10
0
mirror of https://github.com/LCPQ/quantum_package synced 2024-10-19 22:41:48 +02:00

Merge branch 'scemama-master' into toto

This commit is contained in:
Emmanuel Giner 2018-12-20 17:12:31 +01:00
commit 4cadc75ec2
55 changed files with 3344 additions and 524 deletions

1
TODO
View File

@ -5,7 +5,6 @@
* dft_utils_one_body
* dm_for_dft
* integrals_erf
* Faire un README pour les scripts workflow
* Molden format : http://cheminf.cmbi.ru.nl/molden/molden_format.html
* Virer tous les modules qui sont dans plugins
* Permettre aux utilisateurs de facilement deposer des plugins dans plugins via une commande

View File

@ -6,6 +6,7 @@ tags
Makefile
ao_basis
ao_one_e_integrals
ao_two_e_integrals
becke_numerical_grid
bitmask
cis
@ -19,7 +20,6 @@ dummy
electrons
ezfio_files
fci
four_idx
generators_cas
generators_full
hartree_fock
@ -28,6 +28,7 @@ iterations
mo_basis
mo_guess
mo_one_e_integrals
mo_two_e_integrals
mpi
mrpt_utils
nuclei

View File

@ -96,11 +96,11 @@ end = struct
let n_states =
States_number.to_int n
in
(*
let old_nstates, read_wf =
Ezfio.get_determinants_n_states (),
Ezfio.get_determinants_read_wf ()
in
(*
if read_wf && old_nstates <> n_states then
Printf.eprintf "Warning : n_states could not be changed because read_wf is true\n%!"
else

View File

@ -0,0 +1,33 @@
[disk_access_ao_integrals]
type: Disk_access
doc: Read/Write |AO| integrals from/to disk [ Write | Read | None ]
interface: ezfio,provider,ocaml
default: None
[disk_access_ao_integrals_erf]
type: Disk_access
doc: Read/Write |AO| integrals with the long range interaction from/to disk [ Write | Read | None ]
interface: ezfio,provider,ocaml
default: None
[ao_integrals_threshold]
type: Threshold
doc: If | (pq|rs) | < `ao_integrals_threshold` then (pq|rs) is zero
interface: ezfio,provider,ocaml
default: 1.e-15
ezfio_name: threshold_ao
[do_direct_integrals]
type: logical
doc: Compute integrals on the fly (very slow, only for debugging)
interface: ezfio,provider,ocaml
default: False
ezfio_name: direct
[mu_erf]
type: double precision
doc: cutting of the interaction in the range separated model
interface: ezfio,provider,ocaml
default: 0.5
ezfio_name: mu_erf

View File

@ -1,7 +1,5 @@
ao_one_e_integrals
mo_one_e_integrals
pseudo
bitmask
zmq
ao_basis
mo_basis

View File

@ -0,0 +1,194 @@
subroutine ao_bielec_integrals_erf_in_map_slave_tcp(i)
implicit none
integer, intent(in) :: i
BEGIN_DOC
! Computes a buffer of integrals. i is the ID of the current thread.
END_DOC
call ao_bielec_integrals_erf_in_map_slave(0,i)
end
subroutine ao_bielec_integrals_erf_in_map_slave_inproc(i)
implicit none
integer, intent(in) :: i
BEGIN_DOC
! Computes a buffer of integrals. i is the ID of the current thread.
END_DOC
call ao_bielec_integrals_erf_in_map_slave(1,i)
end
subroutine ao_bielec_integrals_erf_in_map_slave(thread,iproc)
use map_module
use f77_zmq
implicit none
BEGIN_DOC
! Computes a buffer of integrals
END_DOC
integer, intent(in) :: thread, iproc
integer :: j,l,n_integrals
integer :: rc
real(integral_kind), allocatable :: buffer_value(:)
integer(key_kind), allocatable :: buffer_i(:)
integer :: worker_id, task_id
character*(512) :: task
integer(ZMQ_PTR),external :: new_zmq_to_qp_run_socket
integer(ZMQ_PTR) :: zmq_to_qp_run_socket
integer(ZMQ_PTR), external :: new_zmq_push_socket
integer(ZMQ_PTR) :: zmq_socket_push
character*(64) :: state
zmq_to_qp_run_socket = new_zmq_to_qp_run_socket()
integer, external :: connect_to_taskserver
if (connect_to_taskserver(zmq_to_qp_run_socket,worker_id,thread) == -1) then
call end_zmq_to_qp_run_socket(zmq_to_qp_run_socket)
return
endif
zmq_socket_push = new_zmq_push_socket(thread)
allocate ( buffer_i(ao_num*ao_num), buffer_value(ao_num*ao_num) )
do
integer, external :: get_task_from_taskserver
if (get_task_from_taskserver(zmq_to_qp_run_socket,worker_id, task_id, task) == -1) then
exit
endif
if (task_id == 0) exit
read(task,*) j, l
integer, external :: task_done_to_taskserver
call compute_ao_integrals_erf_jl(j,l,n_integrals,buffer_i,buffer_value)
if (task_done_to_taskserver(zmq_to_qp_run_socket,worker_id,task_id) == -1) then
stop 'Unable to send task_done'
endif
call push_integrals(zmq_socket_push, n_integrals, buffer_i, buffer_value, task_id)
enddo
integer, external :: disconnect_from_taskserver
if (disconnect_from_taskserver(zmq_to_qp_run_socket,worker_id) == -1) then
continue
endif
deallocate( buffer_i, buffer_value )
call end_zmq_to_qp_run_socket(zmq_to_qp_run_socket)
call end_zmq_push_socket(zmq_socket_push,thread)
end
subroutine ao_bielec_integrals_erf_in_map_collector(zmq_socket_pull)
use map_module
use f77_zmq
implicit none
BEGIN_DOC
! Collects results from the AO integral calculation
END_DOC
integer(ZMQ_PTR), intent(in) :: zmq_socket_pull
integer :: j,l,n_integrals
integer :: rc
real(integral_kind), allocatable :: buffer_value(:)
integer(key_kind), allocatable :: buffer_i(:)
integer(ZMQ_PTR),external :: new_zmq_to_qp_run_socket
integer(ZMQ_PTR) :: zmq_to_qp_run_socket
integer(ZMQ_PTR), external :: new_zmq_pull_socket
integer*8 :: control, accu, sze
integer :: task_id, more
zmq_to_qp_run_socket = new_zmq_to_qp_run_socket()
sze = ao_num*ao_num
allocate ( buffer_i(sze), buffer_value(sze) )
accu = 0_8
more = 1
do while (more == 1)
rc = f77_zmq_recv( zmq_socket_pull, n_integrals, 4, 0)
if (rc == -1) then
n_integrals = 0
return
endif
if (rc /= 4) then
print *, irp_here, ': f77_zmq_recv( zmq_socket_pull, n_integrals, 4, 0)'
stop 'error'
endif
if (n_integrals >= 0) then
if (n_integrals > sze) then
deallocate (buffer_value, buffer_i)
sze = n_integrals
allocate (buffer_value(sze), buffer_i(sze))
endif
rc = f77_zmq_recv( zmq_socket_pull, buffer_i, key_kind*n_integrals, 0)
if (rc /= key_kind*n_integrals) then
print *, rc, key_kind, n_integrals
print *, irp_here, ': f77_zmq_recv( zmq_socket_pull, buffer_i, key_kind*n_integrals, 0)'
stop 'error'
endif
rc = f77_zmq_recv( zmq_socket_pull, buffer_value, integral_kind*n_integrals, 0)
if (rc /= integral_kind*n_integrals) then
print *, irp_here, ': f77_zmq_recv( zmq_socket_pull, buffer_value, integral_kind*n_integrals, 0)'
stop 'error'
endif
rc = f77_zmq_recv( zmq_socket_pull, task_id, 4, 0)
IRP_IF ZMQ_PUSH
IRP_ELSE
rc = f77_zmq_send( zmq_socket_pull, 0, 4, 0)
if (rc /= 4) then
print *, irp_here, ' : f77_zmq_send (zmq_socket_pull,...'
stop 'error'
endif
IRP_ENDIF
call insert_into_ao_integrals_erf_map(n_integrals,buffer_i,buffer_value)
accu += n_integrals
if (task_id /= 0) then
integer, external :: zmq_delete_task
if (zmq_delete_task(zmq_to_qp_run_socket,zmq_socket_pull,task_id,more) == -1) then
stop 'Unable to delete task'
endif
endif
endif
enddo
deallocate( buffer_i, buffer_value )
integer (map_size_kind) :: get_ao_erf_map_size
control = get_ao_erf_map_size(ao_integrals_erf_map)
if (control /= accu) then
print *, ''
print *, irp_here
print *, 'Control : ', control
print *, 'Accu : ', accu
print *, 'Some integrals were lost during the parallel computation.'
print *, 'Try to reduce the number of threads.'
stop
endif
call end_zmq_to_qp_run_socket(zmq_to_qp_run_socket)
end

View File

@ -0,0 +1,375 @@
use map_module
!! AO Map
!! ======
BEGIN_PROVIDER [ type(map_type), ao_integrals_map ]
implicit none
BEGIN_DOC
! AO integrals
END_DOC
integer(key_kind) :: key_max
integer(map_size_kind) :: sze
call bielec_integrals_index(ao_num,ao_num,ao_num,ao_num,key_max)
sze = key_max
call map_init(ao_integrals_map,sze)
print*, 'AO map initialized : ', sze
END_PROVIDER
subroutine bielec_integrals_index(i,j,k,l,i1)
use map_module
implicit none
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
subroutine bielec_integrals_index_reverse(i,j,k,l,i1)
use map_module
implicit none
integer, intent(out) :: i(8),j(8),k(8),l(8)
integer(key_kind), intent(in) :: i1
integer(key_kind) :: i2,i3
i = 0
i2 = ceiling(0.5d0*(dsqrt(8.d0*dble(i1)+1.d0)-1.d0))
l(1) = ceiling(0.5d0*(dsqrt(8.d0*dble(i2)+1.d0)-1.d0))
i3 = i1 - shiftr(i2*i2-i2,1)
k(1) = ceiling(0.5d0*(dsqrt(8.d0*dble(i3)+1.d0)-1.d0))
j(1) = int(i2 - shiftr(l(1)*l(1)-l(1),1),4)
i(1) = int(i3 - shiftr(k(1)*k(1)-k(1),1),4)
!ijkl
i(2) = i(1) !ilkj
j(2) = l(1)
k(2) = k(1)
l(2) = j(1)
i(3) = k(1) !kjil
j(3) = j(1)
k(3) = i(1)
l(3) = l(1)
i(4) = k(1) !klij
j(4) = l(1)
k(4) = i(1)
l(4) = j(1)
i(5) = j(1) !jilk
j(5) = i(1)
k(5) = l(1)
l(5) = k(1)
i(6) = j(1) !jkli
j(6) = k(1)
k(6) = l(1)
l(6) = i(1)
i(7) = l(1) !lijk
j(7) = i(1)
k(7) = j(1)
l(7) = k(1)
i(8) = l(1) !lkji
j(8) = k(1)
k(8) = j(1)
l(8) = i(1)
integer :: ii, jj
do ii=2,8
do jj=1,ii-1
if ( (i(ii) == i(jj)).and. &
(j(ii) == j(jj)).and. &
(k(ii) == k(jj)).and. &
(l(ii) == l(jj)) ) then
i(ii) = 0
exit
endif
enddo
enddo
do ii=1,8
if (i(ii) /= 0) then
call bielec_integrals_index(i(ii),j(ii),k(ii),l(ii),i2)
if (i1 /= i2) then
print *, i1, i2
print *, i(ii), j(ii), k(ii), l(ii)
stop 'bielec_integrals_index_reverse failed'
endif
endif
enddo
end
BEGIN_PROVIDER [ integer, ao_integrals_cache_min ]
&BEGIN_PROVIDER [ integer, ao_integrals_cache_max ]
implicit none
BEGIN_DOC
! Min and max values of the AOs for which the integrals are in the cache
END_DOC
ao_integrals_cache_min = max(1,ao_num - 63)
ao_integrals_cache_max = ao_num
END_PROVIDER
BEGIN_PROVIDER [ double precision, ao_integrals_cache, (0:64*64*64*64) ]
implicit none
BEGIN_DOC
! Cache of AO integrals for fast access
END_DOC
PROVIDE ao_bielec_integrals_in_map
integer :: i,j,k,l,ii
integer(key_kind) :: idx
real(integral_kind) :: integral
!$OMP PARALLEL DO PRIVATE (i,j,k,l,idx,ii,integral)
do l=ao_integrals_cache_min,ao_integrals_cache_max
do k=ao_integrals_cache_min,ao_integrals_cache_max
do j=ao_integrals_cache_min,ao_integrals_cache_max
do i=ao_integrals_cache_min,ao_integrals_cache_max
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,idx)
!DIR$ FORCEINLINE
call map_get(ao_integrals_map,idx,integral)
ii = l-ao_integrals_cache_min
ii = ior( shiftl(ii,6), k-ao_integrals_cache_min)
ii = ior( shiftl(ii,6), j-ao_integrals_cache_min)
ii = ior( shiftl(ii,6), i-ao_integrals_cache_min)
ao_integrals_cache(ii) = integral
enddo
enddo
enddo
enddo
!$OMP END PARALLEL DO
END_PROVIDER
double precision function get_ao_bielec_integral(i,j,k,l,map) result(result)
use map_module
implicit none
BEGIN_DOC
! Gets one AO bi-electronic integral from the AO map
END_DOC
integer, intent(in) :: i,j,k,l
integer(key_kind) :: idx
type(map_type), intent(inout) :: map
integer :: ii
real(integral_kind) :: tmp
PROVIDE ao_bielec_integrals_in_map ao_integrals_cache ao_integrals_cache_min
!DIR$ FORCEINLINE
if (ao_overlap_abs(i,k)*ao_overlap_abs(j,l) < ao_integrals_threshold ) then
tmp = 0.d0
else if (ao_bielec_integral_schwartz(i,k)*ao_bielec_integral_schwartz(j,l) < ao_integrals_threshold) then
tmp = 0.d0
else
ii = l-ao_integrals_cache_min
ii = ior(ii, k-ao_integrals_cache_min)
ii = ior(ii, j-ao_integrals_cache_min)
ii = ior(ii, i-ao_integrals_cache_min)
if (iand(ii, -64) /= 0) then
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,idx)
!DIR$ FORCEINLINE
call map_get(map,idx,tmp)
else
ii = l-ao_integrals_cache_min
ii = ior( shiftl(ii,6), k-ao_integrals_cache_min)
ii = ior( shiftl(ii,6), j-ao_integrals_cache_min)
ii = ior( shiftl(ii,6), i-ao_integrals_cache_min)
tmp = ao_integrals_cache(ii)
endif
endif
result = tmp
end
subroutine get_ao_bielec_integrals(j,k,l,sze,out_val)
use map_module
BEGIN_DOC
! Gets multiple AO bi-electronic integral from the AO map .
! All i are retrieved for j,k,l fixed.
END_DOC
implicit none
integer, intent(in) :: j,k,l, sze
real(integral_kind), intent(out) :: out_val(sze)
integer :: i
integer(key_kind) :: hash
double precision :: thresh
PROVIDE ao_bielec_integrals_in_map ao_integrals_map
thresh = ao_integrals_threshold
if (ao_overlap_abs(j,l) < thresh) then
out_val = 0.d0
return
endif
double precision :: get_ao_bielec_integral
do i=1,sze
out_val(i) = get_ao_bielec_integral(i,j,k,l,ao_integrals_map)
enddo
end
subroutine get_ao_bielec_integrals_non_zero(j,k,l,sze,out_val,out_val_index,non_zero_int)
use map_module
implicit none
BEGIN_DOC
! Gets multiple AO bi-electronic integral from the AO map .
! All non-zero i are retrieved for j,k,l fixed.
END_DOC
integer, intent(in) :: j,k,l, sze
real(integral_kind), intent(out) :: out_val(sze)
integer, intent(out) :: out_val_index(sze),non_zero_int
integer :: i
integer(key_kind) :: hash
double precision :: thresh,tmp
PROVIDE ao_bielec_integrals_in_map
thresh = ao_integrals_threshold
non_zero_int = 0
if (ao_overlap_abs(j,l) < thresh) then
out_val = 0.d0
return
endif
non_zero_int = 0
do i=1,sze
integer, external :: ao_l4
double precision, external :: ao_bielec_integral
!DIR$ FORCEINLINE
if (ao_bielec_integral_schwartz(i,k)*ao_bielec_integral_schwartz(j,l) < thresh) then
cycle
endif
call bielec_integrals_index(i,j,k,l,hash)
call map_get(ao_integrals_map, hash,tmp)
if (dabs(tmp) < thresh ) cycle
non_zero_int = non_zero_int+1
out_val_index(non_zero_int) = i
out_val(non_zero_int) = tmp
enddo
end
function get_ao_map_size()
implicit none
integer (map_size_kind) :: get_ao_map_size
BEGIN_DOC
! Returns the number of elements in the AO map
END_DOC
get_ao_map_size = ao_integrals_map % n_elements
end
subroutine clear_ao_map
implicit none
BEGIN_DOC
! Frees the memory of the AO map
END_DOC
call map_deinit(ao_integrals_map)
FREE ao_integrals_map
end
subroutine insert_into_ao_integrals_map(n_integrals,buffer_i, buffer_values)
use map_module
implicit none
BEGIN_DOC
! Create new entry into AO map
END_DOC
integer, intent(in) :: n_integrals
integer(key_kind), intent(inout) :: buffer_i(n_integrals)
real(integral_kind), intent(inout) :: buffer_values(n_integrals)
call map_append(ao_integrals_map, buffer_i, buffer_values, n_integrals)
end
subroutine dump_ao_integrals(filename)
use map_module
implicit none
BEGIN_DOC
! Save to disk the |AO| integrals
END_DOC
character*(*), intent(in) :: filename
integer(cache_key_kind), pointer :: key(:)
real(integral_kind), pointer :: val(:)
integer*8 :: i,j, n
if (.not.mpi_master) then
return
endif
call ezfio_set_work_empty(.False.)
open(unit=66,file=filename,FORM='unformatted')
write(66) integral_kind, key_kind
write(66) ao_integrals_map%sorted, ao_integrals_map%map_size, &
ao_integrals_map%n_elements
do i=0_8,ao_integrals_map%map_size
write(66) ao_integrals_map%map(i)%sorted, ao_integrals_map%map(i)%map_size,&
ao_integrals_map%map(i)%n_elements
enddo
do i=0_8,ao_integrals_map%map_size
key => ao_integrals_map%map(i)%key
val => ao_integrals_map%map(i)%value
n = ao_integrals_map%map(i)%n_elements
write(66) (key(j), j=1,n), (val(j), j=1,n)
enddo
close(66)
end
integer function load_ao_integrals(filename)
implicit none
BEGIN_DOC
! Read from disk the |AO| integrals
END_DOC
character*(*), intent(in) :: filename
integer*8 :: i
integer(cache_key_kind), pointer :: key(:)
real(integral_kind), pointer :: val(:)
integer :: iknd, kknd
integer*8 :: n, j
load_ao_integrals = 1
open(unit=66,file=filename,FORM='unformatted',STATUS='UNKNOWN')
read(66,err=98,end=98) iknd, kknd
if (iknd /= integral_kind) then
print *, 'Wrong integrals kind in file :', iknd
stop 1
endif
if (kknd /= key_kind) then
print *, 'Wrong key kind in file :', kknd
stop 1
endif
read(66,err=98,end=98) ao_integrals_map%sorted, ao_integrals_map%map_size,&
ao_integrals_map%n_elements
do i=0_8, ao_integrals_map%map_size
read(66,err=99,end=99) ao_integrals_map%map(i)%sorted, &
ao_integrals_map%map(i)%map_size, ao_integrals_map%map(i)%n_elements
call cache_map_reallocate(ao_integrals_map%map(i),ao_integrals_map%map(i)%map_size)
enddo
do i=0_8, ao_integrals_map%map_size
key => ao_integrals_map%map(i)%key
val => ao_integrals_map%map(i)%value
n = ao_integrals_map%map(i)%n_elements
read(66,err=99,end=99) (key(j), j=1,n), (val(j), j=1,n)
enddo
call map_sort(ao_integrals_map)
load_ao_integrals = 0
return
99 continue
call map_deinit(ao_integrals_map)
98 continue
stop 'Problem reading ao_integrals_map file in work/'
end

