10
0
mirror of https://github.com/LCPQ/quantum_package synced 2024-12-23 04:43:50 +01:00
quantum_package/plugins/mrcc_selected/dressing_slave.irp.f

602 lines
19 KiB
Fortran

subroutine mrsc2_dressing_slave_tcp(i)
implicit none
integer, intent(in) :: i
BEGIN_DOC
! Task for parallel MR-SC2
END_DOC
call mrsc2_dressing_slave(0,i)
end
subroutine mrsc2_dressing_slave_inproc(i)
implicit none
integer, intent(in) :: i
BEGIN_DOC
! Task for parallel MR-SC2
END_DOC
call mrsc2_dressing_slave(1,i)
end
subroutine mrsc2_dressing_slave(thread,iproc)
use f77_zmq
implicit none
BEGIN_DOC
! Task for parallel MR-SC2
END_DOC
integer, intent(in) :: thread, iproc
! integer :: j,l
integer :: rc
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
double precision, allocatable :: delta(:,:,:), delta_s2(:,:,:)
integer :: i_state, i, i_I, J, k, k2, k1, kk, ll, degree, degree2, m, l, deg, ni, m2
integer :: n(2)
integer :: p1,p2,h1,h2,s1,s2, blok, I_s, J_s, kn
logical :: ok
double precision :: phase_iI, phase_Ik, phase_Jl, phase_Ji, phase_al
double precision :: diI, hIi, hJi, delta_JI, dkI, HkI, ci_inv(N_states), cj_inv(N_states)
double precision :: contrib, contrib_s2, wall, iwall
double precision, allocatable :: dleat(:,:,:), dleat_s2(:,:,:)
integer, dimension(0:2,2,2) :: exc_iI, exc_Ik, exc_IJ
integer(bit_kind) :: det_tmp(N_int, 2), det_tmp2(N_int, 2), inac, virt
integer, external :: get_index_in_psi_det_sorted_bit, searchDet, detCmp
logical, external :: is_in_wavefunction, isInCassd, detEq
integer,allocatable :: komon(:)
logical :: komoned
!double precision, external :: get_dij
zmq_to_qp_run_socket = new_zmq_to_qp_run_socket()
zmq_socket_push = new_zmq_push_socket(thread)
call connect_to_taskserver(zmq_to_qp_run_socket,worker_id,thread)
allocate (dleat(N_states, N_det_non_ref, 2), delta(N_states,0:N_det_non_ref, 2))
allocate (dleat_s2(N_states, N_det_non_ref, 2), delta_s2(N_states,0:N_det_non_ref, 2))
allocate(komon(0:N_det_non_ref))
do
call get_task_from_taskserver(zmq_to_qp_run_socket,worker_id, task_id, task)
if (task_id == 0) exit
read (task,*) i_I, J, k1, k2
do i_state=1, N_states
ci_inv(i_state) = 1.d0 / psi_ref_coef(i_I,i_state)
cj_inv(i_state) = 1.d0 / psi_ref_coef(J,i_state)
end do
n = 0
delta(:,0,:) = 0d0
delta(:,:nlink(J),1) = 0d0
delta(:,:nlink(i_I),2) = 0d0
delta_s2(:,0,:) = 0d0
delta_s2(:,:nlink(J),1) = 0d0
delta_s2(:,:nlink(i_I),2) = 0d0
komon(0) = 0
komoned = .false.
do kk = k1, k2
k = det_cepa0_idx(linked(kk, i_I))
blok = blokMwen(kk, i_I)
call get_excitation(psi_ref(1,1,i_I),psi_non_ref(1,1,k),exc_Ik,degree,phase_Ik,N_int)
if(J /= i_I) then
call apply_excitation(psi_ref(1,1,J),exc_Ik,det_tmp2,ok,N_int)
if(.not. ok) cycle
l = searchDet(det_cepa0(1,1,cepa0_shortcut(blok)), det_tmp2, cepa0_shortcut(blok+1)-cepa0_shortcut(blok), N_int)
if(l == -1) cycle
ll = cepa0_shortcut(blok)-1+l
l = det_cepa0_idx(ll)
ll = child_num(ll, J)
else
l = k
ll = kk
end if
if(.not. komoned) then
m = 0
m2 = 0
do while(m < nlink(i_I) .and. m2 < nlink(J))
m += 1
m2 += 1
if(linked(m, i_I) < linked(m2, J)) then
m2 -= 1
cycle
else if(linked(m, i_I) > linked(m2, J)) then
m -= 1
cycle
end if
i = det_cepa0_idx(linked(m, i_I))
if(h_cache(J,i) == 0.d0) cycle
if(h_cache(i_I,i) == 0.d0) cycle
komon(0) += 1
kn = komon(0)
komon(kn) = i
do i_state = 1,N_states
dkI = h_cache(J,i) * dij(i_I, i, i_state)
dleat(i_state, kn, 1) = dkI
dleat(i_state, kn, 2) = dkI
dkI = s2_cache(J,i) * dij(i_I, i, i_state)
dleat_s2(i_state, kn, 1) = dkI
dleat_s2(i_state, kn, 2) = dkI
end do
end do
komoned = .true.
