diag_f0.63_01

Update of diag for vp_coord (gkvp_f0.63 or later)
With the update from GKV version gkvp_f0.62 to gkvp_f0.63, the choice of velocity space coordinate has been extended: it is now possible to select either (vl,mu) or (vl,vp). Accordingly, related operations in the post-processing program diag have also been modified.

This change affects the array shape of variables such as gmu, from gmu(0:global_nm) to gmu(-global_nz:global_nz-1, 0:global_nm). Therefore, several related operations in diag—including geometry setup (src/diag_geom.f90), NetCDF read/write routines (src/diag_rb_netcdf.f90, src/out_netcdf.f90), and velocity space analysis (src/out_ffin**.f90)—have been updated for compatibility with the selectable velocity space coordinates in gkvp_f0.63.

What's Changed

• vp_coord compatible by @smaeyama in #12

Full Changelog: f0.61_03...f0.63_01

diag_f0.61_03

To allow changing resolution and MPI decomposition when restarting your run, diag_chgres module (diag_chgres_cnt.f90 and diag_chgres_util.f90, developed by Fujitsu, extended by S. Maeyama) is added.

Example to use (in diag_main.f90):

import diag_chgres_cnt, only : chgres_cnt_netcdf, chgres_cnt_fortran
stpn = 1  ! Run number of existing cnt file, of which you want to change resolution
nnx = 48  ! New resolution of nx in gkvp_header.f90
ngy = 19  ! New resolution of global_ny in gkvp_header.f90
ngz = 32  ! New resolution of global_nz in gkvp_header.f90
ngv = 48  ! New resolution of global_nv in gkvp_header.f90
ngm = 47  ! New resolution of global_nm in gkvp_header.f90
nnpw = 2  ! New MPI decomposition of nprocw in gkvp_header.f90
nnpz = 4  ! New MPI decomposition of nprocz in gkvp_header.f90
nnpv = 4  ! New MPI decomposition of nprocv in gkvp_header.f90
nnpm = 4  ! New MPI decomposition of nprocm in gkvp_header.f90
nnps = 3  ! New MPI decomposition of nprocs in gkvp_header.f90
!### For Fortran binary output, cnt/gkvp.******.cnt.***
call chgres_cnt_fortran(stpn=stpn,nnx=nnx, ngy=ngy, ngz=ngz, ngv=ngv, ngm=ngm, &
                        nnpw=nnpw, nnpz=nnpz, nnpv=nnpv, nnpm=nnpm, nnps=nnps)
!### For NetCDF4 output, cnt/gkvp.cnt.***.nc
!call chgres_cnt_netcdf(stpn=stpn,nnx=nnx, ngy=ngy, ngz=ngz, ngv=ngv, ngm=ngm, &
!                       nnpw=nnpw, nnpz=nnpz, nnpv=nnpv, nnpm=nnpm, nnps=nnps)

Increments of mode number In kx and ky => zero padding.
Decrease of mode number in kx and ky => truncated.
Resolution change in z,v,m directions => linear interpolations.
Increments of plasma species s => zero padding.
Decrease of plasma species s => truncated.

diag_f0.61_02

fluxinvm_fxv, fluxinvm_cnt are modified to output real(vf) and imag(vf), which allows evaluation of phase between v and f.

diag_f0.61_01

• equib_type = "s-alpha-shift" is added. See, A. Ishizawa, et al., PRL 123, 025003 (2019).
• Subroutines fluxinvm_fxv, fluxinvm_cnt are added in src/out_ffinvm.f90, that diagnose the velocity-space-dependent turbulent flux Gamma(zz,vl,mu).
• src/out_netcdf.f90 is changed to cmode=NF90_NETCDF4.
• src/out_mominxmf.f90 is updated. Output phi, Al, mom. Separate header(.xmf) file. Interpolation in zz direction is corrected. n_tht>1 is available.
• src/out_mominrz.f90 is updated. Interpolation in zz direction is corrected. n_tht>1 is available.
• Nonlinear triad transfer analysis based on fluid approximation [S. Maeyama, et al., NF 57, 066036 (2017).] is maintained in src/out_fluidtotaltrans.f90 and src/out_fluiddetailtrans.f90.

diag_f0.59_02

go.diag_fugaku is modified.

diag_f0.59_01

First release of diag on GitHub.

Naming rule: diag_fX.YZ_** basically supports binary outputs from GKV version gkvp_fX.YZ.