View File

@ -0,0 +1,285 @@
use map_module
!! AO Map
!! ======
BEGIN_PROVIDER [ type(map_type), ao_integrals_erf_map ]
implicit none
BEGIN_DOC
! AO integrals
END_DOC
integer(key_kind) :: key_max
integer(map_size_kind) :: sze
call bielec_integrals_index(ao_num,ao_num,ao_num,ao_num,key_max)
sze = key_max
call map_init(ao_integrals_erf_map,sze)
print*, 'AO map initialized : ', sze
END_PROVIDER
BEGIN_PROVIDER [ integer, ao_integrals_erf_cache_min ]
&BEGIN_PROVIDER [ integer, ao_integrals_erf_cache_max ]
implicit none
BEGIN_DOC
! Min and max values of the AOs for which the integrals are in the cache
END_DOC
ao_integrals_erf_cache_min = max(1,ao_num - 63)
ao_integrals_erf_cache_max = ao_num
END_PROVIDER
BEGIN_PROVIDER [ double precision, ao_integrals_erf_cache, (0:64*64*64*64) ]
use map_module
implicit none
BEGIN_DOC
! Cache of AO integrals for fast access
END_DOC
PROVIDE ao_bielec_integrals_erf_in_map
integer :: i,j,k,l,ii
integer(key_kind) :: idx
real(integral_kind) :: integral
!$OMP PARALLEL DO PRIVATE (i,j,k,l,idx,ii,integral)
do l=ao_integrals_erf_cache_min,ao_integrals_erf_cache_max
do k=ao_integrals_erf_cache_min,ao_integrals_erf_cache_max
do j=ao_integrals_erf_cache_min,ao_integrals_erf_cache_max
do i=ao_integrals_erf_cache_min,ao_integrals_erf_cache_max
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,idx)
!DIR$ FORCEINLINE
call map_get(ao_integrals_erf_map,idx,integral)
ii = l-ao_integrals_erf_cache_min
ii = ior( ishft(ii,6), k-ao_integrals_erf_cache_min)
ii = ior( ishft(ii,6), j-ao_integrals_erf_cache_min)
ii = ior( ishft(ii,6), i-ao_integrals_erf_cache_min)
ao_integrals_erf_cache(ii) = integral
enddo
enddo
enddo
enddo
!$OMP END PARALLEL DO
END_PROVIDER
subroutine insert_into_ao_integrals_erf_map(n_integrals,buffer_i, buffer_values)
use map_module
implicit none
BEGIN_DOC
! Create new entry into AO map
END_DOC
integer, intent(in) :: n_integrals
integer(key_kind), intent(inout) :: buffer_i(n_integrals)
real(integral_kind), intent(inout) :: buffer_values(n_integrals)
call map_append(ao_integrals_erf_map, buffer_i, buffer_values, n_integrals)
end
double precision function get_ao_bielec_integral_erf(i,j,k,l,map) result(result)
use map_module
implicit none
BEGIN_DOC
! Gets one AO bi-electronic integral from the AO map
END_DOC
integer, intent(in) :: i,j,k,l
integer(key_kind) :: idx
type(map_type), intent(inout) :: map
integer :: ii
real(integral_kind) :: tmp
PROVIDE ao_bielec_integrals_erf_in_map ao_integrals_erf_cache ao_integrals_erf_cache_min
!DIR$ FORCEINLINE
if (ao_overlap_abs(i,k)*ao_overlap_abs(j,l) < ao_integrals_threshold ) then
tmp = 0.d0
else if (ao_bielec_integral_erf_schwartz(i,k)*ao_bielec_integral_erf_schwartz(j,l) < ao_integrals_threshold) then
tmp = 0.d0
else
ii = l-ao_integrals_erf_cache_min
ii = ior(ii, k-ao_integrals_erf_cache_min)
ii = ior(ii, j-ao_integrals_erf_cache_min)
ii = ior(ii, i-ao_integrals_erf_cache_min)
if (iand(ii, -64) /= 0) then
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,idx)
!DIR$ FORCEINLINE
call map_get(map,idx,tmp)
tmp = tmp
else
ii = l-ao_integrals_erf_cache_min
ii = ior( ishft(ii,6), k-ao_integrals_erf_cache_min)
ii = ior( ishft(ii,6), j-ao_integrals_erf_cache_min)
ii = ior( ishft(ii,6), i-ao_integrals_erf_cache_min)
tmp = ao_integrals_erf_cache(ii)
endif
endif
result = tmp
end
subroutine get_ao_bielec_integrals_erf(j,k,l,sze,out_val)
use map_module
BEGIN_DOC
! Gets multiple AO bi-electronic integral from the AO map .
! All i are retrieved for j,k,l fixed.
END_DOC
implicit none
integer, intent(in) :: j,k,l, sze
real(integral_kind), intent(out) :: out_val(sze)
integer :: i
integer(key_kind) :: hash
double precision :: thresh
PROVIDE ao_bielec_integrals_erf_in_map ao_integrals_erf_map
thresh = ao_integrals_threshold
if (ao_overlap_abs(j,l) < thresh) then
out_val = 0.d0
return
endif
double precision :: get_ao_bielec_integral_erf
do i=1,sze
out_val(i) = get_ao_bielec_integral_erf(i,j,k,l,ao_integrals_erf_map)
enddo
end
subroutine get_ao_bielec_integrals_erf_non_zero(j,k,l,sze,out_val,out_val_index,non_zero_int)
use map_module
implicit none
BEGIN_DOC
! Gets multiple AO bi-electronic integral from the AO map .
! All non-zero i are retrieved for j,k,l fixed.
END_DOC
integer, intent(in) :: j,k,l, sze
real(integral_kind), intent(out) :: out_val(sze)
integer, intent(out) :: out_val_index(sze),non_zero_int
integer :: i
integer(key_kind) :: hash
double precision :: thresh,tmp
PROVIDE ao_bielec_integrals_erf_in_map
thresh = ao_integrals_threshold
non_zero_int = 0
if (ao_overlap_abs(j,l) < thresh) then
out_val = 0.d0
return
endif
non_zero_int = 0
do i=1,sze
integer, external :: ao_l4
double precision, external :: ao_bielec_integral_erf
!DIR$ FORCEINLINE
if (ao_bielec_integral_erf_schwartz(i,k)*ao_bielec_integral_erf_schwartz(j,l) < thresh) then
cycle
endif
call bielec_integrals_index(i,j,k,l,hash)
call map_get(ao_integrals_erf_map, hash,tmp)
if (dabs(tmp) < thresh ) cycle
non_zero_int = non_zero_int+1
out_val_index(non_zero_int) = i
out_val(non_zero_int) = tmp
enddo
end
function get_ao_erf_map_size()
implicit none
integer (map_size_kind) :: get_ao_erf_map_size
BEGIN_DOC
! Returns the number of elements in the AO map
END_DOC
get_ao_erf_map_size = ao_integrals_erf_map % n_elements
end
subroutine clear_ao_erf_map
implicit none
BEGIN_DOC
! Frees the memory of the AO map
END_DOC
call map_deinit(ao_integrals_erf_map)
FREE ao_integrals_erf_map
end
subroutine dump_ao_integrals_erf(filename)
use map_module
implicit none
BEGIN_DOC
! Save to disk the $ao integrals
END_DOC
character*(*), intent(in) :: filename
integer(cache_key_kind), pointer :: key(:)
real(integral_kind), pointer :: val(:)
integer*8 :: i,j, n
call ezfio_set_work_empty(.False.)
open(unit=66,file=filename,FORM='unformatted')
write(66) integral_kind, key_kind
write(66) ao_integrals_erf_map%sorted, ao_integrals_erf_map%map_size, &
ao_integrals_erf_map%n_elements
do i=0_8,ao_integrals_erf_map%map_size
write(66) ao_integrals_erf_map%map(i)%sorted, ao_integrals_erf_map%map(i)%map_size,&
ao_integrals_erf_map%map(i)%n_elements
enddo
do i=0_8,ao_integrals_erf_map%map_size
key => ao_integrals_erf_map%map(i)%key
val => ao_integrals_erf_map%map(i)%value
n = ao_integrals_erf_map%map(i)%n_elements
write(66) (key(j), j=1,n), (val(j), j=1,n)
enddo
close(66)
end
integer function load_ao_integrals_erf(filename)
implicit none
BEGIN_DOC
! Read from disk the $ao integrals
END_DOC
character*(*), intent(in) :: filename
integer*8 :: i
integer(cache_key_kind), pointer :: key(:)
real(integral_kind), pointer :: val(:)
integer :: iknd, kknd
integer*8 :: n, j
load_ao_integrals_erf = 1
open(unit=66,file=filename,FORM='unformatted',STATUS='UNKNOWN')
read(66,err=98,end=98) iknd, kknd
if (iknd /= integral_kind) then
print *, 'Wrong integrals kind in file :', iknd
stop 1
endif
if (kknd /= key_kind) then
print *, 'Wrong key kind in file :', kknd
stop 1
endif
read(66,err=98,end=98) ao_integrals_erf_map%sorted, ao_integrals_erf_map%map_size,&
ao_integrals_erf_map%n_elements
do i=0_8, ao_integrals_erf_map%map_size
read(66,err=99,end=99) ao_integrals_erf_map%map(i)%sorted, &
ao_integrals_erf_map%map(i)%map_size, ao_integrals_erf_map%map(i)%n_elements
call cache_map_reallocate(ao_integrals_erf_map%map(i),ao_integrals_erf_map%map(i)%map_size)
enddo
do i=0_8, ao_integrals_erf_map%map_size
key => ao_integrals_erf_map%map(i)%key
val => ao_integrals_erf_map%map(i)%value
n = ao_integrals_erf_map%map(i)%n_elements
read(66,err=99,end=99) (key(j), j=1,n), (val(j), j=1,n)
enddo
call map_sort(ao_integrals_erf_map)
load_ao_integrals_erf = 0
return
99 continue
call map_deinit(ao_integrals_erf_map)
98 continue
stop 'Problem reading ao_integrals_erf_map file in work/'
end

View File

@ -0,0 +1,125 @@
BEGIN_PROVIDER [ logical, ao_bielec_integrals_erf_in_map ]
implicit none
use f77_zmq
use map_module
BEGIN_DOC
! Map of Atomic integrals
! i(r1) j(r2) 1/r12 k(r1) l(r2)
END_DOC
integer :: i,j,k,l
double precision :: ao_bielec_integral_erf,cpu_1,cpu_2, wall_1, wall_2
double precision :: integral, wall_0
include 'utils/constants.include.F'
! For integrals file
integer(key_kind),allocatable :: buffer_i(:)
integer,parameter :: size_buffer = 1024*64
real(integral_kind),allocatable :: buffer_value(:)
integer :: n_integrals, rc
integer :: kk, m, j1, i1, lmax
character*(64) :: fmt
integral = ao_bielec_integral_erf(1,1,1,1)
double precision :: map_mb
PROVIDE read_ao_integrals_erf disk_access_ao_integrals_erf
if (read_ao_integrals_erf) then
print*,'Reading the AO ERF integrals'
call map_load_from_disk(trim(ezfio_filename)//'/work/ao_ints_erf',ao_integrals_erf_map)
print*, 'AO ERF integrals provided'
ao_bielec_integrals_erf_in_map = .True.
return
endif
print*, 'Providing the AO ERF integrals'
call wall_time(wall_0)
call wall_time(wall_1)
call cpu_time(cpu_1)
integer(ZMQ_PTR) :: zmq_to_qp_run_socket, zmq_socket_pull
call new_parallel_job(zmq_to_qp_run_socket,zmq_socket_pull,'ao_integrals_erf')
character(len=:), allocatable :: task
allocate(character(len=ao_num*12) :: task)
write(fmt,*) '(', ao_num, '(I5,X,I5,''|''))'
do l=1,ao_num
write(task,fmt) (i,l, i=1,l)
integer, external :: add_task_to_taskserver
if (add_task_to_taskserver(zmq_to_qp_run_socket,trim(task)) == -1) then
stop 'Unable to add task to server'
endif
enddo
deallocate(task)
integer, external :: zmq_set_running
if (zmq_set_running(zmq_to_qp_run_socket) == -1) then
print *, irp_here, ': Failed in zmq_set_running'
endif
PROVIDE nproc
!$OMP PARALLEL DEFAULT(shared) private(i) num_threads(nproc+1)
i = omp_get_thread_num()
if (i==0) then
call ao_bielec_integrals_erf_in_map_collector(zmq_socket_pull)
else
call ao_bielec_integrals_erf_in_map_slave_inproc(i)
endif
!$OMP END PARALLEL
call end_parallel_job(zmq_to_qp_run_socket, zmq_socket_pull, 'ao_integrals_erf')
print*, 'Sorting the map'
call map_sort(ao_integrals_erf_map)
call cpu_time(cpu_2)
call wall_time(wall_2)
integer(map_size_kind) :: get_ao_erf_map_size, ao_erf_map_size
ao_erf_map_size = get_ao_erf_map_size()
print*, 'AO ERF integrals provided:'
print*, ' Size of AO ERF map : ', map_mb(ao_integrals_erf_map) ,'MB'
print*, ' Number of AO ERF integrals :', ao_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+tiny(1.d0)), ' )'
ao_bielec_integrals_erf_in_map = .True.
if (write_ao_integrals_erf) then
call ezfio_set_work_empty(.False.)
call map_save_to_disk(trim(ezfio_filename)//'/work/ao_ints_erf',ao_integrals_erf_map)
call ezfio_set_ao_two_e_integrals_disk_access_ao_integrals_erf("Read")
endif
END_PROVIDER
BEGIN_PROVIDER [ double precision, ao_bielec_integral_erf_schwartz,(ao_num,ao_num) ]
implicit none
BEGIN_DOC
! Needed to compute Schwartz inequalities
END_DOC
integer :: i,k
double precision :: ao_bielec_integral_erf,cpu_1,cpu_2, wall_1, wall_2
ao_bielec_integral_erf_schwartz(1,1) = ao_bielec_integral_erf(1,1,1,1)
!$OMP PARALLEL DO PRIVATE(i,k) &
!$OMP DEFAULT(NONE) &
!$OMP SHARED (ao_num,ao_bielec_integral_erf_schwartz) &
!$OMP SCHEDULE(dynamic)
do i=1,ao_num
do k=1,i
ao_bielec_integral_erf_schwartz(i,k) = dsqrt(ao_bielec_integral_erf(i,k,i,k))
ao_bielec_integral_erf_schwartz(k,i) = ao_bielec_integral_erf_schwartz(i,k)
enddo
enddo
!$OMP END PARALLEL DO
END_PROVIDER

View File

@ -0,0 +1,55 @@
BEGIN_PROVIDER [ logical, read_ao_integrals ]
&BEGIN_PROVIDER [ logical, write_ao_integrals ]
BEGIN_DOC
! Flag to read or write the |AO| integrals
END_DOC
implicit none
if (disk_access_ao_integrals.EQ.'Read') then
read_ao_integrals = .True.
write_ao_integrals = .False.
else if (disk_access_ao_integrals.EQ.'Write') then
read_ao_integrals = .False.
write_ao_integrals = .True.
else if (disk_access_ao_integrals.EQ.'None') then
read_ao_integrals = .False.
write_ao_integrals = .False.
else
print *, 'disk_access_ao_integrals has a wrong type'
stop 1
endif
END_PROVIDER
BEGIN_PROVIDER [ logical, read_ao_integrals_erf ]
&BEGIN_PROVIDER [ logical, write_ao_integrals_erf ]
implicit none
BEGIN_DOC
! Flag to read or write the |AO| erf integrals
END_DOC
if (disk_access_ao_integrals_erf.EQ.'Read') then
read_ao_integrals_erf = .True.
write_ao_integrals_erf = .False.
else if (disk_access_ao_integrals_erf.EQ.'Write') then
read_ao_integrals_erf = .False.
write_ao_integrals_erf = .True.
else if (disk_access_ao_integrals_erf.EQ.'None') then
read_ao_integrals_erf = .False.
write_ao_integrals_erf = .False.
else
print *, 'disk_access_ao_integrals_erf has a wrong type'
stop 1
endif
END_PROVIDER

View File

@ -0,0 +1,18 @@
subroutine save_erf_bi_elec_integrals_ao
implicit none
integer :: i,j,k,l
PROVIDE ao_bielec_integrals_erf_in_map
call ezfio_set_work_empty(.False.)
call map_save_to_disk(trim(ezfio_filename)//'/work/ao_ints_erf',ao_integrals_erf_map)
call ezfio_set_integrals_bielec_disk_access_ao_integrals("Read")
end
subroutine save_erf_bielec_ints_ao_into_ints_ao
implicit none
integer :: i,j,k,l
PROVIDE ao_bielec_integrals_erf_in_map
call ezfio_set_work_empty(.False.)
call map_save_to_disk(trim(ezfio_filename)//'/work/ao_ints',ao_integrals_erf_map)
call ezfio_set_integrals_bielec_disk_access_ao_integrals("Read")
end