end if
integer :: hpmin(2)
hpmin(1) = 2 - HP(1,k)
hpmin(2) = 2 - HP(2,k)
do m = 1, komon(0)
i = komon(m)
if(HP(1,i) <= hpmin(1) .and. HP(2,i) <= hpmin(2) ) then
cycle
end if
call apply_excitation(psi_non_ref(1,1,i),exc_Ik,det_tmp,ok,N_int)
if(.not. ok) cycle
do i_state = 1, N_states
contrib = dij(i_I, k, i_state) * dleat(i_state, m, 2)
contrib_s2 = dij(i_I, k, i_state) * dleat_s2(i_state, m, 2)
delta(i_state,ll,1) += contrib
delta_s2(i_state,ll,1) += contrib_s2
if(dabs(psi_ref_coef(i_I,i_state)).ge.5.d-5) then
delta(i_state,0,1) -= contrib * ci_inv(i_state) * psi_non_ref_coef(l,i_state)
delta_s2(i_state,0,1) -= contrib_s2 * ci_inv(i_state) * psi_non_ref_coef(l,i_state)
endif
if(I_i == J) cycle
contrib = dij(J, l, i_state) * dleat(i_state, m, 1)
contrib_s2 = dij(J, l, i_state) * dleat_s2(i_state, m, 1)
delta(i_state,kk,2) += contrib
delta_s2(i_state,kk,2) += contrib_s2
if(dabs(psi_ref_coef(J,i_state)).ge.5.d-5) then
delta(i_state,0,2) -= contrib * cj_inv(i_state) * psi_non_ref_coef(k,i_state)
delta_s2(i_state,0,2) -= contrib_s2 * cj_inv(i_state) * psi_non_ref_coef(k,i_state)
end if
enddo !i_state
end do ! while
end do ! kk
call push_mrsc2_results(zmq_socket_push, I_i, J, delta, delta_s2, task_id)
call task_done_to_taskserver(zmq_to_qp_run_socket,worker_id,task_id)
! end if
enddo
deallocate(delta)
call disconnect_from_taskserver(zmq_to_qp_run_socket,zmq_socket_push,worker_id)
call end_zmq_to_qp_run_socket(zmq_to_qp_run_socket)
call end_zmq_push_socket(zmq_socket_push,thread)
end
subroutine push_mrsc2_results(zmq_socket_push, I_i, J, delta, delta_s2, task_id)
use f77_zmq
implicit none
BEGIN_DOC
! Push integrals in the push socket
END_DOC
integer, intent(in) :: i_I, J
integer(ZMQ_PTR), intent(in) :: zmq_socket_push
double precision,intent(inout) :: delta(N_states, 0:N_det_non_ref, 2)
double precision,intent(inout) :: delta_s2(N_states, 0:N_det_non_ref, 2)
integer, intent(in) :: task_id
integer :: rc , i_state, i, kk, li
integer,allocatable :: idx(:,:)
integer :: n(2)
logical :: ok
allocate(idx(N_det_non_ref,2))
rc = f77_zmq_send( zmq_socket_push, i_I, 4, ZMQ_SNDMORE)
if (rc /= 4) then
print *, irp_here, 'f77_zmq_send( zmq_socket_push, i_I, 4, ZMQ_SNDMORE)'
stop 'error'
endif
rc = f77_zmq_send( zmq_socket_push, J, 4, ZMQ_SNDMORE)
if (rc /= 4) then
print *, irp_here, 'f77_zmq_send( zmq_socket_push, J, 4, ZMQ_SNDMORE)'
stop 'error'
endif
do kk=1,2
n(kk)=0
if(kk == 1) li = nlink(j)
if(kk == 2) li = nlink(i_I)
do i=1, li
ok = .false.
do i_state=1,N_states
if(delta(i_state, i, kk) /= 0d0) then
ok = .true.
