FargoCPT Bootstrap Notebook¶
What is this?¶
This notebook can be used to setup a FargoCPT simulation in an empty directory.
Why is this useful?¶
This is useful when you want to run a simulation on a cluster which you can access using Jupyter notebooks and want to store a full copy of the code alongside your simulation outputs.
Contents¶
It will do the following things:
clone the repository from github
compile the code
give you a command to run the simulation
print the available data
plot the surface density
First we create a new directory and change to it.
example_name = "300_bootstrap"
example_dir = f"example_dirs/{example_name}"
import os
if not os.path.basename(os.getcwd()) == example_name:
!mkdir -p $example_dir
os.chdir(example_dir)
repo_root = os.path.abspath(os.path.join(os.getcwd(), "../../../"))
print(f"Current working directory: {os.getcwd()}")
print(f"Repository root directory: {repo_root}")
Current working directory: /home/rometsch/repo/fargocpt/examples/example_dirs/300_bootstrap
Repository root directory: /home/rometsch/repo/fargocpt
Downloading the code¶
We will clone only the last commit of the code which is enough to run the simulation and faster to download.
!git clone --depth 1 https://github.com/rometsch/fargocpt code
Cloning into 'code'...
remote: Enumerating objects: 749, done.[K
remote: Counting objects: 100% (749/749), done.[K
remote: Compressing objects: 100% (625/625), done.[K
remote: Total 749 (delta 155), reused 459 (delta 88), pack-reused 0[K
Receiving objects: 100% (749/749), 4.95 MiB | 4.37 MiB/s, done.
Resolving deltas: 100% (155/155), done.
Building the code¶
Make sure the code is built by running make again.
%%timeit -n1 -r1
from sys import platform
if platform in ["linux", "darwin"]:
!make -j 4 -C code/src > make.log
else:
raise RuntimeError(f"Seems like you are not running MacOS or Linux but {platform}. This is unsupported. You are on your own, good luck!")
26.7 s ± 0 ns per loop (mean ± std. dev. of 1 run, 1 loop each)
Preparing a setup file¶
We’ll take the example setup file from the examples directory and modify it in python. If you want to create setup files for a parameter study, just copy the code and make your own setup creator script.
configfile = "setup.yml"
!cp code/examples/config.yml $configfile
We’ll use the ruamel.yaml package to read and write the setup file. This can be set up to preserve comments which is very useful if you want to trace your decisions later on.
try:
import ruamel.yaml
except ImportError:
raise ImportError("Please install ruamel.yaml with `python3 -m pip install ruamel.yaml`")
yaml = ruamel.yaml.YAML()
with open(configfile, "r") as infile:
config = yaml.load(infile)
config["nbody"][1]["accretion efficiency"] = "2"
config["MonitorTimestep"] = 0.314 # monitor scalar files around every half orbit
config["Nmonitor"] = 20 # write a snapshot every orbit
config["Nsnapshots"] = 30 # wirte 10 snapshots
# use very low resolution by setting it to 2 cell per scaleheight, cps
del config["Nrad"]
del config["Naz"]
config["cps"] = 2
with open(configfile, "w") as outfile:
yaml.dump(config, outfile)
Running the code¶
We can start fargo using the python interface, but this runs slower when started from within a Jupyter Notebook compared to being executed from without. Even calling a python script using the shell magic “!” does not speed it up. Calling a python script that does the same job from the command line does not have this issue.
If anyone knows why this is, please let me know!
For a production run, even for further testing, please open a terminal and run the output of the following cell:
cwd = os.getcwd()
cmd = f"cd {cwd} && code/run_fargo -np 1 -nt 4 auto {configfile}"
print(cmd)
cd /home/rometsch/repo/fargocpt/examples/example_dirs/300_bootstrap && code/run_fargo -np 1 -nt 4 auto setup.yml
If you plan to run on a cluster, put something similar into a run script to be queued in the queueing system.
Consider generating your queuing script here so you can just copy this notebook and setup a new simulation.
For the sake of this notebook, we just run the simluation here.