View File

@ -0,0 +1,649 @@
double precision function ao_bielec_integral_erf(i,j,k,l)
implicit none
BEGIN_DOC
! integral of the AO basis <ik|jl> or (ij|kl)
! i(r1) j(r1) 1/r12 k(r2) l(r2)
END_DOC
integer,intent(in) :: i,j,k,l
integer :: p,q,r,s
double precision :: I_center(3),J_center(3),K_center(3),L_center(3)
integer :: num_i,num_j,num_k,num_l,dim1,I_power(3),J_power(3),K_power(3),L_power(3)
double precision :: integral
include 'utils/constants.include.F'
double precision :: P_new(0:max_dim,3),P_center(3),fact_p,pp
double precision :: Q_new(0:max_dim,3),Q_center(3),fact_q,qq
integer :: iorder_p(3), iorder_q(3)
double precision :: ao_bielec_integral_schwartz_accel_erf
if (ao_prim_num(i) * ao_prim_num(j) * ao_prim_num(k) * ao_prim_num(l) > 1024 ) then
ao_bielec_integral_erf = ao_bielec_integral_schwartz_accel_erf(i,j,k,l)
return
endif
dim1 = n_pt_max_integrals
num_i = ao_nucl(i)
num_j = ao_nucl(j)
num_k = ao_nucl(k)
num_l = ao_nucl(l)
ao_bielec_integral_erf = 0.d0
if (num_i /= num_j .or. num_k /= num_l .or. num_j /= num_k)then
do p = 1, 3
I_power(p) = ao_power(i,p)
J_power(p) = ao_power(j,p)
K_power(p) = ao_power(k,p)
L_power(p) = ao_power(l,p)
I_center(p) = nucl_coord(num_i,p)
J_center(p) = nucl_coord(num_j,p)
K_center(p) = nucl_coord(num_k,p)
L_center(p) = nucl_coord(num_l,p)
enddo
double precision :: coef1, coef2, coef3, coef4
double precision :: p_inv,q_inv
double precision :: general_primitive_integral_erf
do p = 1, ao_prim_num(i)
coef1 = ao_coef_normalized_ordered_transp(p,i)
do q = 1, ao_prim_num(j)
coef2 = coef1*ao_coef_normalized_ordered_transp(q,j)
call give_explicit_poly_and_gaussian(P_new,P_center,pp,fact_p,iorder_p,&
ao_expo_ordered_transp(p,i),ao_expo_ordered_transp(q,j), &
I_power,J_power,I_center,J_center,dim1)
p_inv = 1.d0/pp
do r = 1, ao_prim_num(k)
coef3 = coef2*ao_coef_normalized_ordered_transp(r,k)
do s = 1, ao_prim_num(l)
coef4 = coef3*ao_coef_normalized_ordered_transp(s,l)
call give_explicit_poly_and_gaussian(Q_new,Q_center,qq,fact_q,iorder_q,&
ao_expo_ordered_transp(r,k),ao_expo_ordered_transp(s,l), &
K_power,L_power,K_center,L_center,dim1)
q_inv = 1.d0/qq
integral = general_primitive_integral_erf(dim1, &
P_new,P_center,fact_p,pp,p_inv,iorder_p, &
Q_new,Q_center,fact_q,qq,q_inv,iorder_q)
ao_bielec_integral_erf = ao_bielec_integral_erf + coef4 * integral
enddo ! s
enddo ! r
enddo ! q
enddo ! p
else
do p = 1, 3
I_power(p) = ao_power(i,p)
J_power(p) = ao_power(j,p)
K_power(p) = ao_power(k,p)
L_power(p) = ao_power(l,p)
enddo
double precision :: ERI_erf
do p = 1, ao_prim_num(i)
coef1 = ao_coef_normalized_ordered_transp(p,i)
do q = 1, ao_prim_num(j)
coef2 = coef1*ao_coef_normalized_ordered_transp(q,j)
do r = 1, ao_prim_num(k)
coef3 = coef2*ao_coef_normalized_ordered_transp(r,k)
do s = 1, ao_prim_num(l)
coef4 = coef3*ao_coef_normalized_ordered_transp(s,l)
integral = ERI_erf( &
ao_expo_ordered_transp(p,i),ao_expo_ordered_transp(q,j),ao_expo_ordered_transp(r,k),ao_expo_ordered_transp(s,l),&
I_power(1),J_power(1),K_power(1),L_power(1), &
I_power(2),J_power(2),K_power(2),L_power(2), &
I_power(3),J_power(3),K_power(3),L_power(3))
ao_bielec_integral_erf = ao_bielec_integral_erf + coef4 * integral
enddo ! s
enddo ! r
enddo ! q
enddo ! p
endif
end
double precision function ao_bielec_integral_schwartz_accel_erf(i,j,k,l)
implicit none
BEGIN_DOC
! integral of the AO basis <ik|jl> or (ij|kl)
! i(r1) j(r1) 1/r12 k(r2) l(r2)
END_DOC
integer,intent(in) :: i,j,k,l
integer :: p,q,r,s
double precision :: I_center(3),J_center(3),K_center(3),L_center(3)
integer :: num_i,num_j,num_k,num_l,dim1,I_power(3),J_power(3),K_power(3),L_power(3)
double precision :: integral
include 'utils/constants.include.F'
double precision :: P_new(0:max_dim,3),P_center(3),fact_p,pp
double precision :: Q_new(0:max_dim,3),Q_center(3),fact_q,qq
integer :: iorder_p(3), iorder_q(3)
double precision, allocatable :: schwartz_kl(:,:)
double precision :: schwartz_ij
dim1 = n_pt_max_integrals
num_i = ao_nucl(i)
num_j = ao_nucl(j)
num_k = ao_nucl(k)
num_l = ao_nucl(l)
ao_bielec_integral_schwartz_accel_erf = 0.d0
double precision :: thr
thr = ao_integrals_threshold*ao_integrals_threshold
allocate(schwartz_kl(0:ao_prim_num(l),0:ao_prim_num(k)))
double precision :: coef3
double precision :: coef2
double precision :: p_inv,q_inv
double precision :: coef1
double precision :: coef4
if (num_i /= num_j .or. num_k /= num_l .or. num_j /= num_k)then
do p = 1, 3
I_power(p) = ao_power(i,p)
J_power(p) = ao_power(j,p)
K_power(p) = ao_power(k,p)
L_power(p) = ao_power(l,p)
I_center(p) = nucl_coord(num_i,p)
J_center(p) = nucl_coord(num_j,p)
K_center(p) = nucl_coord(num_k,p)
L_center(p) = nucl_coord(num_l,p)
enddo
schwartz_kl(0,0) = 0.d0
do r = 1, ao_prim_num(k)
coef1 = ao_coef_normalized_ordered_transp(r,k)*ao_coef_normalized_ordered_transp(r,k)
schwartz_kl(0,r) = 0.d0
do s = 1, ao_prim_num(l)
coef2 = coef1 * ao_coef_normalized_ordered_transp(s,l) * ao_coef_normalized_ordered_transp(s,l)
call give_explicit_poly_and_gaussian(Q_new,Q_center,qq,fact_q,iorder_q,&
ao_expo_ordered_transp(r,k),ao_expo_ordered_transp(s,l), &
K_power,L_power,K_center,L_center,dim1)
q_inv = 1.d0/qq
schwartz_kl(s,r) = general_primitive_integral_erf(dim1, &
Q_new,Q_center,fact_q,qq,q_inv,iorder_q, &
Q_new,Q_center,fact_q,qq,q_inv,iorder_q) &
* coef2
schwartz_kl(0,r) = max(schwartz_kl(0,r),schwartz_kl(s,r))
enddo
schwartz_kl(0,0) = max(schwartz_kl(0,r),schwartz_kl(0,0))
enddo
do p = 1, ao_prim_num(i)
coef1 = ao_coef_normalized_ordered_transp(p,i)
do q = 1, ao_prim_num(j)
coef2 = coef1*ao_coef_normalized_ordered_transp(q,j)
call give_explicit_poly_and_gaussian(P_new,P_center,pp,fact_p,iorder_p,&
ao_expo_ordered_transp(p,i),ao_expo_ordered_transp(q,j), &
I_power,J_power,I_center,J_center,dim1)
p_inv = 1.d0/pp
schwartz_ij = general_primitive_integral_erf(dim1, &
P_new,P_center,fact_p,pp,p_inv,iorder_p, &
P_new,P_center,fact_p,pp,p_inv,iorder_p) * &
coef2*coef2
if (schwartz_kl(0,0)*schwartz_ij < thr) then
cycle
endif
do r = 1, ao_prim_num(k)
if (schwartz_kl(0,r)*schwartz_ij < thr) then
cycle
endif
coef3 = coef2*ao_coef_normalized_ordered_transp(r,k)
do s = 1, ao_prim_num(l)
if (schwartz_kl(s,r)*schwartz_ij < thr) then
cycle
endif
coef4 = coef3*ao_coef_normalized_ordered_transp(s,l)
double precision :: general_primitive_integral_erf
call give_explicit_poly_and_gaussian(Q_new,Q_center,qq,fact_q,iorder_q,&
ao_expo_ordered_transp(r,k),ao_expo_ordered_transp(s,l), &
K_power,L_power,K_center,L_center,dim1)
q_inv = 1.d0/qq
integral = general_primitive_integral_erf(dim1, &
P_new,P_center,fact_p,pp,p_inv,iorder_p, &
Q_new,Q_center,fact_q,qq,q_inv,iorder_q)
ao_bielec_integral_schwartz_accel_erf = ao_bielec_integral_schwartz_accel_erf + coef4 * integral
enddo ! s
enddo ! r
enddo ! q
enddo ! p
else
do p = 1, 3
I_power(p) = ao_power(i,p)
J_power(p) = ao_power(j,p)
K_power(p) = ao_power(k,p)
L_power(p) = ao_power(l,p)
enddo
double precision :: ERI_erf
schwartz_kl(0,0) = 0.d0
do r = 1, ao_prim_num(k)
coef1 = ao_coef_normalized_ordered_transp(r,k)*ao_coef_normalized_ordered_transp(r,k)
schwartz_kl(0,r) = 0.d0
do s = 1, ao_prim_num(l)
coef2 = coef1*ao_coef_normalized_ordered_transp(s,l)*ao_coef_normalized_ordered_transp(s,l)
schwartz_kl(s,r) = ERI_erf( &
ao_expo_ordered_transp(r,k),ao_expo_ordered_transp(s,l),ao_expo_ordered_transp(r,k),ao_expo_ordered_transp(s,l),&
K_power(1),L_power(1),K_power(1),L_power(1), &
K_power(2),L_power(2),K_power(2),L_power(2), &
K_power(3),L_power(3),K_power(3),L_power(3)) * &
coef2
schwartz_kl(0,r) = max(schwartz_kl(0,r),schwartz_kl(s,r))
enddo
schwartz_kl(0,0) = max(schwartz_kl(0,r),schwartz_kl(0,0))
enddo
do p = 1, ao_prim_num(i)
coef1 = ao_coef_normalized_ordered_transp(p,i)
do q = 1, ao_prim_num(j)
coef2 = coef1*ao_coef_normalized_ordered_transp(q,j)
schwartz_ij = ERI_erf( &
ao_expo_ordered_transp(p,i),ao_expo_ordered_transp(q,j),ao_expo_ordered_transp(p,i),ao_expo_ordered_transp(q,j),&
I_power(1),J_power(1),I_power(1),J_power(1), &
I_power(2),J_power(2),I_power(2),J_power(2), &
I_power(3),J_power(3),I_power(3),J_power(3))*coef2*coef2
if (schwartz_kl(0,0)*schwartz_ij < thr) then
cycle
endif
do r = 1, ao_prim_num(k)
if (schwartz_kl(0,r)*schwartz_ij < thr) then
cycle
endif
coef3 = coef2*ao_coef_normalized_ordered_transp(r,k)
do s = 1, ao_prim_num(l)
if (schwartz_kl(s,r)*schwartz_ij < thr) then
cycle
endif
coef4 = coef3*ao_coef_normalized_ordered_transp(s,l)
integral = ERI_erf( &
ao_expo_ordered_transp(p,i),ao_expo_ordered_transp(q,j),ao_expo_ordered_transp(r,k),ao_expo_ordered_transp(s,l),&
I_power(1),J_power(1),K_power(1),L_power(1), &
I_power(2),J_power(2),K_power(2),L_power(2), &
I_power(3),J_power(3),K_power(3),L_power(3))
ao_bielec_integral_schwartz_accel_erf = ao_bielec_integral_schwartz_accel_erf + coef4 * integral
enddo ! s
enddo ! r
enddo ! q
enddo ! p
endif
deallocate (schwartz_kl)
end
subroutine compute_ao_bielec_integrals_erf(j,k,l,sze,buffer_value)
implicit none
use map_module
BEGIN_DOC
! Compute AO 1/r12 integrals for all i and fixed j,k,l
END_DOC
include 'utils/constants.include.F'
integer, intent(in) :: j,k,l,sze
real(integral_kind), intent(out) :: buffer_value(sze)
double precision :: ao_bielec_integral_erf
integer :: i
if (ao_overlap_abs(j,l) < thresh) then
buffer_value = 0._integral_kind
return
endif
if (ao_bielec_integral_erf_schwartz(j,l) < thresh ) then
buffer_value = 0._integral_kind
return
endif
do i = 1, ao_num
if (ao_overlap_abs(i,k)*ao_overlap_abs(j,l) < thresh) then
buffer_value(i) = 0._integral_kind
cycle
endif
if (ao_bielec_integral_erf_schwartz(i,k)*ao_bielec_integral_erf_schwartz(j,l) < thresh ) then
buffer_value(i) = 0._integral_kind
cycle
endif
!DIR$ FORCEINLINE
buffer_value(i) = ao_bielec_integral_erf(i,k,j,l)
enddo
end
double precision function general_primitive_integral_erf(dim, &
P_new,P_center,fact_p,p,p_inv,iorder_p, &
Q_new,Q_center,fact_q,q,q_inv,iorder_q)
implicit none
BEGIN_DOC
! Computes the integral <pq|rs> where p,q,r,s are Gaussian primitives
END_DOC
integer,intent(in) :: dim
include 'utils/constants.include.F'
double precision, intent(in) :: P_new(0:max_dim,3),P_center(3),fact_p,p,p_inv
double precision, intent(in) :: Q_new(0:max_dim,3),Q_center(3),fact_q,q,q_inv
integer, intent(in) :: iorder_p(3)
integer, intent(in) :: iorder_q(3)
double precision :: r_cut,gama_r_cut,rho,dist
double precision :: dx(0:max_dim),Ix_pol(0:max_dim),dy(0:max_dim),Iy_pol(0:max_dim),dz(0:max_dim),Iz_pol(0:max_dim)
integer :: n_Ix,n_Iy,n_Iz,nx,ny,nz
double precision :: bla
integer :: ix,iy,iz,jx,jy,jz,i
double precision :: a,b,c,d,e,f,accu,pq,const
double precision :: pq_inv, p10_1, p10_2, p01_1, p01_2,pq_inv_2
integer :: n_pt_tmp,n_pt_out, iorder
double precision :: d1(0:max_dim),d_poly(0:max_dim),rint,d1_screened(0:max_dim)
general_primitive_integral_erf = 0.d0
!DIR$ ATTRIBUTES ALIGN : $IRP_ALIGN :: dx,Ix_pol,dy,Iy_pol,dz,Iz_pol
!DIR$ ATTRIBUTES ALIGN : $IRP_ALIGN :: d1, d_poly
! Gaussian Product
! ----------------
double precision :: p_plus_q
p_plus_q = (p+q) * ((p*q)/(p+q) + mu_erf*mu_erf)/(mu_erf*mu_erf)
pq = p_inv*0.5d0*q_inv
pq_inv = 0.5d0/p_plus_q
p10_1 = q*pq ! 1/(2p)
p01_1 = p*pq ! 1/(2q)
pq_inv_2 = pq_inv+pq_inv
p10_2 = pq_inv_2 * p10_1*q !0.5d0*q/(pq + p*p)
p01_2 = pq_inv_2 * p01_1*p !0.5d0*p/(q*q + pq)
accu = 0.d0
iorder = iorder_p(1)+iorder_q(1)+iorder_p(1)+iorder_q(1)
!DIR$ VECTOR ALIGNED
do ix=0,iorder
Ix_pol(ix) = 0.d0
enddo
n_Ix = 0
do ix = 0, iorder_p(1)
if (abs(P_new(ix,1)) < thresh) cycle
a = P_new(ix,1)
do jx = 0, iorder_q(1)
d = a*Q_new(jx,1)
if (abs(d) < thresh) cycle
!DEC$ FORCEINLINE
call give_polynom_mult_center_x(P_center(1),Q_center(1),ix,jx,p,q,iorder,pq_inv,pq_inv_2,p10_1,p01_1,p10_2,p01_2,dx,nx)
!DEC$ FORCEINLINE
call add_poly_multiply(dx,nx,d,Ix_pol,n_Ix)
enddo
enddo
if (n_Ix == -1) then
return
endif
iorder = iorder_p(2)+iorder_q(2)+iorder_p(2)+iorder_q(2)
!DIR$ VECTOR ALIGNED
do ix=0, iorder
Iy_pol(ix) = 0.d0
enddo
n_Iy = 0
do iy = 0, iorder_p(2)
if (abs(P_new(iy,2)) > thresh) then
b = P_new(iy,2)
do jy = 0, iorder_q(2)
e = b*Q_new(jy,2)
if (abs(e) < thresh) cycle
!DEC$ FORCEINLINE
call give_polynom_mult_center_x(P_center(2),Q_center(2),iy,jy,p,q,iorder,pq_inv,pq_inv_2,p10_1,p01_1,p10_2,p01_2,dy,ny)
!DEC$ FORCEINLINE
call add_poly_multiply(dy,ny,e,Iy_pol,n_Iy)
enddo
endif
enddo
if (n_Iy == -1) then
return
endif
iorder = iorder_p(3)+iorder_q(3)+iorder_p(3)+iorder_q(3)
do ix=0,iorder
Iz_pol(ix) = 0.d0
enddo
n_Iz = 0
do iz = 0, iorder_p(3)
if (abs(P_new(iz,3)) > thresh) then
c = P_new(iz,3)
do jz = 0, iorder_q(3)
f = c*Q_new(jz,3)
if (abs(f) < thresh) cycle
!DEC$ FORCEINLINE
call give_polynom_mult_center_x(P_center(3),Q_center(3),iz,jz,p,q,iorder,pq_inv,pq_inv_2,p10_1,p01_1,p10_2,p01_2,dz,nz)
!DEC$ FORCEINLINE
call add_poly_multiply(dz,nz,f,Iz_pol,n_Iz)
enddo
endif
enddo
if (n_Iz == -1) then
return
endif
rho = p*q *pq_inv_2 ! le rho qui va bien
dist = (P_center(1) - Q_center(1))*(P_center(1) - Q_center(1)) + &
(P_center(2) - Q_center(2))*(P_center(2) - Q_center(2)) + &
(P_center(3) - Q_center(3))*(P_center(3) - Q_center(3))
const = dist*rho
n_pt_tmp = n_Ix+n_Iy
do i=0,n_pt_tmp
d_poly(i)=0.d0
enddo
!DEC$ FORCEINLINE
call multiply_poly(Ix_pol,n_Ix,Iy_pol,n_Iy,d_poly,n_pt_tmp)
if (n_pt_tmp == -1) then
return
endif
n_pt_out = n_pt_tmp+n_Iz
do i=0,n_pt_out
d1(i)=0.d0
enddo
!DEC$ FORCEINLINE
call multiply_poly(d_poly ,n_pt_tmp ,Iz_pol,n_Iz,d1,n_pt_out)
double precision :: rint_sum
accu = accu + rint_sum(n_pt_out,const,d1)
! change p+q in dsqrt
general_primitive_integral_erf = fact_p * fact_q * accu *pi_5_2*p_inv*q_inv/dsqrt(p_plus_q)
end
double precision function ERI_erf(alpha,beta,delta,gama,a_x,b_x,c_x,d_x,a_y,b_y,c_y,d_y,a_z,b_z,c_z,d_z)
implicit none
BEGIN_DOC
! ATOMIC PRIMTIVE bielectronic integral between the 4 primitives ::
! primitive_1 = x1**(a_x) y1**(a_y) z1**(a_z) exp(-alpha * r1**2)
! primitive_2 = x1**(b_x) y1**(b_y) z1**(b_z) exp(- beta * r1**2)
! primitive_3 = x2**(c_x) y2**(c_y) z2**(c_z) exp(-delta * r2**2)
! primitive_4 = x2**(d_x) y2**(d_y) z2**(d_z) exp(- gama * r2**2)
END_DOC
double precision, intent(in) :: delta,gama,alpha,beta
integer, intent(in) :: a_x,b_x,c_x,d_x,a_y,b_y,c_y,d_y,a_z,b_z,c_z,d_z
integer :: a_x_2,b_x_2,c_x_2,d_x_2,a_y_2,b_y_2,c_y_2,d_y_2,a_z_2,b_z_2,c_z_2,d_z_2
integer :: i,j,k,l,n_pt
integer :: n_pt_sup
double precision :: p,q,denom,coeff
double precision :: I_f
integer :: nx,ny,nz
include 'utils/constants.include.F'
nx = a_x+b_x+c_x+d_x
if(iand(nx,1) == 1) then
ERI_erf = 0.d0
return
endif
ny = a_y+b_y+c_y+d_y
if(iand(ny,1) == 1) then
ERI_erf = 0.d0
return
endif
nz = a_z+b_z+c_z+d_z
if(iand(nz,1) == 1) then
ERI_erf = 0.d0
return
endif
ASSERT (alpha >= 0.d0)
ASSERT (beta >= 0.d0)
ASSERT (delta >= 0.d0)
ASSERT (gama >= 0.d0)
p = alpha + beta
q = delta + gama
double precision :: p_plus_q
p_plus_q = (p+q) * ((p*q)/(p+q) + mu_erf*mu_erf)/(mu_erf*mu_erf)
ASSERT (p+q >= 0.d0)
n_pt = ishft( nx+ny+nz,1 )
coeff = pi_5_2 / (p * q * dsqrt(p_plus_q))
if (n_pt == 0) then
ERI_erf = coeff
return
endif
call integrale_new_erf(I_f,a_x,b_x,c_x,d_x,a_y,b_y,c_y,d_y,a_z,b_z,c_z,d_z,p,q,n_pt)
ERI_erf = I_f * coeff
end
subroutine integrale_new_erf(I_f,a_x,b_x,c_x,d_x,a_y,b_y,c_y,d_y,a_z,b_z,c_z,d_z,p,q,n_pt)
BEGIN_DOC
! calculate the integral of the polynom ::
! I_x1(a_x+b_x, c_x+d_x,p,q) * I_x1(a_y+b_y, c_y+d_y,p,q) * I_x1(a_z+b_z, c_z+d_z,p,q)
! between ( 0 ; 1)
END_DOC
implicit none
include 'utils/constants.include.F'
double precision :: p,q
integer :: a_x,b_x,c_x,d_x,a_y,b_y,c_y,d_y,a_z,b_z,c_z,d_z
integer :: i, n_pt, j
double precision :: I_f, pq_inv, p10_1, p10_2, p01_1, p01_2,rho,pq_inv_2
integer :: ix,iy,iz, jx,jy,jz, sx,sy,sz
j = ishft(n_pt,-1)
ASSERT (n_pt > 1)
double precision :: p_plus_q
p_plus_q = (p+q) * ((p*q)/(p+q) + mu_erf*mu_erf)/(mu_erf*mu_erf)
pq_inv = 0.5d0/(p_plus_q)
pq_inv_2 = pq_inv + pq_inv
p10_1 = 0.5d0/p
p01_1 = 0.5d0/q
p10_2 = 0.5d0 * q /(p * p_plus_q)
p01_2 = 0.5d0 * p /(q * p_plus_q)
double precision :: B00(n_pt_max_integrals)
double precision :: B10(n_pt_max_integrals), B01(n_pt_max_integrals)
double precision :: t1(n_pt_max_integrals), t2(n_pt_max_integrals)
!DIR$ ATTRIBUTES ALIGN : $IRP_ALIGN :: t1, t2, B10, B01, B00
ix = a_x+b_x
jx = c_x+d_x
iy = a_y+b_y
jy = c_y+d_y
iz = a_z+b_z
jz = c_z+d_z
sx = ix+jx
sy = iy+jy
sz = iz+jz
!DIR$ VECTOR ALIGNED
do i = 1,n_pt
B10(i) = p10_1 - gauleg_t2(i,j)* p10_2
B01(i) = p01_1 - gauleg_t2(i,j)* p01_2
B00(i) = gauleg_t2(i,j)*pq_inv
enddo
if (sx > 0) then
call I_x1_new(ix,jx,B10,B01,B00,t1,n_pt)
else
!DIR$ VECTOR ALIGNED
do i = 1,n_pt
t1(i) = 1.d0
enddo
endif
if (sy > 0) then
call I_x1_new(iy,jy,B10,B01,B00,t2,n_pt)
!DIR$ VECTOR ALIGNED
do i = 1,n_pt
t1(i) = t1(i)*t2(i)
enddo
endif
if (sz > 0) then
call I_x1_new(iz,jz,B10,B01,B00,t2,n_pt)
!DIR$ VECTOR ALIGNED
do i = 1,n_pt
t1(i) = t1(i)*t2(i)
enddo
endif
I_f= 0.d0
!DIR$ VECTOR ALIGNED
do i = 1,n_pt
I_f += gauleg_w(i,j)*t1(i)
enddo
end
subroutine compute_ao_integrals_erf_jl(j,l,n_integrals,buffer_i,buffer_value)
implicit none
use map_module
BEGIN_DOC
! Parallel client for AO integrals
END_DOC
integer, intent(in) :: j,l
integer,intent(out) :: n_integrals
integer(key_kind),intent(out) :: buffer_i(ao_num*ao_num)
real(integral_kind),intent(out) :: buffer_value(ao_num*ao_num)
integer :: i,k
double precision :: ao_bielec_integral_erf,cpu_1,cpu_2, wall_1, wall_2
double precision :: integral, wall_0
double precision :: thr
integer :: kk, m, j1, i1
thr = ao_integrals_threshold
n_integrals = 0
j1 = j+ishft(l*l-l,-1)
do k = 1, ao_num ! r1
i1 = ishft(k*k-k,-1)
if (i1 > j1) then
exit
endif
do i = 1, k
i1 += 1
if (i1 > j1) then
exit
endif
if (ao_overlap_abs(i,k)*ao_overlap_abs(j,l) < thr) then
cycle
endif
if (ao_bielec_integral_erf_schwartz(i,k)*ao_bielec_integral_erf_schwartz(j,l) < thr ) then
cycle
endif
!DIR$ FORCEINLINE
integral = ao_bielec_integral_erf(i,k,j,l) ! i,k : r1 j,l : r2
if (abs(integral) < thr) then
cycle
endif
n_integrals += 1
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,buffer_i(n_integrals))
buffer_value(n_integrals) = integral
enddo
enddo
end

