Anders' readme
SUMO - Supernova MOnte Carlo code.
Papers : Jerkstrand, Fransson & Kozma 2011, see also Jerkstrand et al. 2012, Jerkstrand et al. 2015 (A&A).
This file created by A. Jerkstrand anders@mpa-garching.mpg.de.
PREPARE THE ENVIRONMENT¶
-
Define the environmental variable "DD" in your login script (eg .bashrc) For example in bash:
DD should point to some folder for scratch reading and writing (sometimes large files are created).export DD=/disk/myscratchfolder
-
Go to the DD folder, and type
mkdir mcphotons mkdir mcphotons/radfield mkdir modelspectra - The code needs access to the LAPACK and BLAS mathematical libraries. If you do not
have these installed, download them from www.netlib.org and install. Then, open the sumo Makefile and edit the line
LIBS = ...To give the path to your libraries. If you have static libraries of BLAS and LAPACK somewhere, edit .bashrc toand mark inexport SUMOLIBS='..../blas_LINUX.a ..../lapack_LINUX.a' (replace with relevant path and names)
LIBS=$(SUMOLIBS)in the Makefile
To use MKL library files for BLAS and LAPACK, use
LIBS = $(MKLROOT)/lib/em64t/libmkl_lapack95_lp64.a -Wl,--start-group $(MKLROOT)/lib/em64t/libmkl_intel_lp64.a $(MKLROOT)/lib/em64t/libmkl_intel_thread.a $(MKLROOT)/lib/em64t/libmkl_core.a -Wl,--end-group -L$(MKLROOT)/lib/em64t -liomp5 -lpthread
COMPILATION¶
The code compiles with a Makefile.
Compilation is by default with -O3 optimization, edit this in
FFLAGS=... if needed.
Go into the folder where you unpacked the code and run
make all
- "suse" : Calculates temperatures, excitation and ionization statistical equilibria, and emissivities (radiative terms and ionization balance fixed)
- "sumo" : Runs Monte Carlo simulation of radiation field (temperatures, excitation and ionization statistical equilibria fixed)
- "suib" : Calculates ionization statistical equilibria (temperatures, radiative terms fixed).
These three are iterated until convergence is achieved. Information on how to construct wrapped and test for convergence to be added..
Ex suse (don't run yet, see first preparation of input files below):¶
mpirun -np 19 suse example file_modelparameters_example y n 500 10000 1 > outfile_suse
Run model with name example, and run-file named file_modelparameters_example. The y
means compute the gamma-ray deposition (only needed once),
the n means this is not an iterative run (but the first one),
500 means start at 500 A, 10000 means end at 10000 A, 1 means
use resolution of 1 A (at 1000 A). The numer of processes (19 here) has
to be the number of zones in the model plus one; so here the
model must have 18 zones.
In the first iteration one should
use the combination "y n", whereas later on "n y" should be used.
Ex sumo (always after suse):¶
mpirun -np 100 sumo example file_modelparameters_example 500 10000 1 100000000 5 5 50 > outfile_sumo
file_modelparameters_example, 500, 10000, and 1 mean same as before (keep the same as
when running suse).
100000000 means use a (fiducial) total number of 100000000 packets
(this is, in fact, not really the total number of packets, let it
be 100000000 for now). 5 5 50 means, use at least 5 packet per optical/infrared energy bin,
at least 5 packets in the UV, and never use more than 50 packets.
For sumo the number of processors can be anything (100 in the example). Scaling should be relatively linear with number of processors (but depends a lot on the number of packets).
Then run
./gather.sh
Ex suib (always after sumo):¶
mpirun -np 19 suib file_modelparameters_example y > outfile_suib
y means its an iterative run.
PREPARING INPUT FILES¶
The code needs 2 input files to run, the explosion model file, and the model parameters file.
1) The explosion model file.
An example file is provided (DATA/expmodels/file_explmodel_example).
