qp2/src/mo_two_e_ints/mo_bi_integrals_erf.irp.f

659 lines
20 KiB
Fortran

subroutine mo_two_e_integrals_erf_index(i,j,k,l,i1)
use map_module
implicit none
BEGIN_DOC
! Computes an unique index for i,j,k,l integrals
END_DOC
integer, intent(in) :: i,j,k,l
integer(key_kind), intent(out) :: i1
integer(key_kind) :: p,q,r,s,i2
p = min(i,k)
r = max(i,k)
p = p+shiftr(r*r-r,1)
q = min(j,l)
s = max(j,l)
q = q+shiftr(s*s-s,1)
i1 = min(p,q)
i2 = max(p,q)
i1 = i1+shiftr(i2*i2-i2,1)
end
BEGIN_PROVIDER [ logical, mo_two_e_integrals_erf_in_map ]
use map_module
implicit none
BEGIN_DOC
! If True, the map of MO two-electron integrals is provided
END_DOC
integer(bit_kind) :: mask_ijkl(N_int,4)
integer(bit_kind) :: mask_ijk(N_int,3)
double precision :: cpu_1, cpu_2, wall_1, wall_2
PROVIDE mo_class
mo_two_e_integrals_erf_in_map = .True.
if (read_mo_two_e_integrals_erf) then
print*,'Reading the MO integrals_erf'
call map_load_from_disk(trim(ezfio_filename)//'/work/mo_ints_erf',mo_integrals_erf_map)
print*, 'MO integrals_erf provided'
return
endif
PROVIDE ao_two_e_integrals_erf_in_map
print *, ''
print *, 'AO -> MO ERF integrals transformation'
print *, '-------------------------------------'
print *, ''
call wall_time(wall_1)
call cpu_time(cpu_1)
if (dble(ao_num)**4 * 32.d-9 < dble(qp_max_mem)) then
call four_idx_dgemm_erf
else
call add_integrals_to_map_erf(full_ijkl_bitmask_4)
endif
call wall_time(wall_2)
call cpu_time(cpu_2)
integer*8 :: get_mo_erf_map_size, mo_erf_map_size
mo_erf_map_size = get_mo_erf_map_size()
double precision, external :: map_mb
print*,'Molecular integrals provided:'
print*,' Size of MO map ', map_mb(mo_integrals_erf_map) ,'MB'
print*,' Number of MO integrals: ', mo_erf_map_size
print*,' cpu time :',cpu_2 - cpu_1, 's'
print*,' wall time :',wall_2 - wall_1, 's ( x ', (cpu_2-cpu_1)/(wall_2-wall_1), ')'
if (write_mo_two_e_integrals_erf.and.mpi_master) then
call ezfio_set_work_empty(.False.)
call map_save_to_disk(trim(ezfio_filename)//'/work/mo_ints_erf',mo_integrals_erf_map)
call ezfio_set_mo_two_e_ints_io_mo_two_e_integrals_erf('Read')
endif
END_PROVIDER
subroutine four_idx_dgemm_erf
implicit none
integer :: p,q,r,s,i,j,k,l
double precision, allocatable :: a1(:,:,:,:)
double precision, allocatable :: a2(:,:,:,:)
if (ao_num > 1289) then
print *, irp_here, ': Integer overflow in ao_num**3'
endif
allocate (a1(ao_num,ao_num,ao_num,ao_num))
print *, 'Getting AOs'
!$OMP PARALLEL DO DEFAULT(SHARED) PRIVATE(q,r,s)
do s=1,ao_num
do r=1,ao_num
do q=1,ao_num
call get_ao_two_e_integrals_erf(q,r,s,ao_num,a1(1,q,r,s))
enddo
enddo
enddo
!$OMP END PARALLEL DO
print *, '1st transformation'
! 1st transformation
allocate (a2(ao_num,ao_num,ao_num,mo_num))
