Skip to content

Latest commit

 

History

History
331 lines (303 loc) · 16.8 KB

NAMELIST_GUIDE.md

File metadata and controls

331 lines (303 loc) · 16.8 KB

ALaDyn namelist guide

Following here is a brief description of the new input.nml file required for the simulation parameters definition, through an example.

GRID namelist block

&GRID
 nx               = 7168,
 ny               = 3584,
 nz               = 1,
 ny_targ          = 3500,
 k0               = 100.0,
 yx_rat           = 2.0
/
  • nx is the number of grid points in the x direction
    • for PWFA simulations, it is required that this number contains the number of cpus of the y and z domains defined by nprocx, nprocy and nprocz in order to be sure that FFTs are working as expected
  • ny is the number of points in the y direction
    • in order to be sure that the simulation is working as expected, it is better to always use a number that contains nprocx, nprocy and nprocz
  • nz is the number of points in the z direction; for a 2D simulaton, use nz = 1
    • in order to be sure that the 3D simulation is working as expected, it is better to always use here a number that contains the number of cpu assigned to split the z domain (mpi_ntot/pey)
  • ny_targ is the transverse size of the target, expressed in number of grid cell filled with particles (if the simulation is a 3D one, the target is a square in the xy plane)
  • k0 defines the resolution, being the number of points per μm along x, so that Δx = 1/k0 [μm]; please remember to set a value small enough to solve the skin depth
  • yx_rat is the ratio between the resolution along x and y: Δy = yx_rat/k0 [μm].
  • zx_rat is the ratio between the resolution along x and z: Δz = zx_rat/k0 [μm]; the resolution along z is the same as for y if this parameter is not defined.

With those parameters, the full box size (in μm) is: Lx = nx / k0, Ly = yx_rat * ny / k0, Lz = zx_rat * nz / k0

SIMULATION namelist block

&SIMULATION
 LPf_ord          = 2,
 der_ord          = 2,
 str_flag         = 0,
 iform            = 0,
 model_id         = 1,
 dmodel_id        = 3,
 ibx              = 0,
 iby              = 0,
 ibz              = 0,
 ibeam            = 1
/
  • Lpf_ord is the integration scheme order. Lpf_ord=2 for standard leap-frog, Lpf_ord_ord=4 for RK4
  • der_ord is the order of the finite difference scheme. der_ord=2 or 3 for Lpf_ord=2, der_ord=4 for RK4
    • der_ord=3 assures optimized wave (Laser) propagation in under dense plasma.
  • str_flag has three possible different values:
    • 0 for uniform grid
    • 1 to enable stretching along transverse axes
    • 2 to enable stretching along all axes, but for x it is only on the right side of the box (need testing)
  • iform has two possible different values:
    • 0: Esirkepov's scheme for charge conservation (particle by particle) (to be preferred)
    • 1: Esirkepov's scheme for charge conservation (inverted on grid along x) (not allowed for MPI decomposition along the x coordinate)
    • 2: no charge preserving (better energy conservation assured)
  • model_id has five possible different values
    • 1 laser is p-polarized
    • 2 laser is s-polarized
    • 3 laser is circularly polarized
    • 4 laser is described by its envelope approximation model
    • 5 the driver is an electron beam. Its description must be given in a separate namelist block
    • 6 the driver is a proton bunch. Its description must be given in a separate namelist block
  • dmodel_id has five possible different values
    • 1 uniform - the simulation is done using nsp species (electrons, Z_1,Z_2,Z_3) all distributed along the target x-profile
    • 2 empty model
    • 3 preplasma - the simulation is tested in this configuration: three nsp=3 species (electrons, Z_1, Z_2), with a preplasma made of Z_1 + e, a bulk made of Z_1 + e and a contaminant layer made of Z_2 + e
    • 4 foam - the simulation is tested in this configuration: three nsp=3 species (electrons, Z_1, Z_2), with a foam made of Z_2 + e, a bulk made of Z_1 + e and a contaminant layer made of Z_2 + e
    • 5 nanowires target with size parameters lpy(1),lpy(2) and species (e+Z_1) + a uniform bulk (e+Z_2)
    • 6 nanotubes target with size parameters lpy(1),lpy(2) and species (e+Z_1) + a uniform bulk (e+Z_2)
  • ibx, iby, ibz are the boundary conditions:
    • 0 open
    • 1 reflective
    • 2 periodic
  • ibeam: for PWFA (model_id #5/#6):
    • 0
    • 1
    • 2