exit
end if
end do
if(ok) then
n(kk) += 1
! idx(n,kk) = i
if(kk == 1) then
idx(n(1),1) = det_cepa0_idx(linked(i, J))
else
idx(n(2),2) = det_cepa0_idx(linked(i, i_I))
end if
do i_state=1, N_states
delta(i_state, n(kk), kk) = delta(i_state, i, kk)
end do
end if
end do
rc = f77_zmq_send( zmq_socket_push, n(kk), 4, ZMQ_SNDMORE)
if (rc /= 4) then
print *, irp_here, 'f77_zmq_send( zmq_socket_push, n, 4, ZMQ_SNDMORE)'
stop 'error'
endif
if(n(kk) /= 0) then
rc = f77_zmq_send( zmq_socket_push, delta(1,0,kk), (n(kk)+1)*8*N_states, ZMQ_SNDMORE) ! delta(1,0,1) = delta_I delta(1,0,2) = delta_J
if (rc /= (n(kk)+1)*8*N_states) then
print *, irp_here, 'f77_zmq_send( zmq_socket_push, delta, (n(kk)+1)*8*N_states, ZMQ_SNDMORE)'
stop 'error'
endif
rc = f77_zmq_send( zmq_socket_push, delta_s2(1,0,kk), (n(kk)+1)*8*N_states, ZMQ_SNDMORE) ! delta_s2(1,0,1) = delta_I delta_s2(1,0,2) = delta_J
if (rc /= (n(kk)+1)*8*N_states) then
print *, irp_here, 'f77_zmq_send( zmq_socket_push, delta_s2, (n(kk)+1)*8*N_states, ZMQ_SNDMORE)'
stop 'error'
endif
rc = f77_zmq_send( zmq_socket_push, idx(1,kk), n(kk)*4, ZMQ_SNDMORE)
if (rc /= n(kk)*4) then
print *, irp_here, 'f77_zmq_send( zmq_socket_push, delta, 8*n(kk), ZMQ_SNDMORE)'
stop 'error'
endif
end if
end do
rc = f77_zmq_send( zmq_socket_push, task_id, 4, 0)
if (rc /= 4) then
print *, irp_here, 'f77_zmq_send( zmq_socket_push, task_id, 4, 0)'
stop 'error'
endif
! ! Activate is zmq_socket_push is a REQ
! integer :: idummy
! rc = f77_zmq_recv( zmq_socket_push, idummy, 4, 0)
! if (rc /= 4) then
! print *, irp_here, 'f77_zmq_send( zmq_socket_push, idummy, 4, 0)'
! stop 'error'
! endif
end
subroutine pull_mrsc2_results(zmq_socket_pull, I_i, J, n, idx, delta, delta_s2, task_id)
use f77_zmq
implicit none
BEGIN_DOC
! Push integrals in the push socket
END_DOC
integer(ZMQ_PTR), intent(in) :: zmq_socket_pull
integer, intent(out) :: i_I, J, n(2)
double precision, intent(inout) :: delta(N_states, 0:N_det_non_ref, 2)
double precision, intent(inout) :: delta_s2(N_states, 0:N_det_non_ref, 2)
integer, intent(out) :: task_id
integer :: rc , i, kk
integer,intent(inout) :: idx(N_det_non_ref,2)
logical :: ok
rc = f77_zmq_recv( zmq_socket_pull, i_I, 4, ZMQ_SNDMORE)
if (rc /= 4) then
print *, irp_here, 'f77_zmq_recv( zmq_socket_pull, i_I, 4, ZMQ_SNDMORE)'
stop 'error'
endif
rc = f77_zmq_recv( zmq_socket_pull, J, 4, ZMQ_SNDMORE)
if (rc /= 4) then
print *, irp_here, 'f77_zmq_recv( zmq_socket_pull, J, 4, ZMQ_SNDMORE)'
stop 'error'
endif
do kk = 1, 2
rc = f77_zmq_recv( zmq_socket_pull, n(kk), 4, ZMQ_SNDMORE)
if (rc /= 4) then
print *, irp_here, 'f77_zmq_recv( zmq_socket_pull, n, 4, ZMQ_SNDMORE)'
stop 'error'
endif
if(n(kk) /= 0) then
rc = f77_zmq_recv( zmq_socket_pull, delta(1,0,kk), (n(kk)+1)*8*N_states, ZMQ_SNDMORE)
if (rc /= (n(kk)+1)*8*N_states) then
print *, irp_here, 'f77_zmq_recv( zmq_socket_pull, delta, (n(kk)+1)*8*N_states, ZMQ_SNDMORE)'
stop 'error'
endif
rc = f77_zmq_recv( zmq_socket_pull, delta_s2(1,0,kk), (n(kk)+1)*8*N_states, ZMQ_SNDMORE)
if (rc /= (n(kk)+1)*8*N_states) then
print *, irp_here, 'f77_zmq_recv( zmq_socket_pull, delta_s2, (n(kk)+1)*8*N_states, ZMQ_SNDMORE)'
