TABLE OF CONTENTS

ABINIT/alloc_hamilt_gpu [ Functions ]

[ Top ] [ Functions ]

NAME

 alloc_hamilt_gpu

FUNCTION

 allocate several memory pieces on a GPU device for the application
 of Hamiltonian using a GPU

INPUTS

  atindx1(natom)=index table for atoms, inverse of atindx (see gstate.f)
  dtset <type(dataset_type)>=all input variables for this dataset
  gprimd(3,3)=dimensional reciprocal space primitive translations
  mpi_enreg=information about MPI parallelization
  nattyp(ntypat)= # atoms of each type.
  option=0: allocate data for local operator (FFT)
         1: allocate data for nonlocal operator
         2: allocate both
  psps <type(pseudopotential_type)>=variables related to pseudopotentials
  use_gpu_cuda= 0 or 1 to know if we use cuda

OUTPUT

  (no output - only allocation on GPU)

SOURCE

 75 subroutine alloc_hamilt_gpu(atindx1,dtset,gprimd,mpi_enreg,nattyp,npwarr,option,psps,use_gpu_cuda)
 76 
 77 !Arguments ------------------------------------
 78 !scalars
 79  integer,intent(in) :: option,use_gpu_cuda
 80  type(dataset_type),intent(in) :: dtset
 81  type(MPI_type),intent(in) :: mpi_enreg
 82  type(pseudopotential_type),intent(in) :: psps
 83 !arrays
 84  integer,intent(in) :: atindx1(dtset%natom),nattyp(dtset%ntypat),npwarr(dtset%nkpt)
 85  real(dp),intent(in) :: gprimd(3,3)
 86 
 87 !Local variables-------------------------------
 88 !scalars
 89 #if defined HAVE_GPU_CUDA
 90  integer :: dimekb1_max,dimekb2_max,dimffnl_max,ierr,ikpt,npw_max_loc,npw_max_nonloc
 91  integer ::npwarr_tmp(dtset%nkpt)
 92 
 93  integer(kind=c_int32_t) :: proj_dim(3)
 94 #endif
 95 
 96 ! *************************************************************************
 97 
 98  if (use_gpu_cuda==0) return
 99 
100 #if defined(HAVE_GPU_CUDA) && defined(HAVE_FC_ISO_C_BINDING)
101 !=== Local Hamiltonian ===
102  if (option==0.or.option==2) then
103 !  Compute max of total planes waves
104    npw_max_loc=0
105    if(mpi_enreg%paral_kgb==1) then
106      npwarr_tmp=npwarr
107      call xmpi_sum(npwarr_tmp,mpi_enreg%comm_bandfft,ierr)
108      npw_max_loc =maxval(npwarr_tmp)
109    else
110      npw_max_loc=dtset%mpw
111    end if
112 !  Initialize gpu data needed in fourwf
113 !  ndat=bandpp when paral_kgb=1
114    if(mpi_enreg%paral_kgb==1) then
115      call alloc_gpu_fourwf(dtset%ngfft,dtset%bandpp,npw_max_loc,npw_max_loc)
116    else
117      call alloc_gpu_fourwf(dtset%ngfft,1,npw_max_loc,npw_max_loc)
118    end if
119  end if
120 !=== Nonlocal Hamiltonian ===
121  if (option==1.or.option==2) then
122 !  Compute max of total planes waves
123    npw_max_nonloc=0
124    if(mpi_enreg%paral_kgb==1) then
125      npwarr_tmp=npwarr
126      call xmpi_sum(npwarr_tmp,mpi_enreg%comm_band,ierr)
127      npw_max_nonloc =maxval(npwarr_tmp)
128    else
129      npw_max_nonloc=dtset%mpw
130    end if
131 !  Initialize all gpu data needed in nonlop
132    dimffnl_max=4
133 !  if (abs(dtset%berryopt) == 5) dimffnl_max=4
134    dimekb1_max=psps%dimekb
135    if (dtset%usepaw==0) dimekb2_max=psps%ntypat
136    if (dtset%usepaw==1) dimekb2_max=dtset%natom
137    call alloc_nonlop_gpu(npw_max_nonloc,npw_max_nonloc,dtset%nspinor,dtset%natom,&
138 &   dtset%ntypat,psps%lmnmax,psps%indlmn,nattyp,atindx1,gprimd,&
139 &   dimffnl_max,dimffnl_max,dimekb1_max,dimekb2_max)
140  end if
141  call xmpi_barrier(mpi_enreg%comm_cell)
142 #else
143  if (.false.) then
144    write(std_out,*) atindx1(1),dtset%natom,gprimd(1,1),mpi_enreg%me,nattyp(1),option
145  end if
146 #endif
147 
148 end subroutine alloc_hamilt_gpu

ABINIT/dealloc_hamilt_gpu [ Functions ]

[ Top ] [ Functions ]

NAME

 dealloc_hamilt_gpu

FUNCTION

 deallocate several memory pieces on a GPU device used for the application
 of Hamiltonian using a GPU

INPUTS

  option=0: deallocate data for local operator (FFT)
         1: deallocate data for nonlocal operator
         2: deallocate both
  use_gpu_cuda= 0 or 1 to know if we use cuda

OUTPUT

  (no output - only deallocation on GPU)

SOURCE

170 subroutine dealloc_hamilt_gpu(option,use_gpu_cuda)
171 
172 !Arguments ------------------------------------
173 !scalars
174  integer,intent(in) :: option,use_gpu_cuda
175 !arrays
176 
177 !Local variables-------------------------------
178 
179 ! *************************************************************************
180 
181  if (use_gpu_cuda==0) return
182 
183 #if defined(HAVE_GPU_CUDA) && defined(HAVE_FC_ISO_C_BINDING)
184  if (option==0.or.option==2) then
185    call free_gpu_fourwf()
186  end if
187  if (option==1.or.option==2) then
188    call free_nonlop_gpu()
189  end if
190 #else
191  if (.false.) then
192    write(std_out,*) option
193  end if
194 #endif
195 
196 end subroutine dealloc_hamilt_gpu

ABINIT/m_alloc_hamilt_gpu [ Modules ]

[ Top ] [ Modules ]

NAME

  m_alloc_hamilt_gpu

FUNCTION

COPYRIGHT

  Copyright (C) 2000-2022 ABINIT group (MT, FDahm)
  This file is distributed under the terms of the
  GNU General Public License, see ~abinit/COPYING
  or http://www.gnu.org/copyleft/gpl.txt .

SOURCE

15 #if defined HAVE_CONFIG_H
16 #include "config.h"
17 #endif
18 
19 #include "abi_common.h"
20 
21 module m_alloc_hamilt_gpu
22 
23  use defs_basis
24  use m_abicore
25  use m_xmpi
26  use m_dtset
27 #if defined HAVE_GPU_CUDA
28  use m_gpu_toolbox
29 #endif
30 
31 #ifdef HAVE_FC_ISO_C_BINDING
32  use, intrinsic :: iso_c_binding
33 #endif
34 
35  use defs_datatypes, only : pseudopotential_type
36  use defs_abitypes, only : MPI_type
37 
38  implicit none
39 
40  private