The format of these files should be as follows:
Line 1 Name of model and then name of various zones 2 Mass of zone [Msun] 3 Inner velocity [km/s] 4 Outer velocity [km/s] 5 Filling factor. Is normally 1. But the code contains an option to macroscopically mix the first Ncore zones. The velocity limits for all these are then the inner velocity of zone 1 and the outer velocity of zone Ncore. The filling factor for each of the first Ncore zones can then be smaller than 1. 6 Density at 100 days [g cm^-3]. Not used by code, can put values to zero if wished. 7 --------- 8 56Ni+56Co mass fraction (at t=0, before any decay has occurred). 9 57Ni mass fraction (at t=0, before any decay has occured). 10 44Ti mass fraction (at t=0, before any decay has occured). 11 --------- 12 H mass fraction 13 He -- " -- 14 C -- " -- 15 N -- " -- 16 O -- " -- 17 Ne -- " -- 18 Na -- " -- 19 Mg -- " -- 20 Si -- " -- 21 S -- " -- 22 Ar -- " -- 23 Ca -- " -- 24 Fe -- " -- 25 Co -- " -- 26 Ni -- " -- 27 Ti -- " -- 28 Al -- " -- 29 Cr -- " -- 30 Mn -- " -- 31 Sc -- " -- 32 V -- " --
2) The model parameters file.
This file should be created in the code execution directory.
An example file (file_modelparameters_example) is provided.
The philosophy if the code design is to maintain a core source code that should not have to be recompiled depending on which models or parameter space to investigate. If you wish to run 10 different models with 10 different dust parameters, therefore create 10 different "run settings" files, and set up a script to process these in turn.
The format of file_modelparameters is as follows:
Row
1 Time [days since explosion]
2 Distance to SN [MPc]
3 Dust optical depth [dim-less]
4 Explosion model file (see above)
5 Number of zones
6 Ncore : Number of zones to macroscopically mix ("0" for no macroscopic mixing).
7 Ndust : Number of zones to apply dust opacity to
8 Nblobs : Number of clumps in macroscopically mixed model
9 --------------------------
10 VOID
11 VOID
12 VOID
13 VOID
14 VOID
15 VOID
16 VOID
17 VOID
18 -----------------------
19 e+ leakage : 1 => free streaming (no magnetic field), 0 => on-the-spot absorption (complete magnetic confinement).
20 Npis : Maximum number of levels to include photoionization from. 0 => no photoionizations from excited states. Typical value : 50
21 PHOTOEXC : Include photoexcitation in NLTE solutions? 2 => Yes. 1 => Only deexcitations. 0: No. Typical value : 2.
22 beta_Lya : Probability of Lya branching into overlapping metal lines. Typical value : 5E-9.
23 beta_Lyb : Probability of Lyb branching into overlapping metal lines. Typical value : 2.1E-5.
24 tau_tresh : Ignore line optical depths below this value. Typical value : 1E-3.
25 sf_tresh : Ignore computation of non-thermal excitation and ionization for elements with mass fractions below this value. Typical value 1E-3.
26 CTON : .T. => Charge transfer reactions on. .F. => Off. Typical value : .T.
27 MOLSW : 0 => No molecules. 1 => Include molecules. CURRENTLY ONLY 0 supported.
28 OLI : Allow only photoionization by local radiation field? .F. => No, .T.=> Yes. Default : 0.
29 REALRECCONT_H : Use real free-bound continua for HI? .T. => Yes. .F. => No (then uses box-shaped approximation as for other elements). Typical value : .T.
30 FORCECAII : Force Ca to be in the Ca II form everywhere? .F. => no, .T. => yes. Typical value : .F.
31 VOID
32 VOID
33 DOPRINT : Print a lot of output? .T. => Yes, .F. => No. Typical value: .F.
34 -----------------------------------
35 HI : Entry 1 : Allow this atom to emit radiation? Entry 2: Allow this atom to absorb radiation? Entry 3 : How many levels to compute NLTE solutions for? (0=all).
...
81 H2 ... (same as above)
EXAMINING THE OUTPUT¶
Spectrum
The main output of the code is the model spectrum, which goes to the file $DD/modelspectra/spectrum.dat*IDNUMBER*.
The columns in this file are as follows:
1 Wavelength [A] 2 F_lambda [erg s^-1 cm^-1 A^-1] 3 F_lambda, smoothed with Vsmooth 4 component1.. 5 component2.. .. ..
Temperatures (from suse)
Written to ./out/plts/teqs.dat*SHELLNUMBER*
Excitation (from suse)
Written to ./DATA/EXC_CUBE.dat*SHELLNUMBER*
Ionization (from suib)
Written to ./out/ionization/ionfracs.dat*SHELLNUMBER*