I am working on setting up N-body simulations of galaxy mergers. The initial conditions are generated from the DICE code, and looking at them, they seem very realistic. Unfortunately, when I let RAMSES run, it fails to converge. The output looks something like this
Initial mesh structure
Level 1 has 1 grids ( 0, 1, 0,)
Level 2 has 8 grids ( 4, 4, 4,)
Level 3 has 64 grids ( 32, 32, 32,)
Level 4 has 512 grids ( 256, 256, 256,)
Level 5 has 4096 grids ( 2048, 2048, 2048,)
Level 6 has 32768 grids ( 16384, 16384, 16384,)
Level 7 has 262144 grids ( 131072, 131072, 131072,)
Starting time integration
Load balancing AMR grid...
==> Level= 7 Step= 10 Error= NaN
WARN: Fine multigrid Poisson failed to converge...
Fine step= 0 t= 0.00000E+00 dt= 2.714E+04 a= 1.000E+00 mem=17.3% 0.0%
Time elapsed since last coarse step: 5.76 s 5.50 mus/pt 5.50 mus/pt (av)
Used memory: 799.9 MB
Total running time: 97.1599960 s
==> Level= 7 Step= 10 Error= NaN
WARN: Fine multigrid Poisson failed to converge...
Mesh structure
Level 1 has 1 grids ( 0, 1, 0,)
Level 2 has 8 grids ( 4, 4, 4,)
Level 3 has 64 grids ( 32, 32, 32,)
Level 4 has 512 grids ( 256, 256, 256,)
Level 5 has 4096 grids ( 2048, 2048, 2048,)
Level 6 has 32768 grids ( 16384, 16384, 16384,)
Level 7 has 262144 grids ( 131072, 131072, 131072,)
Main step= 1 mcons= 0.00E+00 econs= 0.00E+00 epot= NaN ekin= 0.00E+00
Fine step= 1 t= 2.71350E+04 dt= 2.714E+04 a= 1.000E+00 mem=17.3% 0.0%
Run completed
Total elapsed time: 4.4218359169999957
--------------------------------------------------------------------
minimum average maximum standard dev std/av % rmn rmx TIMER
0.023 0.023 0.023 0.000 0.006 0.2 1 2 coarse levels
0.055 0.055 0.055 0.000 0.000 0.5 2 1 refine
1.397 1.397 1.397 0.000 0.000 13.9 1 2 load balance
0.506 0.509 0.513 0.003 0.007 5.1 2 1 io
7.884 7.884 7.884 0.000 0.000 78.6 1 2 poisson
0.009 0.012 0.016 0.003 0.282 0.1 1 2 rho
0.148 0.148 0.148 0.000 0.000 1.5 1 2 flag
10.034 100.0 TOTAL
Note: The following floating-point exceptions are signalling: IEEE_INVALID_FLAG
Note: The following floating-point exceptions are signalling: IEEE_INVALID_FLAG IEEE_DENORMAL
The namelist file I am running is
#!fortran
&RUN_PARAMS
pic=.true.
poisson=.true.
nrestart=0
nremap=10
nsubcycle=1,2,
ncontrol=1
/
&INIT_PARAMS
filetype='dice'
initfile(1)='/uufs/astro.utah.edu/common/home/u1281896/NBodies/DICE_IC_FILES'
/
&DICE_PARAMS
ic_file='dice_spiral.g2'
ic_nfile=1
ic_ifout=1
ic_format='Gadget2'
ic_center=0.0,0.0,0.0
ic_scale_pos=1.0
ic_scale_vel=1.0
ic_scale_mass=1.0
ic_scale_u=1.0
ic_scale_age=1.0
ic_scale_metal=1.0
ic_head_name='HEAD'
ic_pos_name='POS '
ic_vel_name='VEL '
ic_mass_name='MASS'
ic_id_name='ID '
ic_u_name='U '
ic_metal_name='Z '
ic_age_name='AGE '
IG_rho=1.0D-6
IG_T2=1.0D6
IG_metal=1.0D-4
amr_struct=.false.
gadget_scale_l=3.085677581282D21
gadget_scale_v=1.0D5
gadget_scale_m=1.9891D43
gadget_scale_t=3.15360e+13
ic_skip_type=-1
cosmo_add_gas_index=-1
ic_mag_const=0.0,0.0,0.0
ic_mag_center_x=0.0
ic_mag_center_y=0.0
ic_mag_center_z=0.0
ic_mag_axis_x=0.0
ic_mag_axis_y=0.0
ic_mag_axis_z=1.0
ic_mag_scale_R=1.0
ic_mag_scale_H=1.0
ic_mag_scale_B=0.0
/
&REFINE_PARAMS
m_refine=7*8.,
/
&BOUNDARY_PARAMS
nboundary=6
bound_type=2,2,2,2,2,2
ibound_min=-1,1,-1,-1,-1,-1
ibound_max=-1,1,1,1,1,1
jbound_min=0,0,-1,1,-1,-1
jbound_max=0,0,-1,1,1,1
kbound_min=0,0,0,0,-1,1
kbound_max=0,0,0,0,-1,1
no_inflow=.true.
/
&OUTPUT_PARAMS
noutput=5
tout=0.0,20.0,40.0,60.0,80.0
tend=80.0
/
&AMR_PARAMS
levelmin=7
levelmax=14
ngridtot=2000000
nparttot=3000000
nexpand=1
boxlen=1.0
/
&HYDRO_PARAMS
gamma=1.6666667
courant_factor=0.8
slope_type=1
pressure_fix=.true.
scheme='muscl'
/
&POISSON_PARAMS
epsilon=1.d-2
/
Is anyone familiar enough with the codes to help? I reached out to the author of RAMSES, but he seems to be MIA!