stop 'error'
endif
rc = f77_zmq_recv( zmq_socket_pull, idx(1,kk), n(kk)*4, ZMQ_SNDMORE)
if (rc /= n(kk)*4) then
print *, irp_here, 'f77_zmq_recv( zmq_socket_pull, idx(1,kk), n(kk)*4, ZMQ_SNDMORE)'
stop 'error'
endif
end if
end do
rc = f77_zmq_recv( zmq_socket_pull, task_id, 4, 0)
if (rc /= 4) then
print *, irp_here, 'f77_zmq_recv( zmq_socket_pull, task_id, 4, 0)'
stop 'error'
endif
! ! Activate is zmq_socket_pull is a REP
! integer :: idummy
! rc = f77_zmq_send( zmq_socket_pull, idummy, 4, 0)
! if (rc /= 4) then
! print *, irp_here, 'f77_zmq_send( zmq_socket_pull, idummy, 4, 0)'
! stop 'error'
! endif
end
subroutine mrsc2_dressing_collector(delta_ii_,delta_ij_,delta_ii_s2_,delta_ij_s2_)
use f77_zmq
implicit none
BEGIN_DOC
! Collects results from the AO integral calculation
END_DOC
double precision,intent(inout) :: delta_ij_(N_states,N_det_non_ref,N_det_ref)
double precision,intent(inout) :: delta_ii_(N_states,N_det_ref)
double precision,intent(inout) :: delta_ij_s2_(N_states,N_det_non_ref,N_det_ref)
double precision,intent(inout) :: delta_ii_s2_(N_states,N_det_ref)
! integer :: j,l
integer :: rc
double precision, allocatable :: delta(:,:,:), delta_s2(:,:,:)
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(ZMQ_PTR) :: zmq_socket_pull
integer*8 :: control, accu
integer :: task_id, more
integer :: I_i, J, l, i_state, n(2), kk
integer,allocatable :: idx(:,:)
delta_ii_(:,:) = 0d0
delta_ij_(:,:,:) = 0d0
delta_ii_s2_(:,:) = 0d0
delta_ij_s2_(:,:,:) = 0d0
zmq_to_qp_run_socket = new_zmq_to_qp_run_socket()
zmq_socket_pull = new_zmq_pull_socket()
allocate ( delta(N_states,0:N_det_non_ref,2), delta_s2(N_states,0:N_det_non_ref,2) )
allocate(idx(N_det_non_ref,2))
more = 1
do while (more == 1)
call pull_mrsc2_results(zmq_socket_pull, I_i, J, n, idx, delta, delta_s2, task_id)
do l=1, n(1)
do i_state=1,N_states
delta_ij_(i_state,idx(l,1),i_I) += delta(i_state,l,1)
delta_ij_s2_(i_state,idx(l,1),i_I) += delta_s2(i_state,l,1)
end do
end do
do l=1, n(2)
do i_state=1,N_states
delta_ij_(i_state,idx(l,2),J) += delta(i_state,l,2)
delta_ij_s2_(i_state,idx(l,2),J) += delta_s2(i_state,l,2)
end do
end do
if(n(1) /= 0) then
do i_state=1,N_states
delta_ii_(i_state,i_I) += delta(i_state,0,1)
delta_ii_s2_(i_state,i_I) += delta_s2(i_state,0,1)
end do
end if
if(n(2) /= 0) then
do i_state=1,N_states
delta_ii_(i_state,J) += delta(i_state,0,2)
delta_ii_s2_(i_state,J) += delta_s2(i_state,0,2)
end do
end if
if (task_id /= 0) then
call zmq_delete_task(zmq_to_qp_run_socket,zmq_socket_pull,task_id,more)
endif
enddo
deallocate( delta, delta_s2 )
call end_zmq_to_qp_run_socket(zmq_to_qp_run_socket)
call end_zmq_pull_socket(zmq_socket_pull)
end
BEGIN_PROVIDER [ double precision, delta_ij_old, (N_states,N_det_non_ref,N_det_ref) ]
&BEGIN_PROVIDER [ double precision, delta_ii_old, (N_states,N_det_ref) ]
&BEGIN_PROVIDER [ double precision, delta_ij_s2_old, (N_states,N_det_non_ref,N_det_ref) ]
&BEGIN_PROVIDER [ double precision, delta_ii_s2_old, (N_states,N_det_ref) ]
implicit none
integer :: i_state, i, i_I, J, k, kk, degree, degree2, m, l, deg, ni, m2