call dgemm('T','N', (ao_num*ao_num*ao_num), mo_num, ao_num, 1.d0, a1, ao_num, mo_coef, ao_num, 0.d0, a2, (ao_num*ao_num*ao_num))
! 2nd transformation
print *, '2nd transformation'
deallocate (a1)
allocate (a1(ao_num,ao_num,mo_num,mo_num))
call dgemm('T','N', (ao_num*ao_num*mo_num), mo_num, ao_num, 1.d0, a2, ao_num, mo_coef, ao_num, 0.d0, a1, (ao_num*ao_num*mo_num))
! 3rd transformation
print *, '3rd transformation'
deallocate (a2)
allocate (a2(ao_num,mo_num,mo_num,mo_num))
call dgemm('T','N', (ao_num*mo_num*mo_num), mo_num, ao_num, 1.d0, a1, ao_num, mo_coef, ao_num, 0.d0, a2, (ao_num*mo_num*mo_num))
! 4th transformation
print *, '4th transformation'
deallocate (a1)
allocate (a1(mo_num,mo_num,mo_num,mo_num))
call dgemm('T','N', (mo_num*mo_num*mo_num), mo_num, ao_num, 1.d0, a2, ao_num, mo_coef, ao_num, 0.d0, a1, (mo_num*mo_num*mo_num))
deallocate (a2)
integer :: n_integrals, size_buffer
integer(key_kind) , allocatable :: buffer_i(:)
real(integral_kind), allocatable :: buffer_value(:)
size_buffer = min(ao_num*ao_num*ao_num,16000000)
!$OMP PARALLEL DEFAULT(SHARED) PRIVATE(i,j,k,l,buffer_value,buffer_i,n_integrals)
allocate ( buffer_i(size_buffer), buffer_value(size_buffer) )
n_integrals = 0
!$OMP DO
do l=1,mo_num
do k=1,mo_num
do j=1,l
do i=1,k
if (abs(a1(i,j,k,l)) < mo_integrals_threshold) then
cycle
endif
n_integrals += 1
buffer_value(n_integrals) = a1(i,j,k,l)
!DIR$ FORCEINLINE
call mo_two_e_integrals_index(i,j,k,l,buffer_i(n_integrals))
if (n_integrals == size_buffer) then
call map_append(mo_integrals_erf_map, buffer_i, buffer_value, n_integrals)
n_integrals = 0
endif
enddo
enddo
enddo
enddo
!$OMP END DO
call map_append(mo_integrals_erf_map, buffer_i, buffer_value, n_integrals)
deallocate(buffer_i, buffer_value)
!$OMP END PARALLEL
deallocate (a1)
call map_sort(mo_integrals_erf_map)
call map_unique(mo_integrals_erf_map)
end subroutine
BEGIN_PROVIDER [ double precision, mo_two_e_int_erf_jj_from_ao, (mo_num,mo_num) ]
&BEGIN_PROVIDER [ double precision, mo_two_e_int_erf_jj_exchange_from_ao, (mo_num,mo_num) ]
&BEGIN_PROVIDER [ double precision, mo_two_e_int_erf_jj_anti_from_ao, (mo_num,mo_num) ]
BEGIN_DOC
! mo_two_e_integral_jj_from_ao(i,j) = J_ij
! mo_two_e_integrals_jj_exchange_from_ao(i,j) = J_ij
! mo_two_e_integrals_jj_anti_from_ao(i,j) = J_ij - K_ij
END_DOC
implicit none
integer :: i,j,p,q,r,s
double precision :: c
real(integral_kind) :: integral
integer :: n, pp
real(integral_kind), allocatable :: int_value(:)
integer, allocatable :: int_idx(:)
double precision, allocatable :: iqrs(:,:), iqsr(:,:), iqis(:), iqri(:)
if (.not.do_direct_integrals) then
PROVIDE ao_two_e_integrals_erf_in_map mo_coef
endif
mo_two_e_int_erf_jj_from_ao = 0.d0