View File

@ -40,12 +40,6 @@ doc: Thresholds on generators (fraction of the square of the norm)
interface: ezfio,provider,ocaml
default: 0.99
[threshold_selectors]
type: Threshold
doc: Thresholds on selectors (fraction of the square of the norm)
interface: ezfio,provider,ocaml
default: 0.999
[n_int]
interface: ezfio
doc: Number of integers required to represent bitstrings (set in module :ref:`bitmask`)

View File

@ -1,3 +1,3 @@
mo_basis
mo_one_e_integrals
integrals_bielec
mo_two_e_integrals

View File

@ -16,14 +16,6 @@ doc: Percentage of HF exchange in the DFT model
interface: ezfio,provider,ocaml
default: 0.
[mu_erf]
type: double precision
doc: cutting of the interaction in the range separated model
interface: ezfio,provider,ocaml
default: 0.5
ezfio_name: mu_erf
[density_for_dft]
type: character*(32)
doc: Type of density used for DFT calculation. If WFT it uses the density of the WFT stored in terms of determinants. If input_density it uses the one-body dm stored in data_.../ . If damping_rs_dft it uses the damping density between WFT and input_density

View File

@ -1,4 +1,4 @@
density_for_dft
dft_utils_on_grid
integrals_bielec
integrals_bielec_erf
mo_one_e_integrals
mo_two_e_integrals

View File

@ -0,0 +1,125 @@
BEGIN_PROVIDER [double precision, potential_x_alpha_ao,(ao_num,ao_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_x_beta_ao,(ao_num,ao_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_c_alpha_ao,(ao_num,ao_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_c_beta_ao,(ao_num,ao_num,N_states)]
implicit none
BEGIN_DOC
! alpha/beta exchange/correlation potentials on the AO basis
END_DOC
if(trim(exchange_functional)=="short_range_LDA")then
potential_x_alpha_ao = potential_x_alpha_ao_LDA
potential_x_beta_ao = potential_x_beta_ao_LDA
else if(exchange_functional.EQ."short_range_PBE")then
potential_x_alpha_ao = potential_x_alpha_ao_PBE
potential_x_beta_ao = potential_x_beta_ao_PBE
else if(exchange_functional.EQ."None")then
potential_x_alpha_ao = 0.d0
potential_x_beta_ao = 0.d0
else
print*, 'Exchange functional required does not exist ...'
print*,'exchange_functional',exchange_functional
stop
endif
if(trim(correlation_functional)=="short_range_LDA")then
potential_c_alpha_ao = potential_c_alpha_ao_LDA
potential_c_beta_ao = potential_c_beta_ao_LDA
else if(correlation_functional.EQ."short_range_PBE")then
potential_c_alpha_ao = potential_c_alpha_ao_PBE
potential_c_beta_ao = potential_c_beta_ao_PBE
else if(correlation_functional.EQ."None")then
potential_c_alpha_ao = 0.d0
potential_c_beta_ao = 0.d0
else
print*, 'Correlation functional required does not ecist ...'
print*,'correlation_functional',correlation_functional
stop
endif
END_PROVIDER
BEGIN_PROVIDER [double precision, potential_x_alpha_mo,(mo_tot_num,mo_tot_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_x_beta_mo,(mo_tot_num,mo_tot_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_c_alpha_mo,(mo_tot_num,mo_tot_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_c_beta_mo,(mo_tot_num,mo_tot_num,N_states)]
implicit none
BEGIN_DOC
! alpha/beta exchange/correlation potentials on the MO basis
END_DOC
integer :: istate
do istate = 1, N_states
call ao_to_mo( &
potential_x_alpha_ao(1,1,istate), &
size(potential_x_alpha_ao,1), &
potential_x_alpha_mo(1,1,istate), &
size(potential_x_alpha_mo,1) &
)
call ao_to_mo( &
potential_x_beta_ao(1,1,istate), &
size(potential_x_beta_ao,1), &
potential_x_beta_mo(1,1,istate), &
size(potential_x_beta_mo,1) &
)
call ao_to_mo( &
potential_c_alpha_ao(1,1,istate), &
size(potential_c_alpha_ao,1), &
potential_c_alpha_mo(1,1,istate), &
size(potential_c_alpha_mo,1) &
)
call ao_to_mo( &
potential_c_beta_ao(1,1,istate), &
size(potential_c_beta_ao,1), &
potential_c_beta_mo(1,1,istate), &
size(potential_c_beta_mo,1) &
)
enddo
END_PROVIDER
BEGIN_PROVIDER [double precision, energy_x, (N_states)]
&BEGIN_PROVIDER [double precision, energy_c, (N_states)]
implicit none
if(trim(exchange_functional)=="short_range_LDA")then
energy_x = energy_x_LDA
energy_x = energy_x_LDA
else if(exchange_functional.EQ."short_range_PBE")then
energy_x = energy_x_PBE
energy_x = energy_x_PBE
else if(exchange_functional.EQ."None")then
energy_x = 0.d0
energy_x = 0.d0
else
print*, 'Exchange functional required does not exist ...'
print*,'exchange_functional',exchange_functional
stop
endif
if(trim(correlation_functional)=="short_range_LDA")then
energy_c = energy_c_LDA
energy_c = energy_c_LDA
else if(correlation_functional.EQ."short_range_PBE")then
energy_c = energy_c_PBE
energy_c = energy_c_PBE
else if(correlation_functional.EQ."None")then
energy_c = 0.d0
energy_c = 0.d0
else
print*, 'Correlation functional required does not ecist ...'
print*,'correlation_functional',correlation_functional
stop
endif
END_PROVIDER

View File

@ -0,0 +1,29 @@
BEGIN_PROVIDER [double precision, psi_dft_energy_kinetic, (N_states) ]
&BEGIN_PROVIDER [double precision, psi_dft_energy_nuclear_elec, (N_states) ]
&BEGIN_PROVIDER [double precision, psi_dft_energy_h_core, (N_states) ]
implicit none
BEGIN_DOC
! kinetic, electron-nuclear and total h_core energy computed with the density matrix one_body_dm_mo_beta_for_dft+one_body_dm_mo_alpha_for_dft
END_DOC
integer :: i,j,istate
double precision :: accu
psi_dft_energy_kinetic = 0.d0
psi_dft_energy_nuclear_elec = 0.d0
do istate = 1, N_states
do i = 1, mo_tot_num
do j = 1, mo_tot_num
psi_dft_energy_kinetic(istate) += ( one_body_dm_mo_alpha_for_dft(j,i,istate)+one_body_dm_mo_beta_for_dft(j,i,istate)) * mo_kinetic_integral(j,i)
psi_dft_energy_nuclear_elec(istate) += ( one_body_dm_mo_alpha_for_dft(j,i,istate)+one_body_dm_mo_beta_for_dft(j,i,istate)) * mo_nucl_elec_integral(j,i)
enddo
enddo
enddo
do i = 1, N_states
do j = 1, mo_tot_num
accu += one_body_dm_mo_alpha_for_dft(j,j,i) + one_body_dm_mo_beta_for_dft(j,j,i)
enddo
accu = (elec_alpha_num + elec_beta_num ) / accu
psi_energy_h_core(i) = psi_dft_energy_h_core(i) * accu
enddo
END_PROVIDER