TARGET_DESCRIPTION namelist block

&TARGET_DESCRIPTION
 nsp              = 3,
 nsb              = 1,
 ion_min(1)       = 4,
 ion_max(1)       = 11,
 atomic_number(1) = 13,
 mass_number(1)   = 26.98,
 ion_min(2)       = 1,
 ion_max(2)       = 1,
 atomic_number(2) = 1,
 mass_number(2)   = 1.0,
 ion_min(3)       = 1,
 ion_max(3)       = 1,
 atomic_number(3) = 1,
 mass_number(3)   = 1.0,
 ionz_model       = 0,
 ionz_lev         = 0,
 t0_pl(1)         = 0.0003,
 t0_pl(2)         = 0.0,
 t0_pl(3)         = 0.0,
 t0_pl(4)         = 0.0,
 np_per_xc(1)     = 12,
 np_per_xc(2)     = 4,
 np_per_xc(3)     = 4,
 np_per_xc(4)     = 4,
 np_per_xc(5)     = 4,
 np_per_xc(6)     = 4,
 np_per_yc(1)     = 9,
 np_per_yc(2)     = 3,
 np_per_yc(3)     = 4,
 np_per_yc(4)     = 4,
 np_per_yc(5)     = 4,
 np_per_yc(6)     = 4,
 lpx(1)           = 0.0,
 lpx(2)           = 0.0,
 lpx(3)           = 2.0,
 lpx(4)           = 0.0,
 lpx(5)           = 0.08,
 lpx(6)           = 0.0,
 lpx(7)           = 0.01,
 lpy(1)           = 0.0,
 lpy(2)           = 0.0,
 n_over_nc        = 100.0,
 np1              = 1.0,
 np2              = 10.0
/
  • nsp is the number of species (be careful and coherent with dmodel_id)
  • nsb
  • atomic_number(i) are the atomic number (Z) that define each element species
  • ion_min(i) are the initial ionization status of the element species
  • ion_max(i) are the maximum ionization status of the element species
  • mass_number(i) are the mass number (A) that define the exact isotope of a given atomic_number(i). Here following you can find the only elements known by ALaDyn
Hydrogen  (atomic_number = 1)  - mass_number = 1.0
Helium    (atomic_number = 2)  - mass_number = 4.0
Lithium   (atomic_number = 3)  - mass_number = 6.0
Carbon    (atomic_number = 6)  - mass_number = 12.0
Nitrogen  (atomic_number = 7)  - mass_number = 14.0
Oxygen    (atomic_number = 8)  - mass_number = 16.0
Neon      (atomic_number = 10) - mass_number = 20.0
Aluminium (atomic_number = 13) - mass_number = 26.98
Silicon   (atomic_number = 14) - mass_number = 28.09
Argon     (atomic_number = 18) - mass_number = 39.948
Titanium  (atomic_number = 22) - mass_number = 47.8
Nickel    (atomic_number = 28) - mass_number = 58.7
Copper    (atomic_number = 29) - mass_number = 63.54
  • t0_pl(i), with i from 1 to 4, are the initial temperatures in MeV for the different species
  • np_per_xc(i) Please note that if np_per_xc(1) =0, it just means to ignore the plasma and simply do a laser propagation simulation.
    • dmodel_id=1 : i=1 indicates the number of electrons, i=2 the number of macroparticles of Z_1 species, i=3 the number of macroparticles of Z_2 species and i=4 the number of macroparticles of Z_3 species.
    • dmodel_id=3,4 : i=1,2 are the number of electrons and ions per cell along x/y in the bulk, i=3,4 refer to the front layer and i=5,6 to the contaminants.
  • np_per_yc(i): the same as np_per_xc, this describes the number of particles per cell along transverse directions (valid also for z for 3D simulations)
  • lpx(1) is the length [μm] of the upstream layer (foam or preplasma), having density n1/nc
  • lpx(2) is the length [μm] of the ramp (linear or exponential depending on the mdl) connecting the upstream layer with the central one (made with bulk particles)
  • lpx(3) is the length [μm] of the central layer (bulk), having density n2/nc
  • lpx(4) is the length [μm] of the ramp (linear), connecting the bulk with the contaminants (made with bulk particles)
  • lpx(5) is the length [μm] of the downstream layer (contaminants), having density n3/nc
  • lpx(6) is the angle α of incidence, between the laser axis and the target plane
  • lpx(7) is the offset [μm] between the end of the laser and the beginning of the target (if zero, the target starts right at the end of the laser pulse). The offset is calculated before laser rotation, so mind the transverse size if lx(6) ≠ 0, in order to avoid laser initialization inside the target.
  • lpy(1) defines the wire size [μm]
  • lpy(2) defines the distance [μm] between wires (interwire size)
  • n_over_nc is the density in the central layer (bulk)
    • LWFA case: density is in units of critical density
    • PWFA case: the density is in units of (a nominal value) nc=1e18 cm-3
  • np1 is the density in the upstream layer (foam/preplasma)
  • np2 is the density in the downstream layer (contaminants)
  • ionz_lev: if set to 0, we disable ionization; if 1, only one electron can be extracted per ion, if accessible, per timestep; if 2, it ionizes all the accessible levels in a single timestep
  • ionz_model describes the various ionization models:
    • 1 (pure ADK as in chen et al (2013), the best one for wake sims)
    • 2 (ADK averaged over cycles, as in chen et al (2013), W_AC=<W_DC>, best for envelope simulations)
    • 3 (W_AC+BSI, added barrier suppression ionization)
    • 4 (Minimum between ADK and BSI ionization values. Here the ADK value is computed averaging on m, the magnetic quantum number of the ionized electrons as in Alistair Lawrence-Douglas PhD thesis)