mo_two_e_int_erf_jj_exchange_from_ao = 0.d0
!DIR$ ATTRIBUTES ALIGN : $IRP_ALIGN :: iqrs, iqsr
!$OMP PARALLEL DEFAULT(NONE) &
!$OMP PRIVATE (i,j,p,q,r,s,integral,c,n,pp,int_value,int_idx, &
!$OMP iqrs, iqsr,iqri,iqis) &
!$OMP SHARED(mo_num,mo_coef_transp,ao_num,&
!$OMP ao_integrals_threshold,do_direct_integrals) &
!$OMP REDUCTION(+:mo_two_e_int_erf_jj_from_ao,mo_two_e_int_erf_jj_exchange_from_ao)
allocate( int_value(ao_num), int_idx(ao_num), &
iqrs(mo_num,ao_num), iqis(mo_num), iqri(mo_num),&
iqsr(mo_num,ao_num) )
!$OMP DO SCHEDULE (guided)
do s=1,ao_num
do q=1,ao_num
do j=1,ao_num
!DIR$ VECTOR ALIGNED
do i=1,mo_num
iqrs(i,j) = 0.d0
iqsr(i,j) = 0.d0
enddo
enddo
if (do_direct_integrals) then
double precision :: ao_two_e_integral_erf
do r=1,ao_num
call compute_ao_two_e_integrals_erf(q,r,s,ao_num,int_value)
do p=1,ao_num
integral = int_value(p)
if (abs(integral) > ao_integrals_threshold) then
!DIR$ VECTOR ALIGNED
do i=1,mo_num
iqrs(i,r) += mo_coef_transp(i,p) * integral
enddo
endif
enddo
call compute_ao_two_e_integrals_erf(q,s,r,ao_num,int_value)
do p=1,ao_num
integral = int_value(p)
if (abs(integral) > ao_integrals_threshold) then
!DIR$ VECTOR ALIGNED
do i=1,mo_num
iqsr(i,r) += mo_coef_transp(i,p) * integral
enddo
endif
enddo
enddo
else
do r=1,ao_num
call get_ao_two_e_integrals_erf_non_zero(q,r,s,ao_num,int_value,int_idx,n)
do pp=1,n
p = int_idx(pp)
integral = int_value(pp)
if (abs(integral) > ao_integrals_threshold) then
!DIR$ VECTOR ALIGNED
do i=1,mo_num
iqrs(i,r) += mo_coef_transp(i,p) * integral
enddo
endif
enddo
call get_ao_two_e_integrals_erf_non_zero(q,s,r,ao_num,int_value,int_idx,n)
do pp=1,n
p = int_idx(pp)
integral = int_value(pp)
if (abs(integral) > ao_integrals_threshold) then
!DIR$ VECTOR ALIGNED
do i=1,mo_num
iqsr(i,r) += mo_coef_transp(i,p) * integral
enddo
endif
enddo
enddo
endif
iqis = 0.d0
iqri = 0.d0
do r=1,ao_num
!DIR$ VECTOR ALIGNED
do i=1,mo_num
iqis(i) += mo_coef_transp(i,r) * iqrs(i,r)
iqri(i) += mo_coef_transp(i,r) * iqsr(i,r)
enddo
enddo
do i=1,mo_num
!DIR$ VECTOR ALIGNED
do j=1,mo_num
c = mo_coef_transp(j,q)*mo_coef_transp(j,s)
mo_two_e_int_erf_jj_from_ao(j,i) += c * iqis(i)
mo_two_e_int_erf_jj_exchange_from_ao(j,i) += c * iqri(i)
enddo
enddo
enddo
enddo
!$OMP END DO NOWAIT
deallocate(iqrs,iqsr,int_value,int_idx)
!$OMP END PARALLEL
mo_two_e_int_erf_jj_anti_from_ao = mo_two_e_int_erf_jj_from_ao - mo_two_e_int_erf_jj_exchange_from_ao
! end
END_PROVIDER
BEGIN_PROVIDER [ double precision, mo_two_e_int_erf_jj, (mo_num,mo_num) ]
&BEGIN_PROVIDER [ double precision, mo_two_e_int_erf_jj_exchange, (mo_num,mo_num) ]
&BEGIN_PROVIDER [ double precision, mo_two_e_int_erf_jj_anti, (mo_num,mo_num) ]
implicit none
BEGIN_DOC