View File

@ -0,0 +1,272 @@
BEGIN_PROVIDER[double precision, aos_vc_alpha_LDA_w, (n_points_final_grid,ao_num,N_states)]
&BEGIN_PROVIDER[double precision, aos_vc_beta_LDA_w, (n_points_final_grid,ao_num,N_states)]
&BEGIN_PROVIDER[double precision, aos_vx_alpha_LDA_w, (n_points_final_grid,ao_num,N_states)]
&BEGIN_PROVIDER[double precision, aos_vx_beta_LDA_w, (n_points_final_grid,ao_num,N_states)]
implicit none
BEGIN_DOC
! aos_vxc_alpha_LDA_w(j,i) = ao_i(r_j) * (v^x_alpha(r_j) + v^c_alpha(r_j)) * W(r_j)
END_DOC
integer :: istate,i,j
double precision :: r(3)
double precision :: mu,weight
double precision :: e_c,vc_a,vc_b,e_x,vx_a,vx_b
double precision, allocatable :: rhoa(:),rhob(:)
allocate(rhoa(N_states), rhob(N_states))
do istate = 1, N_states
do i = 1, n_points_final_grid
r(1) = final_grid_points(1,i)
r(2) = final_grid_points(2,i)
r(3) = final_grid_points(3,i)
weight=final_weight_functions_at_final_grid_points(i)
rhoa(istate) = one_body_dm_alpha_at_r(i,istate)
rhob(istate) = one_body_dm_beta_at_r(i,istate)
call ec_LDA_sr(mu_erf,rhoa(istate),rhob(istate),e_c,vc_a,vc_b)
call ex_LDA_sr(mu_erf,rhoa(istate),rhob(istate),e_x,vx_a,vx_b)
do j =1, ao_num
aos_vc_alpha_LDA_w(i,j,istate) = vc_a * aos_in_r_array(j,i)*weight
aos_vc_beta_LDA_w(i,j,istate) = vc_b * aos_in_r_array(j,i)*weight
aos_vx_alpha_LDA_w(i,j,istate) = vx_a * aos_in_r_array(j,i)*weight
aos_vx_beta_LDA_w(i,j,istate) = vx_b * aos_in_r_array(j,i)*weight
enddo
enddo
enddo
END_PROVIDER
BEGIN_PROVIDER[double precision, energy_x_LDA, (N_states) ]
&BEGIN_PROVIDER[double precision, energy_c_LDA, (N_states) ]
implicit none
BEGIN_DOC
! exchange/correlation energy with the short range LDA functional
END_DOC
integer :: istate,i,j
double precision :: r(3)
double precision :: mu,weight
double precision :: e_c,vc_a,vc_b,e_x,vx_a,vx_b
double precision, allocatable :: rhoa(:),rhob(:)
allocate(rhoa(N_states), rhob(N_states))
energy_x_LDA = 0.d0
energy_c_LDA = 0.d0
do istate = 1, N_states
do i = 1, n_points_final_grid
r(1) = final_grid_points(1,i)
r(2) = final_grid_points(2,i)
r(3) = final_grid_points(3,i)
weight=final_weight_functions_at_final_grid_points(i)
rhoa(istate) = one_body_dm_alpha_at_r(i,istate)
rhob(istate) = one_body_dm_beta_at_r(i,istate)
call ec_LDA_sr(mu_erf,rhoa(istate),rhob(istate),e_c,vc_a,vc_b)
call ex_LDA_sr(mu_erf,rhoa(istate),rhob(istate),e_x,vx_a,vx_b)
energy_x_LDA(istate) += weight * e_x
energy_c_LDA(istate) += weight * e_c
enddo
enddo
END_PROVIDER
BEGIN_PROVIDER [double precision, potential_x_alpha_ao_LDA,(ao_num,ao_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_x_beta_ao_LDA,(ao_num,ao_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_c_alpha_ao_LDA,(ao_num,ao_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_c_beta_ao_LDA,(ao_num,ao_num,N_states)]
implicit none
BEGIN_DOC
! short range exchange/correlation alpha/beta potentials with LDA functional on the AO basis
END_DOC
integer :: istate
double precision :: wall_1,wall_2
call wall_time(wall_1)
do istate = 1, N_states
call dgemm('N','N',ao_num,ao_num,n_points_final_grid,1.d0,aos_in_r_array,ao_num,aos_vc_alpha_LDA_w(1,1,istate),n_points_final_grid,0.d0,potential_c_alpha_ao_LDA(1,1,istate),ao_num)
call dgemm('N','N',ao_num,ao_num,n_points_final_grid,1.d0,aos_in_r_array,ao_num,aos_vc_beta_LDA_w(1,1,istate) ,n_points_final_grid,0.d0,potential_c_beta_ao_LDA(1,1,istate),ao_num)
call dgemm('N','N',ao_num,ao_num,n_points_final_grid,1.d0,aos_in_r_array,ao_num,aos_vx_alpha_LDA_w(1,1,istate),n_points_final_grid,0.d0,potential_x_alpha_ao_LDA(1,1,istate),ao_num)
call dgemm('N','N',ao_num,ao_num,n_points_final_grid,1.d0,aos_in_r_array,ao_num,aos_vx_beta_LDA_w(1,1,istate) ,n_points_final_grid,0.d0,potential_x_beta_ao_LDA(1,1,istate),ao_num)
enddo
call wall_time(wall_2)
print*,'time to provide potential_x/c_alpha/beta_ao_LDA = ',wall_2 - wall_1
END_PROVIDER
BEGIN_PROVIDER[double precision, aos_vc_alpha_PBE_w , (ao_num,n_points_final_grid,N_states)] !(n_points_final_grid,ao_num,N_states)]
&BEGIN_PROVIDER[double precision, aos_vc_beta_PBE_w , (ao_num,n_points_final_grid,N_states)]!(n_points_final_grid,ao_num,N_states)]
&BEGIN_PROVIDER[double precision, aos_vx_alpha_PBE_w , (ao_num,n_points_final_grid,N_states)] !(n_points_final_grid,ao_num,N_states)]
&BEGIN_PROVIDER[double precision, aos_vx_beta_PBE_w , (ao_num,n_points_final_grid,N_states)]!(n_points_final_grid,ao_num,N_states)]
&BEGIN_PROVIDER[double precision, aos_dvc_alpha_PBE_w , (ao_num,n_points_final_grid,3,N_states)]
&BEGIN_PROVIDER[double precision, aos_dvc_beta_PBE_w , (ao_num,n_points_final_grid,3,N_states)]
&BEGIN_PROVIDER[double precision, aos_dvx_alpha_PBE_w , (ao_num,n_points_final_grid,3,N_states)]
&BEGIN_PROVIDER[double precision, aos_dvx_beta_PBE_w , (ao_num,n_points_final_grid,3,N_states)]
&BEGIN_PROVIDER[double precision, grad_aos_dvc_alpha_PBE_w , (ao_num,n_points_final_grid,3,N_states)]
&BEGIN_PROVIDER[double precision, grad_aos_dvc_beta_PBE_w , (ao_num,n_points_final_grid,3,N_states)]
&BEGIN_PROVIDER[double precision, grad_aos_dvx_alpha_PBE_w , (ao_num,n_points_final_grid,3,N_states)]
&BEGIN_PROVIDER[double precision, grad_aos_dvx_beta_PBE_w , (ao_num,n_points_final_grid,3,N_states)]
implicit none
BEGIN_DOC
! aos_vxc_alpha_PBE_w(j,i) = ao_i(r_j) * (v^x_alpha(r_j) + v^c_alpha(r_j)) * W(r_j)
END_DOC
integer :: istate,i,j,m
double precision :: r(3)
double precision :: mu,weight
double precision, allocatable :: ex(:), ec(:)
double precision, allocatable :: rho_a(:),rho_b(:),grad_rho_a(:,:),grad_rho_b(:,:),grad_rho_a_2(:),grad_rho_b_2(:),grad_rho_a_b(:)
double precision, allocatable :: contrib_grad_xa(:,:),contrib_grad_xb(:,:),contrib_grad_ca(:,:),contrib_grad_cb(:,:)
double precision, allocatable :: vc_rho_a(:), vc_rho_b(:), vx_rho_a(:), vx_rho_b(:)
double precision, allocatable :: vx_grad_rho_a_2(:), vx_grad_rho_b_2(:), vx_grad_rho_a_b(:), vc_grad_rho_a_2(:), vc_grad_rho_b_2(:), vc_grad_rho_a_b(:)
allocate(vc_rho_a(N_states), vc_rho_b(N_states), vx_rho_a(N_states), vx_rho_b(N_states))
allocate(vx_grad_rho_a_2(N_states), vx_grad_rho_b_2(N_states), vx_grad_rho_a_b(N_states), vc_grad_rho_a_2(N_states), vc_grad_rho_b_2(N_states), vc_grad_rho_a_b(N_states))
allocate(rho_a(N_states), rho_b(N_states),grad_rho_a(3,N_states),grad_rho_b(3,N_states))
allocate(grad_rho_a_2(N_states),grad_rho_b_2(N_states),grad_rho_a_b(N_states), ex(N_states), ec(N_states))
allocate(contrib_grad_xa(3,N_states),contrib_grad_xb(3,N_states),contrib_grad_ca(3,N_states),contrib_grad_cb(3,N_states))
do istate = 1, N_states
do i = 1, n_points_final_grid
r(1) = final_grid_points(1,i)
r(2) = final_grid_points(2,i)
r(3) = final_grid_points(3,i)
weight=final_weight_functions_at_final_grid_points(i)
rho_a(istate) = one_body_dm_alpha_and_grad_at_r(4,i,istate)
rho_b(istate) = one_body_dm_beta_and_grad_at_r(4,i,istate)
grad_rho_a(1:3,istate) = one_body_dm_alpha_and_grad_at_r(1:3,i,istate)
grad_rho_b(1:3,istate) = one_body_dm_beta_and_grad_at_r(1:3,i,istate)
grad_rho_a_2 = 0.d0
grad_rho_b_2 = 0.d0
grad_rho_a_b = 0.d0
do m = 1, 3
grad_rho_a_2(istate) += grad_rho_a(m,istate) * grad_rho_a(m,istate)
grad_rho_b_2(istate) += grad_rho_b(m,istate) * grad_rho_b(m,istate)
grad_rho_a_b(istate) += grad_rho_a(m,istate) * grad_rho_b(m,istate)
enddo
! inputs
call GGA_type_functionals(r,rho_a,rho_b,grad_rho_a_2,grad_rho_b_2,grad_rho_a_b, & ! outputs exchange
ex,vx_rho_a,vx_rho_b,vx_grad_rho_a_2,vx_grad_rho_b_2,vx_grad_rho_a_b, & ! outputs correlation
ec,vc_rho_a,vc_rho_b,vc_grad_rho_a_2,vc_grad_rho_b_2,vc_grad_rho_a_b )
vx_rho_a(istate) *= weight
vc_rho_a(istate) *= weight
vx_rho_b(istate) *= weight
vc_rho_b(istate) *= weight
do m= 1,3
contrib_grad_ca(m,istate) = weight * (2.d0 * vc_grad_rho_a_2(istate) * grad_rho_a(m,istate) + vc_grad_rho_a_b(istate) * grad_rho_b(m,istate))
contrib_grad_xa(m,istate) = weight * (2.d0 * vx_grad_rho_a_2(istate) * grad_rho_a(m,istate) + vx_grad_rho_a_b(istate) * grad_rho_b(m,istate))
contrib_grad_cb(m,istate) = weight * (2.d0 * vc_grad_rho_b_2(istate) * grad_rho_b(m,istate) + vc_grad_rho_a_b(istate) * grad_rho_a(m,istate))
contrib_grad_xb(m,istate) = weight * (2.d0 * vx_grad_rho_b_2(istate) * grad_rho_b(m,istate) + vx_grad_rho_a_b(istate) * grad_rho_a(m,istate))
enddo
do j = 1, ao_num
aos_vc_alpha_PBE_w(j,i,istate) = vc_rho_a(istate) * aos_in_r_array(j,i)
aos_vc_beta_PBE_w (j,i,istate) = vc_rho_b(istate) * aos_in_r_array(j,i)
aos_vx_alpha_PBE_w(j,i,istate) = vx_rho_a(istate) * aos_in_r_array(j,i)
aos_vx_beta_PBE_w (j,i,istate) = vx_rho_b(istate) * aos_in_r_array(j,i)
do m = 1,3
aos_dvc_alpha_PBE_w(j,i,m,istate) = contrib_grad_ca(m,istate) * aos_in_r_array(j,i)
aos_dvc_beta_PBE_w (j,i,m,istate) = contrib_grad_cb(m,istate) * aos_in_r_array(j,i)
aos_dvx_alpha_PBE_w(j,i,m,istate) = contrib_grad_xa(m,istate) * aos_in_r_array(j,i)
aos_dvx_beta_PBE_w (j,i,m,istate) = contrib_grad_xb(m,istate) * aos_in_r_array(j,i)
grad_aos_dvc_alpha_PBE_w (j,i,m,istate) = contrib_grad_ca(m,istate) * aos_grad_in_r_array(j,i,m)
grad_aos_dvc_beta_PBE_w (j,i,m,istate) = contrib_grad_cb(m,istate) * aos_grad_in_r_array(j,i,m)
grad_aos_dvx_alpha_PBE_w (j,i,m,istate) = contrib_grad_xa(m,istate) * aos_grad_in_r_array(j,i,m)
grad_aos_dvx_beta_PBE_w (j,i,m,istate) = contrib_grad_xb(m,istate) * aos_grad_in_r_array(j,i,m)
enddo
enddo
enddo
enddo
END_PROVIDER
BEGIN_PROVIDER[double precision, energy_x_PBE, (N_states) ]
&BEGIN_PROVIDER[double precision, energy_c_PBE, (N_states) ]
implicit none
BEGIN_DOC
! exchange/correlation energy with the short range PBE functional
END_DOC
integer :: istate,i,j,m
double precision :: r(3)
double precision :: mu,weight
double precision, allocatable :: ex(:), ec(:)
double precision, allocatable :: rho_a(:),rho_b(:),grad_rho_a(:,:),grad_rho_b(:,:),grad_rho_a_2(:),grad_rho_b_2(:),grad_rho_a_b(:)
double precision, allocatable :: contrib_grad_xa(:,:),contrib_grad_xb(:,:),contrib_grad_ca(:,:),contrib_grad_cb(:,:)
double precision, allocatable :: vc_rho_a(:), vc_rho_b(:), vx_rho_a(:), vx_rho_b(:)
double precision, allocatable :: vx_grad_rho_a_2(:), vx_grad_rho_b_2(:), vx_grad_rho_a_b(:), vc_grad_rho_a_2(:), vc_grad_rho_b_2(:), vc_grad_rho_a_b(:)
allocate(vc_rho_a(N_states), vc_rho_b(N_states), vx_rho_a(N_states), vx_rho_b(N_states))
allocate(vx_grad_rho_a_2(N_states), vx_grad_rho_b_2(N_states), vx_grad_rho_a_b(N_states), vc_grad_rho_a_2(N_states), vc_grad_rho_b_2(N_states), vc_grad_rho_a_b(N_states))
allocate(rho_a(N_states), rho_b(N_states),grad_rho_a(3,N_states),grad_rho_b(3,N_states))
allocate(grad_rho_a_2(N_states),grad_rho_b_2(N_states),grad_rho_a_b(N_states), ex(N_states), ec(N_states))
energy_x_PBE = 0.d0
energy_c_PBE = 0.d0
do istate = 1, N_states
do i = 1, n_points_final_grid
r(1) = final_grid_points(1,i)
r(2) = final_grid_points(2,i)
r(3) = final_grid_points(3,i)
weight=final_weight_functions_at_final_grid_points(i)
rho_a(istate) = one_body_dm_alpha_and_grad_at_r(4,i,istate)
rho_b(istate) = one_body_dm_beta_and_grad_at_r(4,i,istate)
grad_rho_a(1:3,istate) = one_body_dm_alpha_and_grad_at_r(1:3,i,istate)
grad_rho_b(1:3,istate) = one_body_dm_beta_and_grad_at_r(1:3,i,istate)
grad_rho_a_2 = 0.d0
grad_rho_b_2 = 0.d0
grad_rho_a_b = 0.d0
do m = 1, 3
grad_rho_a_2(istate) += grad_rho_a(m,istate) * grad_rho_a(m,istate)
grad_rho_b_2(istate) += grad_rho_b(m,istate) * grad_rho_b(m,istate)
grad_rho_a_b(istate) += grad_rho_a(m,istate) * grad_rho_b(m,istate)
enddo
! inputs
call GGA_type_functionals(r,rho_a,rho_b,grad_rho_a_2,grad_rho_b_2,grad_rho_a_b, & ! outputs exchange
ex,vx_rho_a,vx_rho_b,vx_grad_rho_a_2,vx_grad_rho_b_2,vx_grad_rho_a_b, & ! outputs correlation
ec,vc_rho_a,vc_rho_b,vc_grad_rho_a_2,vc_grad_rho_b_2,vc_grad_rho_a_b )
energy_x_PBE += ex * weight
energy_c_PBE += ec * weight
enddo
enddo
END_PROVIDER
BEGIN_PROVIDER [double precision, potential_x_alpha_ao_PBE,(ao_num,ao_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_x_beta_ao_PBE,(ao_num,ao_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_c_alpha_ao_PBE,(ao_num,ao_num,N_states)]
&BEGIN_PROVIDER [double precision, potential_c_beta_ao_PBE,(ao_num,ao_num,N_states)]
implicit none
BEGIN_DOC
! exchange/correlation alpha/beta potentials with the short range PBE functional on the AO basis
END_DOC
integer :: istate, m
double precision :: wall_1,wall_2
call wall_time(wall_1)
potential_c_alpha_ao_PBE = 0.d0
potential_x_alpha_ao_PBE = 0.d0
potential_c_beta_ao_PBE = 0.d0
potential_x_beta_ao_PBE = 0.d0
do istate = 1, N_states
! correlation alpha
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,aos_vc_alpha_PBE_w(1,1,istate),size(aos_vc_alpha_PBE_w,1),aos_in_r_array,size(aos_in_r_array,1),1.d0,potential_c_alpha_ao_PBE(1,1,istate),size(potential_c_alpha_ao_PBE,1))
! correlation beta
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,aos_vc_beta_PBE_w(1,1,istate),size(aos_vc_beta_PBE_w,1),aos_in_r_array,size(aos_in_r_array,1),1.d0,potential_c_beta_ao_PBE(1,1,istate),size(potential_c_beta_ao_PBE,1))
! exchange alpha
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,aos_vx_alpha_PBE_w(1,1,istate),size(aos_vx_alpha_PBE_w,1),aos_in_r_array,size(aos_in_r_array,1),1.d0,potential_x_alpha_ao_PBE(1,1,istate),size(potential_x_alpha_ao_PBE,1))
! exchange beta
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,aos_vx_beta_PBE_w(1,1,istate),size(aos_vx_beta_PBE_w,1), aos_in_r_array,size(aos_in_r_array,1),1.d0,potential_x_beta_ao_PBE(1,1,istate), size(potential_x_beta_ao_PBE,1))
do m= 1,3
! correlation alpha
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,aos_dvc_alpha_PBE_w(1,1,m,istate),size(aos_dvc_alpha_PBE_w,1),aos_grad_in_r_array(1,1,m),size(aos_grad_in_r_array,1),1.d0,potential_c_alpha_ao_PBE(1,1,istate),size(potential_c_alpha_ao_PBE,1))
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,grad_aos_dvc_alpha_PBE_w(1,1,m,istate),size(grad_aos_dvc_alpha_PBE_w,1),aos_in_r_array,size(aos_in_r_array,1),1.d0,potential_c_alpha_ao_PBE(1,1,istate),size(potential_c_alpha_ao_PBE,1))
! correlation beta
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,aos_dvc_beta_PBE_w(1,1,m,istate),size(aos_dvc_beta_PBE_w,1),aos_grad_in_r_array(1,1,m),size(aos_grad_in_r_array,1),1.d0,potential_c_beta_ao_PBE(1,1,istate),size(potential_c_beta_ao_PBE,1))
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,grad_aos_dvc_beta_PBE_w(1,1,m,istate),size(grad_aos_dvc_beta_PBE_w,1),aos_in_r_array,size(aos_in_r_array,1),1.d0,potential_c_beta_ao_PBE(1,1,istate),size(potential_c_beta_ao_PBE,1))
! exchange alpha
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,aos_dvx_alpha_PBE_w(1,1,m,istate),size(aos_dvx_alpha_PBE_w,1),aos_grad_in_r_array(1,1,m),size(aos_grad_in_r_array,1),1.d0,potential_x_alpha_ao_PBE(1,1,istate),size(potential_x_alpha_ao_PBE,1))
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,grad_aos_dvx_alpha_PBE_w(1,1,m,istate),size(grad_aos_dvx_alpha_PBE_w,1),aos_in_r_array,size(aos_in_r_array,1),1.d0,potential_x_alpha_ao_PBE(1,1,istate),size(potential_x_alpha_ao_PBE,1))
! exchange beta
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,aos_dvx_beta_PBE_w(1,1,m,istate),size(aos_dvx_beta_PBE_w,1),aos_grad_in_r_array(1,1,m),size(aos_grad_in_r_array,1),1.d0,potential_x_beta_ao_PBE(1,1,istate),size(potential_x_beta_ao_PBE,1))
call dgemm('N','T',ao_num,ao_num,n_points_final_grid,1.d0,grad_aos_dvx_beta_PBE_w(1,1,m,istate),size(grad_aos_dvx_beta_PBE_w,1),aos_in_r_array,size(aos_in_r_array,1),1.d0,potential_x_beta_ao_PBE(1,1,istate),size(potential_x_beta_ao_PBE,1))
enddo
enddo
call wall_time(wall_2)
END_PROVIDER

View File

@ -0,0 +1,16 @@
BEGIN_PROVIDER [double precision, shifting_constant, (N_states)]
implicit none
BEGIN_DOC
! shifting_constant = (E_{Hxc} - <\Psi | V_{Hxc} | \Psi>) / N_elec
! constant to add to the potential in order to obtain the variational energy as
! the eigenvalue of the effective long-range Hamiltonian
! (see original paper of Levy PRL 113, 113002 (2014), equation (17) )
END_DOC
integer :: istate
do istate = 1, N_states
shifting_constant(istate) = energy_x(istate) + energy_c(istate) + short_range_Hartree(istate) - Trace_v_Hxc(istate)
enddo
shifting_constant = shifting_constant / dble(elec_num)
END_PROVIDER

View File

@ -42,10 +42,9 @@ END_PROVIDER
effective_one_e_potential = 0.d0
BEGIN_DOC
! effective_one_e_potential(i,j) = <i| v_{H}^{sr} |j> + <i| h_{core} |j> + <i|v_{xc} |j>
!
! Taking the expectation value does not provide any energy
!
! but effective_one_e_potential(i,j) is the potential coupling DFT and WFT part to be used in any WFT calculation
! shifted_effective_one_e_potential_without_kin = effective_one_e_potential_without_kin + shifting_constant on the diagonal
END_DOC
do istate = 1, N_states
do i = 1, mo_tot_num
@ -61,3 +60,4 @@ END_PROVIDER
enddo
END_PROVIDER

View File

@ -5,10 +5,9 @@ subroutine dress_slave
END_DOC
read_wf = .False.
distributed_davidson = .False.
SOFT_TOUCH read_wf distributed_davidson
threshold_selectors = 1.d0
threshold_generators = 1d0
SOFT_TOUCH read_wf distributed_davidson threshold_generators
call provide_everything
call switch_qp_run_to_master

View File

@ -290,16 +290,13 @@ subroutine ZMQ_dress(E, dress, delta_out, delta_s2_out, relative_error)
if (zmq_put_int(zmq_to_qp_run_socket,1,'dress_stoch_istate',dress_stoch_istate) == -1) then
stop 'Unable to put dress_stoch_istate on ZMQ server'
endif
if (zmq_put_dvector(zmq_to_qp_run_socket,1,'threshold_selectors',threshold_selectors,1) == -1) then
stop 'Unable to put threshold_selectors on ZMQ server'
endif
if (zmq_put_dvector(zmq_to_qp_run_socket,1,'threshold_generators',threshold_generators,1) == -1) then
stop 'Unable to put threshold_generators on ZMQ server'
endif
if (zmq_put_dvector(zmq_to_qp_run_socket,1,'threshold_generators',threshold_generators,1) == -1) then
stop 'Unable to put threshold_generators on ZMQ server'
endif
if (zmq_put_int(zmq_to_qp_run_socket, 1, 'ending', (-1)) == -1) then
stop 'Unable to put initial ending'
endif
if (zmq_put_int(zmq_to_qp_run_socket, 1, 'ending', (-1)) == -1) then
stop 'Unable to put initial ending'
endif

View File

@ -2,11 +2,10 @@ subroutine dress_zmq()
implicit none
double precision, allocatable :: energy(:)
allocate (energy(N_states))
threshold_selectors = 1.d0
threshold_generators = 1d0
read_wf = .True.
SOFT_TOUCH read_wf threshold_generators threshold_selectors
SOFT_TOUCH read_wf threshold_generators
if (.True.) then
integer :: i,j

View File

@ -1,5 +1,6 @@
ao_basis
ao_one_e_integrals
ao_two_e_integrals
becke_numerical_grid
bitmask
cis
@ -13,14 +14,13 @@ dressing
electrons
ezfio_files
fci
four_idx
generators_cas
generators_full
hartree_fock
integrals_bielec
mo_basis
mo_guess
mo_one_e_integrals
mo_two_e_integrals
mpi
mrpt_utils
nuclei

