This documentation aims to help with the reproduction or modification of the results of my thesis. The goal is to calculate the scattering rate of DM-electron scattering. The main software used is QuantumESPRESSO and QEdark:
This is a short incroduction, in how to use the code QEdark_1k.
It is a modification to QEdark with the aim of faster execution on Tesla.
All executables can be found under
/kalinka/home/lukwata/software/My_espresso-5.1.2/bin/
A example is in
/kalinka/home/lukwata/software/QEdark_1k/QEdark_examples/ex_1k
Copy it to your home folder
cp -r /kalinka/home/lukwata/software/QEdark_1k/QEdark_examples/ex_1k ~/
and change in it.
There you find the file dm.in in which you can change the masses of the DM particales and the escape velocity.
There are also the two files ge.243.in and ge.4.in. In ge.243.in a 243 k-point mesh is defined, in ge.4.in a 4 k-point mesh is defined (it is really a reduced 27 k-point mesh, but that does not matter here).
A file with a description of the most important variables can be found here.
Let's use the 4 k-point mesh to generate the crystal formfaktor. The steps for ge.243.in are the same, but the execution time would be much longer.
For the sake of convenience I created a bash-skript qedark_1k.sh.
It lets you control all the following steps.
Before starting the script you need to set the path to the executables and the names of the input-files in qedark_1k.sys.
The content of qedark_1k.sys looks like this:
kpoints=243
GE_IN="ge.243.in"
dm_in="dm.in"
# use absolute path
pseudo_dir="/kalinka/home/lukwata/testQEdark/"
job_f2=job_f2.pbs
job_multi=job_multi.pbs
export EXEC_SUM="/kalinka/home/lukwata/software/test_compile/espresso-5.1.2/bin/qedark_1k_sum.x"
export EXEC_F2_1K="/kalinka/home/lukwata/software/test_compile/espresso-5.1.2/PW/src/pw_dark_1k.x"
export EXEC_MULTI="/kalinka/home/lukwata/software/test_compile/espresso-5.1.2/bin/qedark_1k_multimass.x"
export GE_IN
export KPOINTS="$kpoints"There are some old paths for the executables. So change them to the right ones:
export EXEC_SUM="/kalinka/home/lukwata/software/My_espresso-5.1.2/bin/qedark_1k_sum.x"
export EXEC_F2_1K="/kalinka/home/lukwata/software/My_espresso-5.1.2/PW/src/pw_dark_1k.x"
export EXEC_MULTI="/kalinka/home/lukwata/software/My_espresso-5.1.2/bin/qedark_1k_multi.x"Because we want to use the 4 k-point mesh change the corresponding lines to
GE_IN="ge.4.in"
kpoints=4The path of the pseudo potential is also wrong. Change it to
pseudo_dir="/<your homefolder>/ex_1k/"
Now we can start with the calculation of the crystal formfactor. Run
bash ./qedark_1k.shThe output you will see is:
##########################################################
Control the calculation with qedark_1k over this script.
You can set the path to executables, the names of the input files
and the number of k-points in qedark_1k.sys.
Other settings can be done in the dm.in and ge.##.in
##########################################################
1) Create/Update tmp folders for formfactor calculation and copy dm.in, ge.##.in and job_f2.pbs inside.
2) Run pw_dark_1k.x. Start jobs parallel on Tesla.
3) Sum output of parallel jobs. The output-file is C.dat. It contains the formfactor.
4) Run qedark_1k_multimass.x
5) Clean unused output of pw_darK_1k
6) Remove the tmp folders used for formfactor calculation.
7) Quit
Please enter your choice:
Choose 1 and hit ENTER. This will create 4 folders ge_### , in each a partial sum of the squared formfaktor will be calculated.
Then choose 2. This will send 4 jobs to Tesla. The jobs put ther output in ge_###. Now you can hit 7 and take a coffee brack until all jobs are finished. To check if they are finished run
qstat
If you get no output they are finished. If everything went right there should be a file K.###.dat in every ge_### folder.
If something went wrong look in a folder ge_### for the job_f2.pbs.e######## file.
When changing anything in dm.in, ge.#.in or qedark_1k.sys rerun option 1 with qedark_1k.sh.
So finally everything went well, and there are 4 K.###.dat files.
The content of these files need to be summed up.
Therefor run qedark_1k.sh and choose 3.
This creates the file C.dat which contains the squared formfactor.
You can use 5 or 6 to clean your folder a little bit befor the next step. But be carefull 6 will delet all the partial sums so be shure you have your C.dat file
qedark_1k_multimass.x will use the DM-masses set in dm.in and C.dat to calculate the scattering for each mass.
To start it, again run qedark_1k.sh and choose 4.
It wil only take less then a minute to run then there is a file
C.###.dat for every DM-mass.
The C.###.dat files contain the scattering rate binned with the resolution set in dm.in.
A snippet of a C.###.dat file (less rows then the real one):
01 01 0.139193079E-002 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.390599274E-005
02 01 0.103375018E-002 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.231179746E-005
03 01 0.245999056E-002 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.171677492E-005
01 02 0.139193079E-002 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.390599274E-005
02 02 0.103375018E-002 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.231179746E-005
03 02 0.245999056E-002 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.171677492E-005
01 03 0.139193079E-002 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.390599274E-005
02 03 0.103375018E-002 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.231179746E-005
03 03 0.245999056E-002 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.000000000E+000 0.171677492E-005</verbatim>
The first column stands for the DM-formfaktor ( 1, 1/q, 1/q^2). The second column represents the month and so annual modulation of the velocity of the earth
- December ve_mod = -15kmps
- March ve_mod = 0
- June ve_mod = 15kmps
The third column is the total rate divided by a prefactor the following bins represent the binned rate. If you have not changed anything else then said above the files should have 503 columns.
- plot script