! mo_two_e_integrals_jj(i,j) = J_ij
! mo_two_e_integrals_jj_exchange(i,j) = K_ij
! mo_two_e_integrals_jj_anti(i,j) = J_ij - K_ij
END_DOC
integer :: i,j
double precision :: get_mo_two_e_integral_erf
PROVIDE mo_two_e_integrals_erf_in_map
mo_two_e_int_erf_jj = 0.d0
mo_two_e_int_erf_jj_exchange = 0.d0
do j=1,mo_num
do i=1,mo_num
mo_two_e_int_erf_jj(i,j) = get_mo_two_e_integral_erf(i,j,i,j,mo_integrals_erf_map)
mo_two_e_int_erf_jj_exchange(i,j) = get_mo_two_e_integral_erf(i,j,j,i,mo_integrals_erf_map)
mo_two_e_int_erf_jj_anti(i,j) = mo_two_e_int_erf_jj(i,j) - mo_two_e_int_erf_jj_exchange(i,j)
enddo
enddo
END_PROVIDER
subroutine clear_mo_erf_map
implicit none
BEGIN_DOC
! Frees the memory of the MO map
END_DOC
call map_deinit(mo_integrals_erf_map)
FREE mo_integrals_erf_map mo_two_e_int_erf_jj mo_two_e_int_erf_jj_anti
FREE mo_two_e_int_erf_jj_exchange mo_two_e_integrals_erf_in_map
end
subroutine provide_all_mo_integrals_erf
implicit none
provide mo_integrals_erf_map mo_two_e_int_erf_jj mo_two_e_int_erf_jj_anti
provide mo_two_e_int_erf_jj_exchange mo_two_e_integrals_erf_in_map
end
subroutine add_integrals_to_map_erf(mask_ijkl)
use bitmasks
implicit none
BEGIN_DOC
! Adds integrals to tha MO map according to some bitmask
END_DOC
integer(bit_kind), intent(in) :: mask_ijkl(N_int,4)
integer :: i,j,k,l
integer :: i0,j0,k0,l0
double precision :: c, cpu_1, cpu_2, wall_1, wall_2, wall_0
integer, allocatable :: list_ijkl(:,:)
integer :: n_i, n_j, n_k, n_l
integer, allocatable :: two_e_tmp_0_idx(:)
real(integral_kind), allocatable :: two_e_tmp_0(:,:)
double precision, allocatable :: two_e_tmp_1(:)
double precision, allocatable :: two_e_tmp_2(:,:)
double precision, allocatable :: two_e_tmp_3(:,:,:)
!DIR$ ATTRIBUTES ALIGN : 64 :: two_e_tmp_1, two_e_tmp_2, two_e_tmp_3
integer :: n_integrals
integer :: size_buffer
integer(key_kind),allocatable :: buffer_i(:)
real(integral_kind),allocatable :: buffer_value(:)
double precision :: map_mb
integer :: i1,j1,k1,l1, ii1, kmax, thread_num
integer :: i2,i3,i4
double precision,parameter :: thr_coef = 1.d-10
PROVIDE ao_two_e_integrals_in_map mo_coef
!Get list of MOs for i,j,k and l
!-------------------------------
allocate(list_ijkl(mo_num,4))
call bitstring_to_list( mask_ijkl(1,1), list_ijkl(1,1), n_i, N_int )
call bitstring_to_list( mask_ijkl(1,2), list_ijkl(1,2), n_j, N_int )
call bitstring_to_list( mask_ijkl(1,3), list_ijkl(1,3), n_k, N_int )
call bitstring_to_list( mask_ijkl(1,4), list_ijkl(1,4), n_l, N_int )
character*(2048) :: output(1)
print*, 'i'
call bitstring_to_str( output(1), mask_ijkl(1,1), N_int )
print *, trim(output(1))
j = 0
do i = 1, N_int
j += popcnt(mask_ijkl(i,1))
enddo
if(j==0)then
return
endif