View File

@ -52,8 +52,7 @@ program fci
n_det_before = 0
double precision :: correlation_energy_ratio
double precision :: threshold_selectors_save, threshold_generators_save
threshold_selectors_save = threshold_selectors
double precision :: threshold_generators_save
threshold_generators_save = threshold_generators
double precision :: error(N_states)
@ -71,13 +70,11 @@ program fci
pt2 = 0.d0
variance = 0.d0
norm = 0.d0
threshold_selectors = 1.d0
threshold_generators = 1.d0
SOFT_TOUCH threshold_selectors threshold_generators
SOFT_TOUCH threshold_generators
call ZMQ_pt2(psi_energy_with_nucl_rep,pt2,relative_error,error, variance, norm) ! Stochastic PT2
threshold_selectors = threshold_selectors_save
threshold_generators = threshold_generators_save
SOFT_TOUCH threshold_selectors threshold_generators
SOFT_TOUCH threshold_generators
endif
@ -123,13 +120,11 @@ program fci
pt2 = 0.d0
variance = 0.d0
norm = 0.d0
threshold_selectors = 1.d0
threshold_generators = 1d0
SOFT_TOUCH threshold_selectors threshold_generators
SOFT_TOUCH threshold_generators
call ZMQ_pt2(psi_energy_with_nucl_rep, pt2,relative_error,error,variance,norm) ! Stochastic PT2
threshold_selectors = threshold_selectors_save
threshold_generators = threshold_generators_save
SOFT_TOUCH threshold_selectors threshold_generators
SOFT_TOUCH threshold_generators
call ezfio_set_fci_energy(psi_energy_with_nucl_rep(1:N_states))
call ezfio_set_fci_energy_pt2(psi_energy_with_nucl_rep(1:N_states)+pt2)
endif

View File

@ -5,7 +5,8 @@ program pt2
! EZFIO directory.
END_DOC
read_wf = .True.
SOFT_TOUCH read_wf
threshold_generators = 1.d0
SOFT_TOUCH read_wf threshold_generators
PROVIDE mo_bielec_integrals_in_map
PROVIDE psi_energy
call run
@ -26,8 +27,6 @@ subroutine run
pt2(:) = 0.d0
E_CI_before(:) = psi_energy(:) + nuclear_repulsion
threshold_selectors = 1.d0
threshold_generators = 1.d0
relative_error=PT2_relative_error
call ZMQ_pt2(psi_energy_with_nucl_rep,pt2,relative_error,error, variance, norm) ! Stochastic PT2

View File

@ -154,9 +154,6 @@ subroutine ZMQ_pt2(E, pt2,relative_error, error, variance, norm)
if (zmq_put_ivector(zmq_to_qp_run_socket,1,'pt2_stoch_istate',pt2_stoch_istate,1) == -1) then
stop 'Unable to put pt2_stoch_istate on ZMQ server'
endif
if (zmq_put_dvector(zmq_to_qp_run_socket,1,'threshold_selectors',threshold_selectors,1) == -1) then
stop 'Unable to put threshold_selectors on ZMQ server'
endif
if (zmq_put_dvector(zmq_to_qp_run_socket,1,'threshold_generators',threshold_generators,1) == -1) then
stop 'Unable to put threshold_generators on ZMQ server'
endif

View File

@ -1212,28 +1212,19 @@ subroutine get_d0(gen, phasemask, bannedOrb, banned, mat, mask, h, p, sp, coefs)
if(bannedOrb(p2,2)) cycle
call get_mo_bielec_integrals(p2,h1,h2,mo_tot_num,hij_cache1,mo_integrals_map)
do p1=1, mo_tot_num
if(bannedOrb(p1, 1)) cycle
if(banned(p1, p2, bant)) cycle ! rentable?
if(p1 == h1 .or. p2 == h2) then
call apply_particles(mask, 1,p1,2,p2, det, ok, N_int)
call i_h_j(gen, det, N_int, hij)
else
if(bannedOrb(p1, 1) .or. banned(p1, p2, bant)) cycle
if(p1 /= h1 .and. p2 /= h2) then
if (hij_cache1(p1) == 0.d0) cycle
phase = get_phase_bi(phasemask, 1, 2, h1, p1, h2, p2, N_int)
hij = hij_cache1(p1) * phase
else
call apply_particles(mask, 1,p1,2,p2, det, ok, N_int)
call i_h_j(gen, det, N_int, hij)
if (hij == 0.d0) cycle
end if
! if( (bannedOrb(p1, 1)).or.(banned(p1, p2, bant)) ) then
! hij = 0.d0
! else if(p1 /= h1 .and. p2 /= h2) then
! hij = hij_cache1(p1) * get_phase_bi(phasemask, 1, 2, h1, p1, h2, p2, N_int)
! else
! call apply_particles(mask, 1,p1,2,p2, det, ok, N_int)
! call i_h_j(gen, det, N_int, hij)
! end if
if (hij /= 0.d0) then
do k=1,N_states
mat(k, p1, p2) = mat(k, p1, p2) + coefs(k) * hij ! HOTSPOT
enddo
endif
do k=1,N_states
mat(k, p1, p2) = mat(k, p1, p2) + coefs(k) * hij ! HOTSPOT
enddo
end do
end do
@ -1245,20 +1236,19 @@ subroutine get_d0(gen, phasemask, bannedOrb, banned, mat, mask, h, p, sp, coefs)
call get_mo_bielec_integrals(puti,p2,p1,mo_tot_num,hij_cache1,mo_integrals_map)
call get_mo_bielec_integrals(puti,p1,p2,mo_tot_num,hij_cache2,mo_integrals_map)
do putj=puti+1, mo_tot_num
if(bannedOrb(putj, sp)) cycle
if(banned(puti, putj, bant)) cycle ! rentable?
if(puti == p1 .or. putj == p2 .or. puti == p2 .or. putj == p1) then
call apply_particles(mask, sp,puti,sp,putj, det, ok, N_int)
call i_h_j(gen, det, N_int, hij)
if(bannedOrb(putj, sp) .or. banned(putj, sp, bant)) cycle
if(puti /= p1 .and. putj /= p2 .and. puti /= p2 .and. putj /= p1) then
hij = hij_cache1(putj) - hij_cache2(putj)
if (hij /= 0.d0) then
hij = hij * get_phase_bi(phasemask, sp, sp, puti, p1 , putj, p2, N_int)
do k=1,N_states
mat(k, puti, putj) = mat(k, puti, putj) + coefs(k) * hij
enddo
endif
else
hij = hij_cache1(putj) - hij_cache2(putj)
call apply_particles(mask, sp,puti,sp,putj, det, ok, N_int)
call i_h_j(gen, det, N_int, hij)
if (hij /= 0.d0) then
hij = hij * get_phase_bi(phasemask, sp, sp, puti, p1 , putj, p2, N_int)
do k=1,N_states
mat(k, puti, putj) = mat(k, puti, putj) + coefs(k) * hij
enddo
@ -1385,3 +1375,4 @@ subroutine bitstring_to_list_in_selection( string, list, n_elements, Nint)
enddo
end

View File

@ -42,9 +42,6 @@ subroutine ZMQ_selection(N_in, pt2, variance, norm)
if (zmq_put_dvector(zmq_to_qp_run_socket,1,'energy',pt2_e0_denominator,size(pt2_e0_denominator)) == -1) then
stop 'Unable to put energy on ZMQ server'
endif
if (zmq_put_dvector(zmq_to_qp_run_socket,1,'threshold_selectors',threshold_selectors,1) == -1) then
stop 'Unable to put threshold_selectors on ZMQ server'
endif
if (zmq_put_dvector(zmq_to_qp_run_socket,1,'state_average_weight',state_average_weight,N_states) == -1) then
stop 'Unable to put state_average_weight on ZMQ server'
endif

View File

@ -1 +0,0 @@
integrals_bielec

View File

@ -1,6 +0,0 @@
=======
FourIdx
=======
Four-index transformation.

View File

@ -1,4 +1,4 @@
integrals_bielec
ao_one_e_integrals
ao_two_e_integrals
mo_guess
bitmask

View File

@ -1,22 +0,0 @@
BEGIN_PROVIDER [double precision, big_array_coulomb_integrals, (mo_tot_num,mo_tot_num, mo_tot_num)]
&BEGIN_PROVIDER [double precision, big_array_exchange_integrals,(mo_tot_num,mo_tot_num, mo_tot_num)]
implicit none
integer :: i,j,k,l
double precision :: get_mo_bielec_integral
double precision :: integral
do k = 1, mo_tot_num
do i = 1, mo_tot_num
do j = 1, mo_tot_num
l = j
integral = get_mo_bielec_integral(i,j,k,l,mo_integrals_map)
big_array_coulomb_integrals(j,i,k) = integral
l = j
integral = get_mo_bielec_integral(i,j,l,k,mo_integrals_map)
big_array_exchange_integrals(j,i,k) = integral
enddo
enddo
enddo
END_PROVIDER

View File

@ -1,47 +0,0 @@
BEGIN_PROVIDER [ logical, read_ao_integrals ]
&BEGIN_PROVIDER [ logical, read_mo_integrals ]
&BEGIN_PROVIDER [ logical, write_ao_integrals ]
&BEGIN_PROVIDER [ logical, write_mo_integrals ]
BEGIN_DOC
! One level of abstraction for disk_access_ao_integrals and disk_access_mo_integrals
END_DOC
implicit none
if (disk_access_ao_integrals.EQ.'Read') then
read_ao_integrals = .True.
write_ao_integrals = .False.
else if (disk_access_ao_integrals.EQ.'Write') then
read_ao_integrals = .False.
write_ao_integrals = .True.
else if (disk_access_ao_integrals.EQ.'None') then
read_ao_integrals = .False.
write_ao_integrals = .False.
else
print *, 'bielec_integrals/disk_access_ao_integrals has a wrong type'
stop 1
endif
if (disk_access_mo_integrals.EQ.'Read') then
read_mo_integrals = .True.
write_mo_integrals = .False.
else if (disk_access_mo_integrals.EQ.'Write') then
read_mo_integrals = .False.
write_mo_integrals = .True.
else if (disk_access_mo_integrals.EQ.'None') then
read_mo_integrals = .False.
write_mo_integrals = .False.
else
print *, 'bielec_integrals/disk_access_mo_integrals has a wrong type'
stop 1
endif
END_PROVIDER

View File

@ -4,19 +4,12 @@ doc: Read/Write |MO| integrals from/to disk [ Write | Read | None ]
interface: ezfio,provider,ocaml
default: None
[disk_access_ao_integrals]
[disk_access_mo_integrals_erf]
type: Disk_access
doc: Read/Write |AO| integrals from/to disk [ Write | Read | None ]
doc: Read/Write MO integrals with the long range interaction from/to disk [ Write | Read | None ]
interface: ezfio,provider,ocaml
default: None
[ao_integrals_threshold]
type: Threshold
doc: If | (pq|rs) | < `ao_integrals_threshold` then (pq|rs) is zero
interface: ezfio,provider,ocaml
default: 1.e-15
ezfio_name: threshold_ao
[mo_integrals_threshold]
type: Threshold
doc: If | <ij|kl> | < `mo_integrals_threshold` then <ij|kl> is zero
@ -24,7 +17,6 @@ interface: ezfio,provider,ocaml
default: 1.e-15
ezfio_name: threshold_mo
[no_vvvv_integrals]
type: logical
doc: If `True`, computes all integrals except for the integrals having 4 virtual indices
@ -46,10 +38,3 @@ interface: ezfio,provider,ocaml
default: False
ezfio_name: no_vvv_integrals
[do_direct_integrals]
type: logical
doc: Compute integrals on the fly (very slow, only for debugging)
interface: ezfio,provider,ocaml
default: False
ezfio_name: direct

View File

@ -0,0 +1,6 @@
ao_one_e_integrals
ao_two_e_integrals
pseudo
bitmask
zmq
mo_basis

View File

@ -0,0 +1,46 @@
BEGIN_PROVIDER [double precision, big_array_coulomb_integrals, (mo_tot_num,mo_tot_num, mo_tot_num)]
&BEGIN_PROVIDER [double precision, big_array_exchange_integrals,(mo_tot_num,mo_tot_num, mo_tot_num)]
implicit none
integer :: i,j,k,l
double precision :: get_mo_bielec_integral
double precision :: integral
do k = 1, mo_tot_num
do i = 1, mo_tot_num
do j = 1, mo_tot_num
l = j
integral = get_mo_bielec_integral(i,j,k,l,mo_integrals_map)
big_array_coulomb_integrals(j,i,k) = integral
l = j
integral = get_mo_bielec_integral(i,j,l,k,mo_integrals_map)
big_array_exchange_integrals(j,i,k) = integral
enddo
enddo
enddo
END_PROVIDER
BEGIN_PROVIDER [double precision, big_array_coulomb_integrals_erf, (mo_tot_num,mo_tot_num, mo_tot_num)]
&BEGIN_PROVIDER [double precision, big_array_exchange_integrals_erf,(mo_tot_num,mo_tot_num, mo_tot_num)]
implicit none
integer :: i,j,k,l
double precision :: get_mo_bielec_integral_erf
double precision :: integral
do k = 1, mo_tot_num
do i = 1, mo_tot_num
do j = 1, mo_tot_num
l = j
integral = get_mo_bielec_integral_erf(i,j,k,l,mo_integrals_erf_map)
big_array_coulomb_integrals_erf(j,i,k) = integral
l = j
integral = get_mo_bielec_integral_erf(i,j,l,k,mo_integrals_erf_map)
big_array_exchange_integrals_erf(j,i,k) = integral
enddo
enddo
enddo
END_PROVIDER