integer :: p1,p2,h1,h2,s1,s2, blok, I_s, J_s, nex, nzer, ntot
! integer, allocatable :: linked(:,:), blokMwen(:, :), nlink(:)
logical :: ok
double precision :: phase_iI, phase_Ik, phase_Jl, phase_Ji, phase_al, diI, hIi, hJi, delta_JI, dkI(N_states), HkI, ci_inv(N_states), dia_hla(N_states)
double precision :: contrib, wall, iwall ! , searchance(N_det_ref)
integer, dimension(0:2,2,2) :: exc_iI, exc_Ik, exc_IJ
integer(bit_kind) :: det_tmp(N_int, 2), det_tmp2(N_int, 2), inac, virt
integer, external :: get_index_in_psi_det_sorted_bit, searchDet, detCmp
logical, external :: is_in_wavefunction, isInCassd, detEq
character*(512) :: task
integer(ZMQ_PTR) :: zmq_to_qp_run_socket
integer :: KKsize = 1000000
call new_parallel_job(zmq_to_qp_run_socket,'mrsc2')
call wall_time(iwall)
! allocate(linked(N_det_non_ref, N_det_ref), blokMwen(N_det_non_ref, N_det_ref), nlink(N_det_ref))
! searchance = 0d0
! do J = 1, N_det_ref
! nlink(J) = 0
! do blok=1,cepa0_shortcut(0)
! do k=cepa0_shortcut(blok), cepa0_shortcut(blok+1)-1
! call get_excitation_degree(psi_ref(1,1,J),det_cepa0(1,1,k),degree,N_int)
! if(degree <= 2) then
! nlink(J) += 1
! linked(nlink(J),J) = k
! blokMwen(nlink(J),J) = blok
! searchance(J) += 1d0 + log(dfloat(cepa0_shortcut(blok+1) - cepa0_shortcut(blok)))
! end if
! end do
! end do
! end do
! stop
nzer = 0
ntot = 0
do nex = 3, 0, -1
print *, "los ",nex
do I_s = N_det_ref, 1, -1
! if(mod(I_s,1) == 0) then
! call wall_time(wall)
! wall = wall-iwall
! print *, I_s, "/", N_det_ref, wall * (dfloat(N_det_ref) / dfloat(I_s)), wall, wall * (dfloat(N_det_ref) / dfloat(I_s))-wall
! end if
do J_s = 1, I_s
call get_excitation_degree(psi_ref(1,1,J_s), psi_ref(1,1,I_s), degree, N_int)
if(degree /= nex) cycle
if(nex == 3) nzer = nzer + 1
ntot += 1
! if(degree > 3) then
! deg += 1
! cycle
! else if(degree == -10) then
! KKsize = 100000
! else
! KKsize = 1000000
! end if
if(searchance(I_s) < searchance(J_s)) then
i_I = I_s
J = J_s
else
i_I = J_s
J = I_s
end if
KKsize = nlink(1)
if(nex == 0) KKsize = int(float(nlink(1)) / float(nlink(i_I)) * (float(nlink(1)) / 64d0))
!if(KKsize == 0) stop "ZZEO"
do kk = 1 , nlink(i_I), KKsize
write(task,*) I_i, J, kk, int(min(kk+KKsize-1, nlink(i_I)))
call add_task_to_taskserver(zmq_to_qp_run_socket,task)
end do
! do kk = 1 , nlink(i_I)
! k = linked(kk,i_I)
! blok = blokMwen(kk,i_I)
! write(task,*) I_i, J, k, blok
! call add_task_to_taskserver(zmq_to_qp_run_socket,task)
!
! enddo !kk
enddo !J
enddo !I
end do ! nex
print *, "tasked"
! integer(ZMQ_PTR) ∷ collector_thread
! external ∷ ao_bielec_integrals_in_map_collector
! rc = pthread_create(collector_thread, mrsc2_dressing_collector)
print *, nzer, ntot, float(nzer) / float(ntot)
provide nproc
!$OMP PARALLEL DEFAULT(none) SHARED(delta_ii_old,delta_ij_old,delta_ii_s2_old,delta_ij_s2_old) PRIVATE(i) NUM_THREADS(nproc+1)
i = omp_get_thread_num()
if (i==0) then
call mrsc2_dressing_collector(delta_ii_old,delta_ij_old,delta_ii_s2_old,delta_ij_s2_old)
else
call mrsc2_dressing_slave_inproc(i)
endif
!$OMP END PARALLEL
! rc = pthread_join(collector_thread)
call end_parallel_job(zmq_to_qp_run_socket, 'mrsc2')
END_PROVIDER