fm_v16
cd HeCO2fm_v16/run1       # here are the jobs

A json file with input paramters is required. We start from par15.json:

cp par15.json par16.json
# vi par16.json                       # edit parameter file

Next, edit par16.json to set the required parameters. To run:

./bat1.sh par16.json NS:HeCO2fm_v16
# The script echos the input
# You will get some questions - simply press <enter> for the default:

Reset journal HeCO2fm_v16_JOURNAL [Y/n] ?

Fd: Re-initialize table "files" [Y/n]

# job2.sh and job3.sh will be run by bat1.sh, should take about a minute

Three files will have been defined, to check enter:

Fd ls
Fd ls -l              # gives more detail

One of the files is the .SETTINGS file, which defines some bash variables:

cat .SETTINGS
# this shout give
TERM_TITLE=HeCO2fm_v16
export TODO=HeCO2fm_v16_TODO
export FILES=HeCO2fm_v16_FILES
export JOURNAL=HeCO2fm_v16_JOURNAL

If you use gi_prompt.sh, then the title of your terminal should change to $TERM_TITLE when you enter the directory with the .SETTINGS file.

Define more files and setup table with jobs:

bat2.sh
bat3.sh               # press enter to this question:

Fd: Re-initialize table "HeCO2fm_v16_TODO" [Y/n]?
#
Jb ls                 # show table with jobs
Jb ls -l              # with more details
Jb stat               # count number of jobs for each value of status

Next, we run a single job to test:

./bat4.sh |& tee bat4.out     # output appears on screen and is written to bat4.otu

In another terminal go to the same directory, and monitor progress:

# (other terminal)
cd HeCO2fm_v16/run1
# To monitor progress use any of these
./w.sh                        # uses watch to update, shows status
./started.sh                  # shows started jobs
Jb stat                       # shows status (once)
Jb ls -w status = started +c Jtot,parity,iE
Fd ls                         # show files
Log ls                        # show log files
cd scratch                    # this links is created by bat1.sh
ls -l                         # shows directories with output
cd  48even                    # If first job has Jtot=48 and parity=-1
ls -l
# job4_331.out                # should be there

If the status of the jobs turns into done then submit the other jobs using slurm:

./sbat4.sh                    # will submit one job for each Jtop and parity
                              # and loop over energies
squeue -u me                  # show slurm jobs

Use the above commands to monitor progress. Use the Mb command to check use of cluster. To start extra jobs, use:

./sbat4.sh njob:100            # start hundred extra jobs

Other settings can be passed to this script:

./sbat4.sh mem:4G loop:4 njob:10 w:cn21

Files with cross sections can be seen with:

Fd ls %CS_%.dat               # "%" is SQL wildcard

The underscore _ is also a SQL wildcard, it may be escaped with a backslash:

Fd ls %CS\_%.dat

When all jobs are done the script alldone.sh is started. This script will runs sbat6.sh, which will start a bat6.sh on each host that has output. This job collects all cross sections computed on that host into a spec. They can be seen with:

Fd ls %Call%

When all jobs are done, output is written to alldone.log.

To collect all cross sections in a single file on the current host:

./bat7.sh
./csall.sh            # this shows where on /scratch Call is written

Instructions for plotting will be added soon.