View File

@ -1,285 +1,5 @@
use map_module
!! AO Map
!! ======
BEGIN_PROVIDER [ type(map_type), ao_integrals_map ]
implicit none
BEGIN_DOC
! AO integrals
END_DOC
integer(key_kind) :: key_max
integer(map_size_kind) :: sze
call bielec_integrals_index(ao_num,ao_num,ao_num,ao_num,key_max)
sze = key_max
call map_init(ao_integrals_map,sze)
print*, 'AO map initialized : ', sze
END_PROVIDER
subroutine bielec_integrals_index(i,j,k,l,i1)
use map_module
implicit none
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
subroutine bielec_integrals_index_reverse(i,j,k,l,i1)
use map_module
implicit none
integer, intent(out) :: i(8),j(8),k(8),l(8)
integer(key_kind), intent(in) :: i1
integer(key_kind) :: i2,i3
i = 0
i2 = ceiling(0.5d0*(dsqrt(8.d0*dble(i1)+1.d0)-1.d0))
l(1) = ceiling(0.5d0*(dsqrt(8.d0*dble(i2)+1.d0)-1.d0))
i3 = i1 - shiftr(i2*i2-i2,1)
k(1) = ceiling(0.5d0*(dsqrt(8.d0*dble(i3)+1.d0)-1.d0))
j(1) = int(i2 - shiftr(l(1)*l(1)-l(1),1),4)
i(1) = int(i3 - shiftr(k(1)*k(1)-k(1),1),4)
!ijkl
i(2) = i(1) !ilkj
j(2) = l(1)
k(2) = k(1)
l(2) = j(1)
i(3) = k(1) !kjil
j(3) = j(1)
k(3) = i(1)
l(3) = l(1)
i(4) = k(1) !klij
j(4) = l(1)
k(4) = i(1)
l(4) = j(1)
i(5) = j(1) !jilk
j(5) = i(1)
k(5) = l(1)
l(5) = k(1)
i(6) = j(1) !jkli
j(6) = k(1)
k(6) = l(1)
l(6) = i(1)
i(7) = l(1) !lijk
j(7) = i(1)
k(7) = j(1)
l(7) = k(1)
i(8) = l(1) !lkji
j(8) = k(1)
k(8) = j(1)
l(8) = i(1)
integer :: ii, jj
do ii=2,8
do jj=1,ii-1
if ( (i(ii) == i(jj)).and. &
(j(ii) == j(jj)).and. &
(k(ii) == k(jj)).and. &
(l(ii) == l(jj)) ) then
i(ii) = 0
exit
endif
enddo
enddo
do ii=1,8
if (i(ii) /= 0) then
call bielec_integrals_index(i(ii),j(ii),k(ii),l(ii),i2)
if (i1 /= i2) then
print *, i1, i2
print *, i(ii), j(ii), k(ii), l(ii)
stop 'bielec_integrals_index_reverse failed'
endif
endif
enddo
end
BEGIN_PROVIDER [ integer, ao_integrals_cache_min ]
&BEGIN_PROVIDER [ integer, ao_integrals_cache_max ]
implicit none
BEGIN_DOC
! Min and max values of the AOs for which the integrals are in the cache
END_DOC
ao_integrals_cache_min = max(1,ao_num - 63)
ao_integrals_cache_max = ao_num
END_PROVIDER
BEGIN_PROVIDER [ double precision, ao_integrals_cache, (0:64*64*64*64) ]
implicit none
BEGIN_DOC
! Cache of AO integrals for fast access
END_DOC
PROVIDE ao_bielec_integrals_in_map
integer :: i,j,k,l,ii
integer(key_kind) :: idx
real(integral_kind) :: integral
!$OMP PARALLEL DO PRIVATE (i,j,k,l,idx,ii,integral)
do l=ao_integrals_cache_min,ao_integrals_cache_max
do k=ao_integrals_cache_min,ao_integrals_cache_max
do j=ao_integrals_cache_min,ao_integrals_cache_max
do i=ao_integrals_cache_min,ao_integrals_cache_max
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,idx)
!DIR$ FORCEINLINE
call map_get(ao_integrals_map,idx,integral)
ii = l-ao_integrals_cache_min
ii = ior( shiftl(ii,6), k-ao_integrals_cache_min)
ii = ior( shiftl(ii,6), j-ao_integrals_cache_min)
ii = ior( shiftl(ii,6), i-ao_integrals_cache_min)
ao_integrals_cache(ii) = integral
enddo
enddo
enddo
enddo
!$OMP END PARALLEL DO
END_PROVIDER
double precision function get_ao_bielec_integral(i,j,k,l,map) result(result)
use map_module
implicit none
BEGIN_DOC
! Gets one AO bi-electronic integral from the AO map
END_DOC
integer, intent(in) :: i,j,k,l
integer(key_kind) :: idx
type(map_type), intent(inout) :: map
integer :: ii
real(integral_kind) :: tmp
PROVIDE ao_bielec_integrals_in_map ao_integrals_cache ao_integrals_cache_min
!DIR$ FORCEINLINE
if (ao_overlap_abs(i,k)*ao_overlap_abs(j,l) < ao_integrals_threshold ) then
tmp = 0.d0
else if (ao_bielec_integral_schwartz(i,k)*ao_bielec_integral_schwartz(j,l) < ao_integrals_threshold) then
tmp = 0.d0
else
ii = l-ao_integrals_cache_min
ii = ior(ii, k-ao_integrals_cache_min)
ii = ior(ii, j-ao_integrals_cache_min)
ii = ior(ii, i-ao_integrals_cache_min)
if (iand(ii, -64) /= 0) then
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,idx)
!DIR$ FORCEINLINE
call map_get(map,idx,tmp)
else
ii = l-ao_integrals_cache_min
ii = ior( shiftl(ii,6), k-ao_integrals_cache_min)
ii = ior( shiftl(ii,6), j-ao_integrals_cache_min)
ii = ior( shiftl(ii,6), i-ao_integrals_cache_min)
tmp = ao_integrals_cache(ii)
endif
endif
result = tmp
end
subroutine get_ao_bielec_integrals(j,k,l,sze,out_val)
use map_module
BEGIN_DOC
! Gets multiple AO bi-electronic integral from the AO map .
! All i are retrieved for j,k,l fixed.
END_DOC
implicit none
integer, intent(in) :: j,k,l, sze
real(integral_kind), intent(out) :: out_val(sze)
integer :: i
integer(key_kind) :: hash
double precision :: thresh
PROVIDE ao_bielec_integrals_in_map ao_integrals_map
thresh = ao_integrals_threshold
if (ao_overlap_abs(j,l) < thresh) then
out_val = 0.d0
return
endif
double precision :: get_ao_bielec_integral
do i=1,sze
out_val(i) = get_ao_bielec_integral(i,j,k,l,ao_integrals_map)
enddo
end
subroutine get_ao_bielec_integrals_non_zero(j,k,l,sze,out_val,out_val_index,non_zero_int)
use map_module
implicit none
BEGIN_DOC
! Gets multiple AO bi-electronic integral from the AO map .
! All non-zero i are retrieved for j,k,l fixed.
END_DOC
integer, intent(in) :: j,k,l, sze
real(integral_kind), intent(out) :: out_val(sze)
integer, intent(out) :: out_val_index(sze),non_zero_int
integer :: i
integer(key_kind) :: hash
double precision :: thresh,tmp
PROVIDE ao_bielec_integrals_in_map
thresh = ao_integrals_threshold
non_zero_int = 0
if (ao_overlap_abs(j,l) < thresh) then
out_val = 0.d0
return
endif
non_zero_int = 0
do i=1,sze
integer, external :: ao_l4
double precision, external :: ao_bielec_integral
!DIR$ FORCEINLINE
if (ao_bielec_integral_schwartz(i,k)*ao_bielec_integral_schwartz(j,l) < thresh) then
cycle
endif
call bielec_integrals_index(i,j,k,l,hash)
call map_get(ao_integrals_map, hash,tmp)
if (dabs(tmp) < thresh ) cycle
non_zero_int = non_zero_int+1
out_val_index(non_zero_int) = i
out_val(non_zero_int) = tmp
enddo
end
function get_ao_map_size()
implicit none
integer (map_size_kind) :: get_ao_map_size
BEGIN_DOC
! Returns the number of elements in the AO map
END_DOC
get_ao_map_size = ao_integrals_map % n_elements
end
subroutine clear_ao_map
implicit none
BEGIN_DOC
! Frees the memory of the AO map
END_DOC
call map_deinit(ao_integrals_map)
FREE ao_integrals_map
end
!! MO Map
!! ======
@ -296,20 +16,6 @@ BEGIN_PROVIDER [ type(map_type), mo_integrals_map ]
print*, 'MO map initialized: ', sze
END_PROVIDER
subroutine insert_into_ao_integrals_map(n_integrals,buffer_i, buffer_values)
use map_module
implicit none
BEGIN_DOC
! Create new entry into AO map
END_DOC
integer, intent(in) :: n_integrals
integer(key_kind), intent(inout) :: buffer_i(n_integrals)
real(integral_kind), intent(inout) :: buffer_values(n_integrals)
call map_append(ao_integrals_map, buffer_i, buffer_values, n_integrals)
end
subroutine insert_into_mo_integrals_map(n_integrals, &
buffer_i, buffer_values, thr)
use map_module
@ -663,13 +369,12 @@ integer*8 function get_mo_map_size()
get_mo_map_size = mo_integrals_map % n_elements
end
BEGIN_TEMPLATE
subroutine dump_$ao_integrals(filename)
subroutine dump_mo_integrals(filename)
use map_module
implicit none
BEGIN_DOC
! Save to disk the $ao integrals
! Save to disk the |MO| integrals
END_DOC
character*(*), intent(in) :: filename
integer(cache_key_kind), pointer :: key(:)
@ -681,16 +386,16 @@ subroutine dump_$ao_integrals(filename)
call ezfio_set_work_empty(.False.)
open(unit=66,file=filename,FORM='unformatted')
write(66) integral_kind, key_kind
write(66) $ao_integrals_map%sorted, $ao_integrals_map%map_size, &
$ao_integrals_map%n_elements
do i=0_8,$ao_integrals_map%map_size
write(66) $ao_integrals_map%map(i)%sorted, $ao_integrals_map%map(i)%map_size,&
$ao_integrals_map%map(i)%n_elements
write(66) mo_integrals_map%sorted, mo_integrals_map%map_size, &
mo_integrals_map%n_elements
do i=0_8,mo_integrals_map%map_size
write(66) mo_integrals_map%map(i)%sorted, mo_integrals_map%map(i)%map_size,&
mo_integrals_map%map(i)%n_elements
enddo
do i=0_8,$ao_integrals_map%map_size
key => $ao_integrals_map%map(i)%key
val => $ao_integrals_map%map(i)%value
n = $ao_integrals_map%map(i)%n_elements
do i=0_8,mo_integrals_map%map_size
key => mo_integrals_map%map(i)%key
val => mo_integrals_map%map(i)%value
n = mo_integrals_map%map(i)%n_elements
write(66) (key(j), j=1,n), (val(j), j=1,n)
enddo
close(66)
@ -698,10 +403,10 @@ subroutine dump_$ao_integrals(filename)
end
integer function load_$ao_integrals(filename)
integer function load_mo_integrals(filename)
implicit none
BEGIN_DOC
! Read from disk the $ao integrals
! Read from disk the |MO| integrals
END_DOC
character*(*), intent(in) :: filename
integer*8 :: i
@ -709,7 +414,7 @@ integer function load_$ao_integrals(filename)
real(integral_kind), pointer :: val(:)
integer :: iknd, kknd
integer*8 :: n, j
load_$ao_integrals = 1
load_mo_integrals = 1
open(unit=66,file=filename,FORM='unformatted',STATUS='UNKNOWN')
read(66,err=98,end=98) iknd, kknd
if (iknd /= integral_kind) then
@ -720,30 +425,26 @@ integer function load_$ao_integrals(filename)
print *, 'Wrong key kind in file :', kknd
stop 1
endif
read(66,err=98,end=98) $ao_integrals_map%sorted, $ao_integrals_map%map_size,&
$ao_integrals_map%n_elements
do i=0_8, $ao_integrals_map%map_size
read(66,err=99,end=99) $ao_integrals_map%map(i)%sorted, &
$ao_integrals_map%map(i)%map_size, $ao_integrals_map%map(i)%n_elements
call cache_map_reallocate($ao_integrals_map%map(i),$ao_integrals_map%map(i)%map_size)
read(66,err=98,end=98) mo_integrals_map%sorted, mo_integrals_map%map_size,&
mo_integrals_map%n_elements
do i=0_8, mo_integrals_map%map_size
read(66,err=99,end=99) mo_integrals_map%map(i)%sorted, &
mo_integrals_map%map(i)%map_size, mo_integrals_map%map(i)%n_elements
call cache_map_reallocate(mo_integrals_map%map(i),mo_integrals_map%map(i)%map_size)
enddo
do i=0_8, $ao_integrals_map%map_size
key => $ao_integrals_map%map(i)%key
val => $ao_integrals_map%map(i)%value
n = $ao_integrals_map%map(i)%n_elements
do i=0_8, mo_integrals_map%map_size
key => mo_integrals_map%map(i)%key
val => mo_integrals_map%map(i)%value
n = mo_integrals_map%map(i)%n_elements
read(66,err=99,end=99) (key(j), j=1,n), (val(j), j=1,n)
enddo
call map_sort($ao_integrals_map)
load_$ao_integrals = 0
call map_sort(mo_integrals_map)
load_mo_integrals = 0
return
99 continue
call map_deinit($ao_integrals_map)
call map_deinit(mo_integrals_map)
98 continue
stop 'Problem reading $ao_integrals_map file in work/'
stop 'Problem reading mo_integrals_map file in work/'
end
SUBST [ ao_integrals_map, ao_integrals, ao_num ]
ao_integrals_map ; ao_integrals ; ao_num ;;
mo_integrals_map ; mo_integrals ; mo_tot_num ;;
END_TEMPLATE

View File

@ -0,0 +1,386 @@
use map_module
integer function load_mo_integrals_erf(filename)
implicit none
BEGIN_DOC
! Read from disk the $ao integrals
END_DOC
character*(*), intent(in) :: filename
integer*8 :: i
integer(cache_key_kind), pointer :: key(:)
real(integral_kind), pointer :: val(:)
integer :: iknd, kknd
integer*8 :: n, j
load_mo_integrals_erf = 1
open(unit=66,file=filename,FORM='unformatted',STATUS='UNKNOWN')
read(66,err=98,end=98) iknd, kknd
if (iknd /= integral_kind) then
print *, 'Wrong integrals kind in file :', iknd
stop 1
endif
if (kknd /= key_kind) then
print *, 'Wrong key kind in file :', kknd
stop 1
endif
read(66,err=98,end=98) mo_integrals_erf_map%sorted, mo_integrals_erf_map%map_size,&
mo_integrals_erf_map%n_elements
do i=0_8, mo_integrals_erf_map%map_size
read(66,err=99,end=99) mo_integrals_erf_map%map(i)%sorted, &
mo_integrals_erf_map%map(i)%map_size, mo_integrals_erf_map%map(i)%n_elements
call cache_map_reallocate(mo_integrals_erf_map%map(i),mo_integrals_erf_map%map(i)%map_size)
enddo
do i=0_8, mo_integrals_erf_map%map_size
key => mo_integrals_erf_map%map(i)%key
val => mo_integrals_erf_map%map(i)%value
n = mo_integrals_erf_map%map(i)%n_elements
read(66,err=99,end=99) (key(j), j=1,n), (val(j), j=1,n)
enddo
call map_sort(mo_integrals_erf_map)
load_mo_integrals_erf = 0
return
99 continue
call map_deinit(mo_integrals_erf_map)
98 continue
stop 'Problem reading mo_integrals_erf_map file in work/'
end
BEGIN_PROVIDER [ type(map_type), mo_integrals_erf_map ]
implicit none
BEGIN_DOC
! MO integrals
END_DOC
integer(key_kind) :: key_max
integer(map_size_kind) :: sze
call bielec_integrals_index(mo_tot_num,mo_tot_num,mo_tot_num,mo_tot_num,key_max)
sze = key_max
call map_init(mo_integrals_erf_map,sze)
print*, 'MO ERF map initialized'
END_PROVIDER
subroutine insert_into_mo_integrals_erf_map(n_integrals, &
buffer_i, buffer_values, thr)
use map_module
implicit none
BEGIN_DOC
! Create new entry into MO map, or accumulate in an existing entry
END_DOC
integer, intent(in) :: n_integrals
integer(key_kind), intent(inout) :: buffer_i(n_integrals)
real(integral_kind), intent(inout) :: buffer_values(n_integrals)
real(integral_kind), intent(in) :: thr
call map_update(mo_integrals_erf_map, buffer_i, buffer_values, n_integrals, thr)
end
BEGIN_PROVIDER [ integer, mo_integrals_erf_cache_min ]
&BEGIN_PROVIDER [ integer, mo_integrals_erf_cache_max ]
implicit none
BEGIN_DOC
! Min and max values of the MOs for which the integrals are in the cache
END_DOC
mo_integrals_erf_cache_min = max(1,elec_alpha_num - 31)
mo_integrals_erf_cache_max = min(mo_tot_num,mo_integrals_erf_cache_min+63)
END_PROVIDER
BEGIN_PROVIDER [ double precision, mo_integrals_erf_cache, (0:64*64*64*64) ]
implicit none
BEGIN_DOC
! Cache of MO integrals for fast access
END_DOC
PROVIDE mo_bielec_integrals_erf_in_map
integer :: i,j,k,l
integer :: ii
integer(key_kind) :: idx
real(integral_kind) :: integral
FREE ao_integrals_erf_cache
!$OMP PARALLEL DO PRIVATE (i,j,k,l,idx,ii,integral)
do l=mo_integrals_erf_cache_min,mo_integrals_erf_cache_max
do k=mo_integrals_erf_cache_min,mo_integrals_erf_cache_max
do j=mo_integrals_erf_cache_min,mo_integrals_erf_cache_max
do i=mo_integrals_erf_cache_min,mo_integrals_erf_cache_max
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,idx)
!DIR$ FORCEINLINE
call map_get(mo_integrals_erf_map,idx,integral)
ii = l-mo_integrals_erf_cache_min
ii = ior( ishft(ii,6), k-mo_integrals_erf_cache_min)
ii = ior( ishft(ii,6), j-mo_integrals_erf_cache_min)
ii = ior( ishft(ii,6), i-mo_integrals_erf_cache_min)
mo_integrals_erf_cache(ii) = integral
enddo
enddo
enddo
enddo
!$OMP END PARALLEL DO
END_PROVIDER
double precision function get_mo_bielec_integral_erf(i,j,k,l,map)
use map_module
implicit none
BEGIN_DOC
! Returns one integral <ij|kl> in the MO basis
END_DOC
integer, intent(in) :: i,j,k,l
integer(key_kind) :: idx
integer :: ii
type(map_type), intent(inout) :: map
real(integral_kind) :: tmp
PROVIDE mo_bielec_integrals_erf_in_map mo_integrals_erf_cache
ii = l-mo_integrals_erf_cache_min
ii = ior(ii, k-mo_integrals_erf_cache_min)
ii = ior(ii, j-mo_integrals_erf_cache_min)
ii = ior(ii, i-mo_integrals_erf_cache_min)
if (iand(ii, -64) /= 0) then
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,idx)
!DIR$ FORCEINLINE
call map_get(map,idx,tmp)
get_mo_bielec_integral_erf = dble(tmp)
else
ii = l-mo_integrals_erf_cache_min
ii = ior( ishft(ii,6), k-mo_integrals_erf_cache_min)
ii = ior( ishft(ii,6), j-mo_integrals_erf_cache_min)
ii = ior( ishft(ii,6), i-mo_integrals_erf_cache_min)
get_mo_bielec_integral_erf = mo_integrals_erf_cache(ii)
endif
end
double precision function mo_bielec_integral_erf(i,j,k,l)
implicit none
BEGIN_DOC
! Returns one integral <ij|kl> in the MO basis
END_DOC
integer, intent(in) :: i,j,k,l
double precision :: get_mo_bielec_integral_erf
PROVIDE mo_bielec_integrals_erf_in_map mo_integrals_erf_cache
!DIR$ FORCEINLINE
PROVIDE mo_bielec_integrals_erf_in_map
mo_bielec_integral_erf = get_mo_bielec_integral_erf(i,j,k,l,mo_integrals_erf_map)
return
end
subroutine get_mo_bielec_integrals_erf(j,k,l,sze,out_val,map)
use map_module
implicit none
BEGIN_DOC
! Returns multiple integrals <ij|kl> in the MO basis, all
! i for j,k,l fixed.
END_DOC
integer, intent(in) :: j,k,l, sze
double precision, intent(out) :: out_val(sze)
type(map_type), intent(inout) :: map
integer :: i
integer(key_kind) :: hash(sze)
real(integral_kind) :: tmp_val(sze)
PROVIDE mo_bielec_integrals_erf_in_map
do i=1,sze
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,hash(i))
enddo
if (key_kind == 8) then
call map_get_many(map, hash, out_val, sze)
else
call map_get_many(map, hash, tmp_val, sze)
! Conversion to double precision
do i=1,sze
out_val(i) = dble(tmp_val(i))
enddo
endif
end
subroutine get_mo_bielec_integrals_erf_ij(k,l,sze,out_array,map)
use map_module
implicit none
BEGIN_DOC
! Returns multiple integrals <ij|kl> in the MO basis, all
! i(1)j(2) 1/r12 k(1)l(2)
! i, j for k,l fixed.
END_DOC
integer, intent(in) :: k,l, sze
double precision, intent(out) :: out_array(sze,sze)
type(map_type), intent(inout) :: map
integer :: i,j,kk,ll,m
integer(key_kind),allocatable :: hash(:)
integer ,allocatable :: pairs(:,:), iorder(:)
real(integral_kind), allocatable :: tmp_val(:)
PROVIDE mo_bielec_integrals_erf_in_map
allocate (hash(sze*sze), pairs(2,sze*sze),iorder(sze*sze), &
tmp_val(sze*sze))
kk=0
out_array = 0.d0
do j=1,sze
do i=1,sze
kk += 1
!DIR$ FORCEINLINE
call bielec_integrals_index(i,j,k,l,hash(kk))
pairs(1,kk) = i
pairs(2,kk) = j
iorder(kk) = kk
enddo
enddo
logical :: integral_is_in_map
if (key_kind == 8) then
call i8radix_sort(hash,iorder,kk,-1)
else if (key_kind == 4) then
call iradix_sort(hash,iorder,kk,-1)
else if (key_kind == 2) then
call i2radix_sort(hash,iorder,kk,-1)
endif
call map_get_many(mo_integrals_erf_map, hash, tmp_val, kk)
do ll=1,kk
m = iorder(ll)
i=pairs(1,m)
j=pairs(2,m)
out_array(i,j) = tmp_val(ll)
enddo
deallocate(pairs,hash,iorder,tmp_val)
end
subroutine get_mo_bielec_integrals_erf_i1j1(k,l,sze,out_array,map)
use map_module
implicit none
BEGIN_DOC
! Returns multiple integrals <ik|jl> in the MO basis, all
! i(1)j(1) erf(mu_erf * r12) /r12 k(2)l(2)
! i, j for k,l fixed.
END_DOC
integer, intent(in) :: k,l, sze
double precision, intent(out) :: out_array(sze,sze)
type(map_type), intent(inout) :: map
integer :: i,j,kk,ll,m
integer(key_kind),allocatable :: hash(:)
integer ,allocatable :: pairs(:,:), iorder(:)
real(integral_kind), allocatable :: tmp_val(:)
PROVIDE mo_bielec_integrals_erf_in_map
allocate (hash(sze*sze), pairs(2,sze*sze),iorder(sze*sze), &
tmp_val(sze*sze))
kk=0
out_array = 0.d0
do j=1,sze
do i=1,sze
kk += 1
!DIR$ FORCEINLINE
call bielec_integrals_index(i,k,j,l,hash(kk))
pairs(1,kk) = i
pairs(2,kk) = j
iorder(kk) = kk
enddo
enddo
logical :: integral_is_in_map
if (key_kind == 8) then
call i8radix_sort(hash,iorder,kk,-1)
else if (key_kind == 4) then
call iradix_sort(hash,iorder,kk,-1)
else if (key_kind == 2) then
call i2radix_sort(hash,iorder,kk,-1)
endif
call map_get_many(mo_integrals_erf_map, hash, tmp_val, kk)
do ll=1,kk
m = iorder(ll)
i=pairs(1,m)
j=pairs(2,m)
out_array(i,j) = tmp_val(ll)
enddo
deallocate(pairs,hash,iorder,tmp_val)
end
subroutine get_mo_bielec_integrals_erf_coulomb_ii(k,l,sze,out_val,map)
use map_module
implicit none
BEGIN_DOC
! Returns multiple integrals <ki|li>
! k(1)i(2) 1/r12 l(1)i(2) :: out_val(i1)
! for k,l fixed.
END_DOC
integer, intent(in) :: k,l, sze
double precision, intent(out) :: out_val(sze)
type(map_type), intent(inout) :: map
integer :: i
integer(key_kind) :: hash(sze)
real(integral_kind) :: tmp_val(sze)
PROVIDE mo_bielec_integrals_erf_in_map
integer :: kk
do i=1,sze
!DIR$ FORCEINLINE
call bielec_integrals_index(k,i,l,i,hash(i))
enddo
if (key_kind == 8) then
call map_get_many(map, hash, out_val, sze)
else
call map_get_many(map, hash, tmp_val, sze)
! Conversion to double precision
do i=1,sze
out_val(i) = dble(tmp_val(i))
enddo
endif
end
subroutine get_mo_bielec_integrals_erf_exch_ii(k,l,sze,out_val,map)
use map_module
implicit none
BEGIN_DOC
! Returns multiple integrals <ki|il>
! k(1)i(2) 1/r12 i(1)l(2) :: out_val(i1)
! for k,l fixed.
END_DOC
integer, intent(in) :: k,l, sze
double precision, intent(out) :: out_val(sze)
type(map_type), intent(inout) :: map
integer :: i
integer(key_kind) :: hash(sze)
real(integral_kind) :: tmp_val(sze)
PROVIDE mo_bielec_integrals_erf_in_map
integer :: kk
do i=1,sze
!DIR$ FORCEINLINE
call bielec_integrals_index(k,i,i,l,hash(i))
enddo
if (key_kind == 8) then
call map_get_many(map, hash, out_val, sze)
else
call map_get_many(map, hash, tmp_val, sze)
! Conversion to double precision
do i=1,sze
out_val(i) = dble(tmp_val(i))
enddo
endif
end
integer*8 function get_mo_erf_map_size()
implicit none
BEGIN_DOC
! Return the number of elements in the MO map
END_DOC
get_mo_erf_map_size = mo_integrals_erf_map % n_elements
end