We’ll use the python module from within the directory.
import sys
sys.path = ["code/python"] + sys.path
from fargocpt import run
run(["start", configfile], np=2, nt=1, exe="code/bin/fargocpt_exe", detach=False)
Running command: mpirun -np 2 --report-pid /tmp/tmpcqh1chwn -x OMP_NUM_THREADS=1 code/bin/fargocpt_exe start setup.yml
fargo process pid 1403025
[0] MPI rank # 0 runs as process 1403029
[1] MPI rank # 1 runs as process 1403030
[0] MPI rank # 0 OpenMP thread # 0 of 1 on cpt-kamino
[1] MPI rank # 1 OpenMP thread # 0 of 1 on cpt-kamino
[0] fargo: This file was compiled on Nov 14 2023, 15:52:04.
[0] fargo: This version of FARGO used _GNU_SOURCE
[0] fargo: This version of FARGO used NDEBUG. So no assertion checks!
[0] Using parameter file setup.yml
[0] Computing disk quantities within 5.00000e+00 L0 from coordinate center
[0] BC: Inner composite = reflecting
[0] BC: Outer composite = reflecting
[0] BC: Sigma inner = zerogradient
[0] BC: Sigma outer = zerogradient
[0] BC: Energy inner = zerogradient
[0] BC: Energy outer = zerogradient
[0] BC: Vrad inner = reflecting
[0] BC: Vrad outer = reflecting
[0] BC: Vaz inner = keplerian
[0] BC: Vaz outer = keplerian
[0] DampingTimeFactor: 1.00000e-01 Outer damping time is computed at radius of 2.50000e+00
[0] Damping VRadial to reference value at inner boundary.
[0] Damping VRadial to reference value at outer boundary.
[0] Damping VAzimuthal to reference value at inner boundary.
[0] Damping VAzimuthal to reference value at outer boundary.
[0] Damping SurfaceDensity to reference value at inner boundary.
[0] Damping SurfaceDensity to reference value at outer boundary.
[0] Damping Energy to reference value at inner boundary.
[0] Damping Energy to reference value at outer boundary.
[0] Radiative diffusion is disabled. Using fixed omega = 1.500000 with a maximum 50000 interations.
[0] Indirect Term computed as effective Hydro center acceleratrion with shifting the Nbody system to the center.
[0] Body force on gas computed via potential.
[0] Using FARGO algorithm for azimuthal advection.
[0] Using standard forward euler scheme for source terms.
[0] Grid resolution set using cps = 2.000000
[0] The grid has (Nrad, Naz) = (74, 251) cells with (1.994113, 1.997395) cps.
[0] Computing scale height with respect to primary object.
[0] Using isothermal equation of state. AdiabaticIndex = 1.400.
[0] Viscosity is of alpha type with alpha = 1.000e-03
[0] Defaulting to VanLeer flux limiter
[0] Output information:
[0] Output directory: output/out/
[0] Number of files: 240
[0] Total output size: 0.00 GB
[0] Space Available: 31.18 GB
[0] Initializing 1 RNGs per MPI process.
[0] Warning : no `radii.dat' file found. Using default.
[0] The first 1 planets are used to calculate the hydro frame center.
[0] The mass of the planets used as hydro frame center is 1.000000e+00.
[0] 2 planet(s) initialized.
[0] Planet overview:
[0]
[0] # | name | mass [m0] | x [l0] | y [l0] | vx | vy |
[0] -----+-------------------------+------------+------------+------------+------------+------------+
[0] 0 | Star | 1 | 0 | -0 | 0 | 0 |
[0] 1 | Jupiter | 0.0009546033 | 1 | 0 | -0 | 1.000477 |
[0]
[0] # | e | a | T [t0] | T [a] | accreting | Accretion Type |
[0] -----+------------+------------+------------+------------+------------+----------------+
[0] 0 | 6.368246e-17 | 1 | 6.280188 | 0.999548 | 0 | No Accretion |
[0] 1 | 6.368246e-17 | 1 | 6.280188 | 0.999548 | 2 | Kley Accret. |
[0]
[0] # | Temp [K] | R [l0] | irradiates | rampuptime |
[0] -----+------------+------------+------------+------------+
[0] 0 | 5778 | 0.0046505 | yes | 0 |
[0] 1 | 0 | 4.6505e-05 | no | 0 |
[0]
[0] Using Tscharnuter-Winkler (1979) artificial viscosity with C = 1.410000.
[0] Artificial viscosity is used for dissipation.
[0] Surface density factor: 2.50663
[0] Tau factor: 0.5
[0] Tau min: 0.01
[0] Kappa factor: 1
[0] Minimum temperature: 2.81162e-05 K = 3.00000e+00
[0] Maximum temperature: 9.37206e+94 K = 1.00000e+100
[0] Heating from viscous dissipation is enabled. Using a total factor of 1.