LASER namelist block (only for ibeam=1)

&LASER
 G_prof         = .true.,
 nb_laser       = 1,
 t0_lp          = 16.5,
 xc_lp          = 16.5,
 tau_fwhm       = 33.0,
 w0_y           = 6.2,
 a0             = 3.0
 lam0           = 0.8
 lp_delay       = 20.59
 lp_offset      = 0,
 t1_lp          = 200.0,
 tau1_fwhm      = 24.74,
 w1_y           = 3.5,
 a1             = 0.45,
 lam1           = 0.4,
/
  • G_prof logical flag: if true the pulse is temporally gaussian, if false it has a cos^2 shape
  • nb_laser number of (identical) laser pulses injected
  • xc_lp is the position (in μm) along x of the central point of the laser envelope
  • t0_lp is the distance (in μm) after which the laser is focused; so the focus is at xf = xc_lp + t0_lp
  • tau_fwhm is the FWHM pulse duration (in fs)
  • w0_y is the transverse waist FWHM (in μm)
  • a_0 is the laser adimensional parameter: a0=eA/(me c2) of all the pulses injected
  • lam0 is the laser wavelength (in μm) of all the pulses injected
  • lp_delayis the distance between the center of every injected laser pulse
  • lp_offset is the distance between the center of the last injected pulse and the center of another different pulse injected (if different from 0)
  • t1_lp same as t0_lp, but for the additional pulse injected with lp_offset!=0
  • tau1_fwhm same as tau_fwhm, but for the additional pulse injected with lp_offset!=0
  • w1_y same as w0_y, but for the additional pulse injected with lp_offset!=0
  • a1 same as a0, but for the additional pulse injected with lp_offset!=0
  • lam1 same as lam0, but for the additional pulse injected with lp_offset!=0

MOVING_WINDOW namelist block

&MOVING_WINDOW
 w_sh           = 20,
 wi_time        = 100.0,
 wf_time        = 150.0,
 w_speed        = 1.0
/
  • w_sh is the number of time steps after which the moving-window is called, repeatedly
  • wi_time is the absolute time at which the window starts moving. nb: in order to block the MW from a certain simulation, the only tested path is to set a wi_time greater than the ending time of the simulation itself. w_speed=0 should also work, but is not tested for now
  • wf_time is the absolute time at which the window stops moving
  • w_speed is the speed over c of the moving window

