Running Optab example

February 15, 2026 ยท View on GitHub

CAUTION

It is advisable to begin with a minimal number of layers and closely observe the memory consumption before proceeding to full-scale calculations. For instance, in the sample execution below, the peak memory consumption can reach approximately 1.5GB or more, varying with the level of parallelization.

This section details the creation of opacity tables using a chemical abundance table from FastChem. Ensure FastChem is set up before proceeding. The variables $FASTCHEM and $OPTAB denote the root directories of FastChem and Optab, respectively.

  1. Generate a Chemical Abundance Table with FastChem
    Follow the instructions in work/FastChem-lnk_interpolate_dev/README.md. By default, the grid is 21 x 21 and evenly spaced in the log T - log P plane (see the image below). (If you plan to use an existing chemical-equilibrium dataset, see eos/README.md.)
    Warning: FastChem built with quadruple precision does not run correctly on arm64 Macs due to ABI constraints. arm64 Mac users must run FastChem on a separate x86_64 platform; see Step 2 for a fallback.

  2. Convert FastChem Output to an optab-compatible HDF5 File

    cd $OPTAB/eos/FastChem
../src/convert_Fastchem $FASTCHEM/output/sample.dat

    If you do not have access to an x86_64 platform but still want to test optab, a sample FastChem output is available at $OPTAB/work/FastChem-lnk_interpolate_dev/output/sample.dat. In that case, run:

    ../src/convert_Fastchem $OPTAB/work/FastChem-lnk_interpolate_dev/output/sample.dat

    Visualize the converted HDF5 file sample.h5 using the following Python script:

    python3 ../python/eos.py sample.h5 mmw --syms=100

  3. Execute optab with the sample.sh Script

    cd $OPTAB/sample/
bash sample.sh

    This example takes approximately 68 minutes on a 3.3 GHz 12-Core Intel Xeon W processor using 12 MPI processes.

    NOTE
    Adjust the following parameters to align with your system before running the script:

    • EOS : path to the chemical abundance table in the HDF5 format created in the above step
    • OPTAB : path to the top directory of Optab
    • DATABASE: path to the top directory of Optab database
    • MPIBIN: path to the directory where mpirun is installed
    • jprc: number of MPI processes

    Consider reviewing and modifying additional parameters within the script as shown below:

    ...

#####
export EOS='/Volumes/Storage/optab/eos/FastChem/sample.h5'
export OPTAB='/Volumes/Storage/optab/'
export DATABASE='/Volumes/Storage/optab/database/'
export MPIBIN='/opt/local/bin'
...

##### SELECT A SINGLE LINE-SOURCE FOR EACH MOLECULAR ISOTOPOLOGUE
cat <<EOF > input/species_id.dat
     Species   Isotopologue  HITRAN    HITEMP        Exomol
1    H2O       1H2-16O       0 HITRAN  1 HITEMP      0 POKAZATEL  0 BT2
1    H2O       1H2-18O       0 HITRAN  1 HITEMP      0 HotWat78
1    H2O       1H2-17O       0 HITRAN  1 HITEMP      0 HotWat78
...
56   H3+       1H3_p         1 HITRAN                0 MiZATeP
56   H3+       1H2-2H_p                              0 ST
999  dummy     dummy                                  0 dummy
EOF

##### SELECT OPACITY SOURCES TO BE CONSIDERED (1: SELECTED, 0: NOT SELECTED)
cat <<EOF > input/fort.5
&switches ! selection of opacity sources
line_molecules = 1           ! molecular lines
line_kurucz_gfpred = 1       ! Kurucz gfpred lines
line_kurucz_gfall = 1        ! Kurucz gfall lines
rayleigh_scattering_h2 = 1   ! Rayleigh scattering by H2
rayleigh_scattering_he = 1   ! Rayleigh scattering by He
rayleigh_scattering_h = 1    ! Rayleigh scattering by H
electron_scattering = 1      ! electron scattering
cia = 0                      ! Collision-induced absorption (EXPERIMENTAL)
photoion_h2 = 1              ! Photoionization by H2
photoion_topbase = 1         ! TOPbase photoionization (exclusive with mathisen)
photoion_mathisen = 0        ! Mathisen photoionization (exclusive with topbase)
photoion_verner = 1          ! Verner photoionization
photoion_h_minus = 1         ! Photoionization by H-
brems_h_minus = 1            ! Bremsstrahlung by H-
brems_h2_minus = 1           ! Bremsstrahlung by H2-
brems_atomicions = 1         ! Bremsstrahlung by atomic ions
/
&cutoffs ! for line evaluation
cutoff0_voigt = 1d2 ! cutoff for Voigt profile [in wavelenth(cm^-1)]
cutoff0_gauss = 3d0 ! cutoff for Gaussian profile [in gaussian width]
delta_crit = 1d-4   ! criteria for discarding weak lines
delta_voigt = 1d0   ! criteria for adopting Voigt profile
/
&radtemp ! radiation temperature for 2-temp Planck-mean opacity
temp2 = 6000d0
/
&grid_log_const ! logarithmic wavenumber grid
k_total = 100000 ! total number of grid points
grd_min = 1d0    ! min value of wavenumber grid
grd_max = 9d0    ! max value of wavenumber grid
/
&mpi_decomp ! total number of MPI processes = kprc * jprc * mprc * jprc
kprc = 1  ! number of processes in wavenumber grid (EXPERIMENTAL)
lprc = 1  ! number of processes in line loop (EXPERIMENTAL)
mprc = 1  ! number of processes in reading line-block loop (EXPERIMENTAL)
jprc = 8  ! number of processes in layer loop
/
...
EOF

...

  4. Visualize optab Output Files
    For each set of (T, P, Ni), optab produces an HDF5 file, mono_?????.h5 in the directory $OPTAB/sample/sample/output/, containing:

    • Rosseland-mean opacity
    • Planck-mean opacity
    • two-temperature Planck-mean opacity
    • monochromatic absorption and scattering opacities

    Use the opac.py and mono.py scripts within $OPTAB/sample/python to visualize the data as follows. To see the available script arguments, execute them without any arguments.

    python3 python/opac.py sample ross 150
python3 python/opac.py sample pla 150
python3 python/opac.py sample pla2 150
python3 python/mono.py sample 30
python3 python/mono.py sample 300