[0] Cooling (beta) is disabled and reference temperature is floor. Using beta = 10.
[0] Cooling (radiative) is enabled. Using a total factor of 1.
[0] S-curve cooling is disabled.
[0] CFL parameter: 0.5
[0] Opacity uses tables from Lin & Papaloizou, 1985
[0] Particles are disabled.
[0] Initializing Sigma(r) = 2.25093e-05 = 200 g cm^-2 * [r/(1 AU)]^(-0.5)
[0] Total disk is mass is 0.000348841 = 6.9366e+29 g.
[0] Writing output output/out/snapshots/0, Snapshot Number 0, Time 0.000000.
[0] Writing output output/out/snapshots/reference, Snapshot Number 0, Time 0.000000.
[0] Writing output output/out/snapshots/1, Snapshot Number 1, Time 6.280000.
[0] Writing output output/out/snapshots/2, Snapshot Number 2, Time 12.560000.
[0] Writing output output/out/snapshots/3, Snapshot Number 3, Time 18.840000.
[0] Writing output output/out/snapshots/4, Snapshot Number 4, Time 25.120000.
[0] Writing output output/out/snapshots/5, Snapshot Number 5, Time 31.400000.
[0] Writing output output/out/snapshots/6, Snapshot Number 6, Time 37.680000.
[0] Writing output output/out/snapshots/7, Snapshot Number 7, Time 43.960000.
[0] Writing output output/out/snapshots/8, Snapshot Number 8, Time 50.240000.
[0] Writing output output/out/snapshots/9, Snapshot Number 9, Time 56.520000.
[0] Writing output output/out/snapshots/10, Snapshot Number 10, Time 62.800000.
[0] Writing output output/out/snapshots/11, Snapshot Number 11, Time 69.080000.
[0] Writing output output/out/snapshots/12, Snapshot Number 12, Time 75.360000.
[0] Writing output output/out/snapshots/13, Snapshot Number 13, Time 81.640000.
[0] Writing output output/out/snapshots/14, Snapshot Number 14, Time 87.920000.
[0] Writing output output/out/snapshots/15, Snapshot Number 15, Time 94.200000.
[0] Writing output output/out/snapshots/16, Snapshot Number 16, Time 100.480000.
[0] Writing output output/out/snapshots/17, Snapshot Number 17, Time 106.760000.
[0] Writing output output/out/snapshots/18, Snapshot Number 18, Time 113.040000.
[0] Writing output output/out/snapshots/19, Snapshot Number 19, Time 119.320000.
[0] Writing output output/out/snapshots/20, Snapshot Number 20, Time 125.600000.
[0] Writing output output/out/snapshots/21, Snapshot Number 21, Time 131.880000.
[0] Writing output output/out/snapshots/22, Snapshot Number 22, Time 138.160000.
[0] Writing output output/out/snapshots/23, Snapshot Number 23, Time 144.440000.
[0] Writing output output/out/snapshots/24, Snapshot Number 24, Time 150.720000.
[0] Writing output output/out/snapshots/25, Snapshot Number 25, Time 157.000000.
[0] Logging info: snapshot 25, monitor 514, hydrostep 4249, time inside simulation 161.619190, dt 4.524e-02, realtime 10.00 s, timeperstep 2.35 ms
[0] Writing output output/out/snapshots/26, Snapshot Number 26, Time 163.280000.
[0] Writing output output/out/snapshots/27, Snapshot Number 27, Time 169.560000.
[0] Writing output output/out/snapshots/28, Snapshot Number 28, Time 175.840000.
[0] Writing output output/out/snapshots/29, Snapshot Number 29, Time 182.120000.
[0] Writing output output/out/snapshots/30, Snapshot Number 30, Time 188.400000.
[0] -- Final: Total Hydrosteps 4909, Time 188.40, Walltime 11.56 seconds, Time per Step: 2.35 milliseconds
0
Data¶
Let’s check whether there is data.
from fargocpt import Loader
l = Loader("output/out")
l
Loader
====================
| output_dir: output/out
| snapshots: 0 ... 30
| special_snapshots: ['reference']
| snapshot_time: 0.0 5.02257e+06 s ... 188.4 5.02257e+06 s
| monitor_number: 0 ... 600
| units: Units
| target_units = None
| gas: Hydro
| nbody: Nbody
| params: Params
| particles = None
====================