OUTPUT namelist block

&OUTPUT
 nouts          = 10,
 iene           = 200,
 nvout          = 3,
 nden           = 1,
 npout          = 0,
 nbout          = 0,
 jump           = 1,
 pjump          = 1,
 xp0_out        = 0.0,
 xp1_out        = 100.0,
 yp_out         = 20.0,
 tmax           = 100.0,
 cfl            = 0.8,
 new_sim        = 0,
 id_new         = 0,
 dump           = 0,
 L_env_modulus  = .true.,
 P_tracking     = .true.,
/
  • nouts is the number of binary outputs during the relative time of the simulation
  • iene is the number of text outputs during the relative time of the simulation
  • nvout is the number of fields written. For a 2D P-polarized case, we have just 3 fields, Ex, Ey and Bz; in all the other cases there are 6 fields with these IDs: 1=Ex, 2=Ey, 3=Ez, 4=Bx, 5=By, 6=Bz. At each nouts step, every field with ID ≤ nvout will be dumped.
  • nden can have three different values; every output is divided by species on different files:
    • 1: writes only the particle density n on the grid
    • 2: writes also the energy density n ・ gamma on the grid
    • 3: writes also the currents J on the grid
  • npout
    • 1: writes only the electron phase space
    • 2: writes only the proton phase space
    • in general it writes only the phase space of the n-th species, with n=npv; if n=npv>nps, it writes the phase spaces of all the particle species
  • nbout (only for PWFA)
    • 1: writes only the electron phase space
    • 2: writes only the proton phase space
    • in general it writes only the phase space of the n-th species, with n=npv; if n=npv>nps, it writes the phase spaces of all the particle species
  • jump: jump on grid points; if jump=2, it means that each processor is going to write only one point every two. Mind that it's extremely important that jump is a divisor of the number of grid points per processor, otherwise the grid output will be corrupted
  • pjump: jump on particles; if pjump=2, it means that each processor is going to write the phase space of just one particle every two.
  • xp0_out, xp1_out and yp_out only particles contained inside the box defined by xp0 < x < xp1, |y| < yp_out will be printed in the output
  • tmax is the relative time (in μm, because it is multiplied by c) that the simulation is going to be evolved. Being relative, it's going to be added to the time eventually already reached before the restart. To obtain the time in fs, you have to divide it by 0.299792458 [speed of light in μm/fs]
  • cfl is the Courant–Friedrichs–Lewy parameter ( Wikipedia CFL-parameter page )
  • new_sim: identifies if a simulation is new new=0 (data is recreated from initial conditions) or if it is a restart new=1 (dump are going to be read)
  • id_new identifies the starting number of the output files (if new=0) or the last one written in the previous step (if new=1)
  • dump
    • 1: each processor will dump a binary file for every nout in order to enable restart
    • 0: dumps are suppressed
  • L_env_modulus logical flag, only if model_id=4: if true the code generates the absolute value of the laser envelope amplitude, otherwise gives the real and imaginary part in two separate files
  • P_tracking logical flag, if true the particle tracking is enabled

TRACKING namelist block

&TRACKING
 nkjump           = 1,
 tkjump           = 4,
 txmin            = 55.,
 txmax            = 75.,
 tymin            = -80.,
 tymax            = 80.,
 tzmin            = -20.,
 tzmax            = 20.,
 t_in             = 0.,
 t_out            = 200.,
/
  • nkjump a tracked particle every nkjump is written in the output file
  • tkjump a snapshot of the tracked particles phase space is taken every tkjump timestep
  • txmin to select particles with initial longitudinal coordinate x > txmin to be tracked
  • txmax to select particles with initial longitudinal coordinate x < txmax to be tracked
  • tymin to select particles with initial transverse coordinate y > tymin to be tracked
  • tymax to select particles with initial transverse coordinate y < tymax to be tracked
  • tzmin to select particles with initial transverse coordinate z > tzmin to be tracked
  • tzmax to select particles with initial transverse coordinate z < tzmax to be tracked
  • t_in initial tracking time
  • t_out final tracking time

MPIPARAMS namelist block

&MPIPARAMS
 nprocx         = 1,
 nprocy         = 64,
 nprocz         = 1
/
  • nprocx defines the number of processor the simulation will be splitted along the x coordinate
  • nprocy defines the number of processor the simulation will be splitted along the y coordinate
  • nprocz defines the number of processor the simulation will be splitted along the z coordinate
    • mind that usually we do not split along x and we use the same number of processor along y and z (for 3D simulations) or we define nprocz=1 explicitly for 2D simulations. No guarantees are given if you want to explore other configurations (that should work anyway).