print*, 'j'
call bitstring_to_str( output(1), mask_ijkl(1,2), N_int )
print *, trim(output(1))
j = 0
do i = 1, N_int
j += popcnt(mask_ijkl(i,2))
enddo
if(j==0)then
return
endif
print*, 'k'
call bitstring_to_str( output(1), mask_ijkl(1,3), N_int )
print *, trim(output(1))
j = 0
do i = 1, N_int
j += popcnt(mask_ijkl(i,3))
enddo
if(j==0)then
return
endif
print*, 'l'
call bitstring_to_str( output(1), mask_ijkl(1,4), N_int )
print *, trim(output(1))
j = 0
do i = 1, N_int
j += popcnt(mask_ijkl(i,4))
enddo
if(j==0)then
return
endif
size_buffer = min(ao_num*ao_num*ao_num,16000000)
print*, 'Providing the ERF molecular integrals '
print*, 'Buffers : ', 8.*(mo_num*(n_j)*(n_k+1) + mo_num+&
ao_num+ao_num*ao_num+ size_buffer*3)/(1024*1024), 'MB / core'
call wall_time(wall_1)
call cpu_time(cpu_1)
double precision :: accu_bis
accu_bis = 0.d0
!$OMP PARALLEL PRIVATE(l1,k1,j1,i1,i2,i3,i4,i,j,k,l,c, ii1,kmax, &
!$OMP two_e_tmp_0_idx, two_e_tmp_0, two_e_tmp_1,two_e_tmp_2,two_e_tmp_3,&
!$OMP buffer_i,buffer_value,n_integrals,wall_2,i0,j0,k0,l0, &
!$OMP wall_0,thread_num,accu_bis) &
!$OMP DEFAULT(NONE) &
!$OMP SHARED(size_buffer,ao_num,mo_num,n_i,n_j,n_k,n_l, &
!$OMP mo_coef_transp, &
!$OMP mo_coef_transp_is_built, list_ijkl, &
!$OMP mo_coef_is_built, wall_1, &
!$OMP mo_coef,mo_integrals_threshold,mo_integrals_erf_map)
n_integrals = 0
wall_0 = wall_1
allocate(two_e_tmp_3(mo_num, n_j, n_k), &
two_e_tmp_1(mo_num), &
two_e_tmp_0(ao_num,ao_num), &
two_e_tmp_0_idx(ao_num), &
two_e_tmp_2(mo_num, n_j), &
buffer_i(size_buffer), &
buffer_value(size_buffer) )
thread_num = 0
!$ thread_num = omp_get_thread_num()
!$OMP DO SCHEDULE(guided)
do l1 = 1,ao_num
two_e_tmp_3 = 0.d0
do k1 = 1,ao_num
two_e_tmp_2 = 0.d0
do j1 = 1,ao_num
call get_ao_two_e_integrals_erf(j1,k1,l1,ao_num,two_e_tmp_0(1,j1)) ! all integrals for a given l1, k1
! call compute_ao_two_e_integrals(j1,k1,l1,ao_num,two_e_tmp_0(1,j1))
enddo
do j1 = 1,ao_num
kmax = 0
do i1 = 1,ao_num
c = two_e_tmp_0(i1,j1)
if (c == 0.d0) then
cycle
endif
kmax += 1
two_e_tmp_0(kmax,j1) = c
two_e_tmp_0_idx(kmax) = i1
enddo
if (kmax==0) then
cycle
endif
two_e_tmp_1 = 0.d0
ii1=1
! sum_m c_m^i (m)
do ii1 = 1,kmax-4,4
i1 = two_e_tmp_0_idx(ii1)
i2 = two_e_tmp_0_idx(ii1+1)
i3 = two_e_tmp_0_idx(ii1+2)
i4 = two_e_tmp_0_idx(ii1+3)
do i = list_ijkl(1,1), list_ijkl(n_i,1)
two_e_tmp_1(i) = two_e_tmp_1(i) + &
mo_coef_transp(i,i1) * two_e_tmp_0(ii1,j1) + &
mo_coef_transp(i,i2) * two_e_tmp_0(ii1+1,j1) + &
mo_coef_transp(i,i3) * two_e_tmp_0(ii1+2,j1) + &
mo_coef_transp(i,i4) * two_e_tmp_0(ii1+3,j1)
enddo ! i
enddo ! ii1
i2 = ii1
do ii1 = i2,kmax
i1 = two_e_tmp_0_idx(ii1)
do i = list_ijkl(1,1), list_ijkl(n_i,1)