View File

@ -152,7 +152,7 @@ BEGIN_PROVIDER [ logical, mo_bielec_integrals_in_map ]
if (write_mo_integrals.and.mpi_master) then
call ezfio_set_work_empty(.False.)
call map_save_to_disk(trim(ezfio_filename)//'/work/mo_ints',mo_integrals_map)
call ezfio_set_integrals_bielec_disk_access_mo_integrals("Read")
call ezfio_set_mo_two_e_integrals_disk_access_mo_integrals_erf('Read')
endif
END_PROVIDER

View File

@ -0,0 +1,549 @@
subroutine mo_bielec_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+ishft(r*r-r,-1)
q = min(j,l)
s = max(j,l)
q = q+ishft(s*s-s,-1)
i1 = min(p,q)
i2 = max(p,q)
i1 = i1+ishft(i2*i2-i2,-1)
end
BEGIN_PROVIDER [ logical, mo_bielec_integrals_erf_in_map ]
use map_module
implicit none
integer(bit_kind) :: mask_ijkl(N_int,4)
integer(bit_kind) :: mask_ijk(N_int,3)
BEGIN_DOC
! If True, the map of MO bielectronic integrals is provided
END_DOC
real :: map_mb
mo_bielec_integrals_erf_in_map = .True.
if (read_mo_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
else
PROVIDE ao_bielec_integrals_erf_in_map
endif
! call four_index_transform_block(ao_integrals_erf_map,mo_integrals_erf_map, &
! mo_coef, size(mo_coef,1), &
! 1, 1, 1, 1, ao_num, ao_num, ao_num, ao_num, &
! 1, 1, 1, 1, mo_num, mo_num, mo_num, mo_num)
call add_integrals_to_map_erf(full_ijkl_bitmask_4)
integer*8 :: get_mo_erf_map_size, mo_erf_map_size
mo_erf_map_size = get_mo_erf_map_size()
! print*,'Molecular integrals ERF provided:'
! print*,' Size of MO ERF map ', map_mb(mo_integrals_erf_map) ,'MB'
! print*,' Number of MO ERF integrals: ', mo_erf_map_size
if (write_mo_integrals_erf) 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_integrals_bielec_erf_disk_access_mo_integrals_erf("Read")
endif
END_PROVIDER
BEGIN_PROVIDER [ double precision, mo_bielec_integral_erf_jj_from_ao, (mo_tot_num,mo_tot_num) ]
&BEGIN_PROVIDER [ double precision, mo_bielec_integral_erf_jj_exchange_from_ao, (mo_tot_num,mo_tot_num) ]
&BEGIN_PROVIDER [ double precision, mo_bielec_integral_erf_jj_anti_from_ao, (mo_tot_num,mo_tot_num) ]
BEGIN_DOC
! mo_bielec_integral_jj_from_ao(i,j) = J_ij
! mo_bielec_integral_jj_exchange_from_ao(i,j) = J_ij
! mo_bielec_integral_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_bielec_integrals_erf_in_map mo_coef
endif
mo_bielec_integral_erf_jj_from_ao = 0.d0
mo_bielec_integral_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) &
!$OMP SHARED(mo_tot_num,mo_coef_transp,ao_num,&
!$OMP ao_integrals_threshold,do_direct_integrals) &
!$OMP REDUCTION(+:mo_bielec_integral_erf_jj_from_ao,mo_bielec_integral_erf_jj_exchange_from_ao)
allocate( int_value(ao_num), int_idx(ao_num), &
iqrs(mo_tot_num,ao_num), iqis(mo_tot_num), iqri(mo_tot_num),&
iqsr(mo_tot_num,ao_num) )
!$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_tot_num
iqrs(i,j) = 0.d0
iqsr(i,j) = 0.d0
enddo
enddo
if (do_direct_integrals) then
double precision :: ao_bielec_integral_erf
do r=1,ao_num
call compute_ao_bielec_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_tot_num
iqrs(i,r) += mo_coef_transp(i,p) * integral
enddo
endif
enddo
call compute_ao_bielec_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_tot_num
iqsr(i,r) += mo_coef_transp(i,p) * integral
enddo
endif
enddo
enddo
else
do r=1,ao_num
call get_ao_bielec_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_tot_num
iqrs(i,r) += mo_coef_transp(i,p) * integral
enddo
endif
enddo
call get_ao_bielec_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_tot_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_tot_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_tot_num
!DIR$ VECTOR ALIGNED
do j=1,mo_tot_num
c = mo_coef_transp(j,q)*mo_coef_transp(j,s)
mo_bielec_integral_erf_jj_from_ao(j,i) += c * iqis(i)
mo_bielec_integral_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_bielec_integral_erf_jj_anti_from_ao = mo_bielec_integral_erf_jj_from_ao - mo_bielec_integral_erf_jj_exchange_from_ao
! end
END_PROVIDER
BEGIN_PROVIDER [ double precision, mo_bielec_integral_erf_jj, (mo_tot_num,mo_tot_num) ]
&BEGIN_PROVIDER [ double precision, mo_bielec_integral_erf_jj_exchange, (mo_tot_num,mo_tot_num) ]
&BEGIN_PROVIDER [ double precision, mo_bielec_integral_erf_jj_anti, (mo_tot_num,mo_tot_num) ]
implicit none
BEGIN_DOC
! mo_bielec_integral_jj(i,j) = J_ij
! mo_bielec_integral_jj_exchange(i,j) = K_ij
! mo_bielec_integral_jj_anti(i,j) = J_ij - K_ij
END_DOC
integer :: i,j
double precision :: get_mo_bielec_integral_erf
PROVIDE mo_bielec_integrals_erf_in_map
mo_bielec_integral_erf_jj = 0.d0
mo_bielec_integral_erf_jj_exchange = 0.d0
do j=1,mo_tot_num
do i=1,mo_tot_num
mo_bielec_integral_erf_jj(i,j) = get_mo_bielec_integral_erf(i,j,i,j,mo_integrals_erf_map)
mo_bielec_integral_erf_jj_exchange(i,j) = get_mo_bielec_integral_erf(i,j,j,i,mo_integrals_erf_map)
mo_bielec_integral_erf_jj_anti(i,j) = mo_bielec_integral_erf_jj(i,j) - mo_bielec_integral_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_bielec_integral_erf_jj mo_bielec_integral_erf_jj_anti
FREE mo_bielec_integral_Erf_jj_exchange mo_bielec_integrals_erf_in_map
end
subroutine provide_all_mo_integrals_erf
implicit none
provide mo_integrals_erf_map mo_bielec_integral_erf_jj mo_bielec_integral_erf_jj_anti
provide mo_bielec_integral_erf_jj_exchange mo_bielec_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 :: bielec_tmp_0_idx(:)
real(integral_kind), allocatable :: bielec_tmp_0(:,:)
double precision, allocatable :: bielec_tmp_1(:)
double precision, allocatable :: bielec_tmp_2(:,:)
double precision, allocatable :: bielec_tmp_3(:,:,:)
!DIR$ ATTRIBUTES ALIGN : 64 :: bielec_tmp_1, bielec_tmp_2, bielec_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_bielec_integrals_in_map mo_coef
!Get list of MOs for i,j,k and l
!-------------------------------
allocate(list_ijkl(mo_tot_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_tot_num*(n_j)*(n_k+1) + mo_tot_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 bielec_tmp_0_idx, bielec_tmp_0, bielec_tmp_1,bielec_tmp_2,bielec_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_tot_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(bielec_tmp_3(mo_tot_num, n_j, n_k), &
bielec_tmp_1(mo_tot_num), &
bielec_tmp_0(ao_num,ao_num), &
bielec_tmp_0_idx(ao_num), &
bielec_tmp_2(mo_tot_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
bielec_tmp_3 = 0.d0
do k1 = 1,ao_num
bielec_tmp_2 = 0.d0
do j1 = 1,ao_num
call get_ao_bielec_integrals_erf(j1,k1,l1,ao_num,bielec_tmp_0(1,j1)) ! all integrals for a given l1, k1
! call compute_ao_bielec_integrals(j1,k1,l1,ao_num,bielec_tmp_0(1,j1))
enddo
do j1 = 1,ao_num
kmax = 0
do i1 = 1,ao_num
c = bielec_tmp_0(i1,j1)
if (c == 0.d0) then
cycle
endif
kmax += 1
bielec_tmp_0(kmax,j1) = c
bielec_tmp_0_idx(kmax) = i1
enddo
if (kmax==0) then
cycle
endif
bielec_tmp_1 = 0.d0
ii1=1
! sum_m c_m^i (m)
do ii1 = 1,kmax-4,4
i1 = bielec_tmp_0_idx(ii1)
i2 = bielec_tmp_0_idx(ii1+1)
i3 = bielec_tmp_0_idx(ii1+2)
i4 = bielec_tmp_0_idx(ii1+3)
do i = list_ijkl(1,1), list_ijkl(n_i,1)
bielec_tmp_1(i) = bielec_tmp_1(i) + &
mo_coef_transp(i,i1) * bielec_tmp_0(ii1,j1) + &
mo_coef_transp(i,i2) * bielec_tmp_0(ii1+1,j1) + &
mo_coef_transp(i,i3) * bielec_tmp_0(ii1+2,j1) + &
mo_coef_transp(i,i4) * bielec_tmp_0(ii1+3,j1)
enddo ! i
enddo ! ii1
i2 = ii1
do ii1 = i2,kmax
i1 = bielec_tmp_0_idx(ii1)
do i = list_ijkl(1,1), list_ijkl(n_i,1)
bielec_tmp_1(i) = bielec_tmp_1(i) + mo_coef_transp(i,i1) * bielec_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(bielec_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)
bielec_tmp_2(i,j0) = bielec_tmp_2(i,j0) + c * bielec_tmp_1(i)
enddo ! i
enddo ! j
enddo !j1
if ( maxval(abs(bielec_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
bielec_tmp_3(i,j0,k0) = bielec_tmp_3(i,j0,k0) + c* bielec_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
bielec_tmp_1 = 0.d0
do i0 = 1, n_i
i = list_ijkl(i0,1)
if (i>k) then
exit
endif
bielec_tmp_1(i) = c*bielec_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(bielec_tmp_1(i)) < mo_integrals_threshold) then
cycle
endif
n_integrals += 1
buffer_value(n_integrals) = bielec_tmp_1(i)
!DIR$ FORCEINLINE
call mo_bielec_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
!$OMP END DO NOWAIT
deallocate (bielec_tmp_1,bielec_tmp_2,bielec_tmp_3)
integer :: index_needed
call insert_into_mo_integrals_erf_map(n_integrals,buffer_i,buffer_value,&
real(mo_integrals_threshold,integral_kind))
deallocate(buffer_i, buffer_value)
!$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

View File

@ -0,0 +1,55 @@
BEGIN_PROVIDER [ logical, read_mo_integrals ]
&BEGIN_PROVIDER [ logical, write_mo_integrals ]
BEGIN_DOC
! Flag to read or write the |MO| integrals
END_DOC
implicit none
if (disk_access_mo_integrals.EQ.'Read') then
read_mo_integrals = .True.
write_mo_integrals = .False.
else if (disk_access_mo_integrals.EQ.'Write') then
read_mo_integrals = .False.
write_mo_integrals = .True.
else if (disk_access_mo_integrals.EQ.'None') then
read_mo_integrals = .False.
write_mo_integrals = .False.
else
print *, 'disk_access_mo_integrals has a wrong type'
stop 1
endif
END_PROVIDER
BEGIN_PROVIDER [ logical, read_mo_integrals_erf ]
&BEGIN_PROVIDER [ logical, write_mo_integrals_erf ]
implicit none
BEGIN_DOC
! Flag to read or write the |MO| erf integrals
END_DOC
if (disk_access_mo_integrals_erf.EQ.'Read') then
read_mo_integrals_erf = .True.
write_mo_integrals_erf = .False.
else if (disk_access_mo_integrals_erf.EQ.'Write') then
read_mo_integrals_erf = .False.
write_mo_integrals_erf = .True.
else if (disk_access_mo_integrals_erf.EQ.'None') then
read_mo_integrals_erf = .False.
write_mo_integrals_erf = .False.
else
print *, 'disk_access_mo_integrals_erf has a wrong type'
stop 1
endif
END_PROVIDER

View File

@ -0,0 +1,18 @@
subroutine save_erf_bi_elec_integrals_mo
implicit none
integer :: i,j,k,l
PROVIDE mo_bielec_integrals_erf_in_map
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_integrals_bielec_disk_access_mo_integrals("Read")
end
subroutine save_erf_bielec_ints_mo_into_ints_mo
implicit none
integer :: i,j,k,l
PROVIDE mo_bielec_integrals_erf_in_map
call ezfio_set_work_empty(.False.)
call map_save_to_disk(trim(ezfio_filename)//'/work/mo_ints',mo_integrals_erf_map)
call ezfio_set_integrals_bielec_disk_access_mo_integrals("Read")
end

View File

@ -169,7 +169,6 @@ subroutine pt2_epstein_nesbet_2x2 ($arguments)
endif
H_pert_diag(i) = h*c_pert(i)*c_pert(i)
! print*, 'N_det,N_det_selectors = ',N_det,N_det_selectors
! print*, 'threshold_selectors',threshold_selectors
! print*, delta_e,i_H_psi_array(1)
! double precision :: hij,accu
! accu = 0.d0

View File

@ -1,5 +1,13 @@
use bitmasks
BEGIN_PROVIDER [ double precision, threshold_selectors ]
implicit none
BEGIN_DOC
! Thresholds on selectors (fraction of the square of the norm)
END_DOC
threshold_selectors = dsqrt(threshold_generators)
END_PROVIDER
BEGIN_PROVIDER [ integer, N_det_selectors]
implicit none
BEGIN_DOC
@ -10,9 +18,6 @@ BEGIN_PROVIDER [ integer, N_det_selectors]
double precision :: norm, norm_max
call write_time(6)
N_det_selectors = N_det
! if (threshold_selectors == 1.d0) then
! return
! endif
norm = 1.d0
do i=1,N_det
norm = norm - psi_average_norm_contrib_sorted(i)

View File

@ -16,7 +16,7 @@ end
subroutine provide_everything
PROVIDE H_apply_buffer_allocated mo_bielec_integrals_in_map psi_det_generators psi_coef_generators psi_det_sorted_bit psi_selectors n_det_generators n_states generators_bitmask zmq_context N_states_diag
PROVIDE pt2_e0_denominator mo_tot_num N_int ci_energy mpi_master zmq_state zmq_context
PROVIDE psi_det psi_coef threshold_generators threshold_selectors state_average_weight
PROVIDE psi_det psi_coef threshold_generators state_average_weight
PROVIDE N_det_selectors pt2_stoch_istate N_det
end
@ -51,7 +51,7 @@ subroutine run_wf
zmq_to_qp_run_socket = new_zmq_to_qp_run_socket()
PROVIDE psi_det psi_coef threshold_generators threshold_selectors state_average_weight mpi_master
PROVIDE psi_det psi_coef threshold_generators state_average_weight mpi_master
PROVIDE zmq_state N_det_selectors pt2_stoch_istate N_det pt2_e0_denominator
PROVIDE N_det_generators N_states N_states_diag psi_energy
@ -101,10 +101,6 @@ subroutine run_wf
call mpi_print('zmq_get_dvector threshold_generators')
IRP_ENDIF
if (zmq_get_dvector(zmq_to_qp_run_socket,1,'threshold_generators',threshold_generators,1) == -1) cycle
IRP_IF MPI_DEBUG
call mpi_print('zmq_get_dvector threshold_selectors')
IRP_ENDIF
if (zmq_get_dvector(zmq_to_qp_run_socket,1,'threshold_selectors',threshold_selectors,1) == -1) cycle
IRP_IF MPI_DEBUG
call mpi_print('zmq_get_dvector energy')
IRP_ENDIF
@ -122,7 +118,7 @@ subroutine run_wf
IRP_ENDIF
if (zmq_get_dvector(zmq_to_qp_run_socket,1,'state_average_weight',state_average_weight,N_states) == -1) cycle
psi_energy(1:N_states) = energy(1:N_states)
TOUCH psi_energy state_average_weight threshold_selectors threshold_generators
TOUCH psi_energy state_average_weight threshold_generators
if (mpi_master) then
print *, 'N_det', N_det
@ -213,10 +209,6 @@ subroutine run_wf
call mpi_print('zmq_get_dvector threshold_generators')
IRP_ENDIF
if (zmq_get_dvector(zmq_to_qp_run_socket,1,'threshold_generators',threshold_generators,1) == -1) cycle
IRP_IF MPI_DEBUG
call mpi_print('zmq_get_dvector threshold_selectors')
IRP_ENDIF
if (zmq_get_dvector(zmq_to_qp_run_socket,1,'threshold_selectors',threshold_selectors,1) == -1) cycle
IRP_IF MPI_DEBUG
call mpi_print('zmq_get_dvector energy')
IRP_ENDIF
@ -230,7 +222,7 @@ subroutine run_wf
IRP_ENDIF
if (zmq_get_dvector(zmq_to_qp_run_socket,1,'state_average_weight',state_average_weight,N_states) == -1) cycle
psi_energy(1:N_states) = energy(1:N_states)
TOUCH psi_energy state_average_weight pt2_stoch_istate threshold_selectors threshold_generators
TOUCH psi_energy state_average_weight pt2_stoch_istate threshold_generators
if (mpi_master) then
print *, 'N_det', N_det
print *, 'N_det_generators', N_det_generators

View File

@ -0,0 +1,20 @@
program write_integrals
implicit none
BEGIN_DOC
! Saves the bielec erf integrals into the EZFIO
END_DOC
disk_access_mo_integrals = 'None'
touch disk_access_mo_integrals
disk_access_ao_integrals = 'None'
touch disk_access_ao_integrals
call routine
end
subroutine routine
implicit none
call save_erf_bi_elec_integrals_ao
call save_erf_bi_elec_integrals_mo
end

View File

@ -28,7 +28,7 @@ function run_FCI() {
}
@test "FCI H2O cc-pVDZ" {
run_FCI h2o.ezfio 2000 -76.1250464119599 -76.1258876912360
run_FCI h2o.ezfio 2000 -76.1249493055975 -76.1258779921351
}