two_e_tmp_1(i) = two_e_tmp_1(i) + mo_coef_transp(i,i1) * two_e_tmp_0(ii1,j1)
enddo ! i
enddo ! ii1
c = 0.d0
do i = list_ijkl(1,1), list_ijkl(n_i,1)
c = max(c,abs(two_e_tmp_1(i)))
if (c>mo_integrals_threshold) exit
enddo
if ( c < mo_integrals_threshold ) then
cycle
endif
do j0 = 1, n_j
j = list_ijkl(j0,2)
c = mo_coef_transp(j,j1)
if (abs(c) < thr_coef) then
cycle
endif
do i = list_ijkl(1,1), list_ijkl(n_i,1)
two_e_tmp_2(i,j0) = two_e_tmp_2(i,j0) + c * two_e_tmp_1(i)
enddo ! i
enddo ! j
enddo !j1
if ( maxval(abs(two_e_tmp_2)) < mo_integrals_threshold ) then
cycle
endif
do k0 = 1, n_k
k = list_ijkl(k0,3)
c = mo_coef_transp(k,k1)
if (abs(c) < thr_coef) then
cycle
endif
do j0 = 1, n_j
j = list_ijkl(j0,2)
do i = list_ijkl(1,1), k
two_e_tmp_3(i,j0,k0) = two_e_tmp_3(i,j0,k0) + c* two_e_tmp_2(i,j0)
enddo!i
enddo !j
enddo !k
enddo !k1
do l0 = 1,n_l
l = list_ijkl(l0,4)
c = mo_coef_transp(l,l1)
if (abs(c) < thr_coef) then
cycle
endif
j1 = ishft((l*l-l),-1)
do j0 = 1, n_j
j = list_ijkl(j0,2)
if (j > l) then
exit
endif
j1 += 1
do k0 = 1, n_k
k = list_ijkl(k0,3)
i1 = ishft((k*k-k),-1)
if (i1<=j1) then
continue
else
exit
endif
two_e_tmp_1 = 0.d0
do i0 = 1, n_i
i = list_ijkl(i0,1)
if (i>k) then
exit
endif
two_e_tmp_1(i) = c*two_e_tmp_3(i,j0,k0)
! i1+=1
enddo
do i0 = 1, n_i
i = list_ijkl(i0,1)
if(i> min(k,j1-i1+list_ijkl(1,1)-1))then
exit
endif
if (abs(two_e_tmp_1(i)) < mo_integrals_threshold) then
cycle
endif
n_integrals += 1
buffer_value(n_integrals) = two_e_tmp_1(i)
!DIR$ FORCEINLINE
call mo_two_e_integrals_index(i,j,k,l,buffer_i(n_integrals))
if (n_integrals == size_buffer) then
call insert_into_mo_integrals_erf_map(n_integrals,buffer_i,buffer_value,&
real(mo_integrals_threshold,integral_kind))
n_integrals = 0
endif
enddo
enddo
enddo
enddo
call wall_time(wall_2)
if (thread_num == 0) then
if (wall_2 - wall_0 > 1.d0) then
wall_0 = wall_2
print*, 100.*float(l1)/float(ao_num), '% in ', &
wall_2-wall_1, 's', map_mb(mo_integrals_erf_map) ,'MB'
endif
endif
enddo
!$OMP END DO NOWAIT
deallocate (two_e_tmp_1,two_e_tmp_2,two_e_tmp_3)
integer :: index_needed
call insert_into_mo_integrals_erf_map(n_integrals,buffer_i,buffer_value,&
real(mo_integrals_threshold,integral_kind))
deallocate(buffer_i, buffer_value)
!$OMP END PARALLEL
call map_merge(mo_integrals_erf_map)
call wall_time(wall_2)
call cpu_time(cpu_2)
integer*8 :: get_mo_erf_map_size, mo_map_size
mo_map_size = get_mo_erf_map_size()
deallocate(list_ijkl)
print*,'Molecular ERF integrals provided:'
print*,' Size of MO ERF map ', map_mb(mo_integrals_erf_map) ,'MB'
print*,' Number of MO ERF integrals: ', mo_map_size
print*,' cpu time :',cpu_2 - cpu_1, 's'
print*,' wall time :',wall_2 - wall_1, 's ( x ', (cpu_2-cpu_1)/(wall_2-wall_1), ')'
end