Merge branch 'dev' of github.com:QuantumPackage/qp2 into dev

etc/qp.rc

@@ -93,7 +93,7 @@ function qp()
 
     "prompt")
       shift
-      python2 $QP_ROOT/scripts/hello.py
+      python3 $QP_ROOT/scripts/hello.py
       function _check_ezfio() {
         if [[ -d ${EZFIO_FILE} ]] ; then
           printf "\e[0;32m|${EZFIO_FILE}>\e[m"

src/cipsi/pt2_stoch_routines.irp.f

@@ -15,7 +15,7 @@ END_PROVIDER
   pt2_n_tasks_max = min(pt2_n_tasks_max,1+N_det_generators/10000)
   call write_int(6,pt2_n_tasks_max,'pt2_n_tasks_max')
 
-  pt2_F(:) = int(sqrt(float(pt2_n_tasks_max)))
+  pt2_F(:) = max(int(sqrt(float(pt2_n_tasks_max))),1)
   do i=1,pt2_n_0(1+pt2_N_teeth/4)
     pt2_F(i) = pt2_n_tasks_max*pt2_min_parallel_tasks
   enddo
@@ -26,7 +26,6 @@ END_PROVIDER
     pt2_F(i) = 1
   enddo
 
-
 END_PROVIDER
 
  BEGIN_PROVIDER [ integer, pt2_N_teeth ]

src/cipsi/run_pt2_slave.irp.f

@@ -103,7 +103,7 @@ subroutine run_pt2_slave_small(thread,iproc,energy)
     if (n_tasks == 0) exit
 
     do k=1,n_tasks
-      read (task(k),*) subset(k), i_generator(k), N
+      call sscanf_ddd(task(k), subset(k), i_generator(k), N)
     enddo
     if (b%N == 0) then
       ! Only first time
@@ -183,9 +183,9 @@ subroutine run_pt2_slave_large(thread,iproc,energy)
   type(selection_buffer) :: b
   logical :: done, buffer_ready
 
-  type(pt2_type) :: pt2_data(1)
+  type(pt2_type) :: pt2_data
   integer :: n_tasks, k, N
-  integer :: i_generator(1), subset
+  integer :: i_generator, subset
 
   integer :: bsize ! Size of selection buffers
   logical :: sending
@@ -220,7 +220,11 @@ subroutine run_pt2_slave_large(thread,iproc,energy)
     endif
     if (n_tasks == 0) exit
 
-    read (task,*) subset, i_generator(1), N
+    call sscanf_ddd(task, subset, i_generator, N)
+    if( pt2_F(i_generator) <= 0 .or. pt2_F(i_generator) > N_det ) then
+      print *,  irp_here
+      stop 'bug in selection'
+    endif
     if (b%N == 0) then
       ! Only first time
       bsize = min(N, (elec_alpha_num * (mo_num-elec_alpha_num))**2)
@@ -232,15 +236,10 @@ subroutine run_pt2_slave_large(thread,iproc,energy)
 
     double precision :: time0, time1
     call wall_time(time0)
-    call pt2_alloc(pt2_data(1),N_states)
+    call pt2_alloc(pt2_data,N_states)
     b%cur = 0
-!double precision :: time2
+    call select_connected(i_generator,energy,pt2_data,b,subset,pt2_F(i_generator))
-!call wall_time(time2)
-    call select_connected(i_generator(1),energy,pt2_data(1),b,subset,pt2_F(i_generator(1)))
-!call wall_time(time1)
-!print *,  i_generator(1), time1-time2, n_tasks, pt2_F(i_generator(1))
     call wall_time(time1)
-!print *,  '-->', i_generator(1), time1-time0, n_tasks
 
     integer, external :: tasks_done_to_taskserver
     if (tasks_done_to_taskserver(zmq_to_qp_run_socket,worker_id,task_id,n_tasks) == -1) then
@@ -255,14 +254,14 @@ subroutine run_pt2_slave_large(thread,iproc,energy)
     call omp_unset_lock(global_selection_buffer_lock)
     if ( iproc == 1 ) then
       call omp_set_lock(global_selection_buffer_lock)
-      call push_pt2_results_async_send(zmq_socket_push, i_generator, pt2_data, global_selection_buffer, task_id, n_tasks,sending)
+      call push_pt2_results_async_send(zmq_socket_push, i_generator, pt2_data, global_selection_buffer, task_id, 1,sending)
       global_selection_buffer%cur = 0
       call omp_unset_lock(global_selection_buffer_lock)
     else
-      call push_pt2_results_async_send(zmq_socket_push, i_generator, pt2_data, b, task_id, n_tasks,sending)
+      call push_pt2_results_async_send(zmq_socket_push, i_generator, pt2_data, b, task_id, 1,sending)
     endif
 
-    call pt2_dealloc(pt2_data(1))
+    call pt2_dealloc(pt2_data)
   end do
   call push_pt2_results_async_recv(zmq_socket_push,b%mini,sending)
 

src/cipsi/slave_cipsi.irp.f

@@ -122,7 +122,7 @@ subroutine run_slave_main
       IRP_ENDIF
       if (zmq_get_dvector(zmq_to_qp_run_socket,1,'selection_weight',selection_weight,N_states) == -1) cycle
       pt2_e0_denominator(1:N_states) = energy(1:N_states)
-      TOUCH pt2_e0_denominator state_average_weight threshold_generators selection_weight
+      TOUCH pt2_e0_denominator state_average_weight threshold_generators selection_weight psi_det psi_coef
 
       if (mpi_master) then
         print *,  'N_det', N_det
@@ -230,7 +230,8 @@ subroutine run_slave_main
       IRP_ENDIF
       if (zmq_get_dvector(zmq_to_qp_run_socket,1,'selection_weight',selection_weight,N_states) == -1) cycle
       pt2_e0_denominator(1:N_states) = energy(1:N_states)
-      SOFT_TOUCH pt2_e0_denominator state_average_weight pt2_stoch_istate threshold_generators selection_weight
+      SOFT_TOUCH pt2_e0_denominator state_average_weight pt2_stoch_istate threshold_generators selection_weight psi_det psi_coef N_det_generators N_det_selectors
+
 
       call wall_time(t1)
       call write_double(6,(t1-t0),'Broadcast time')
@@ -300,7 +301,13 @@ subroutine run_slave_main
             PROVIDE det_to_configuration
           endif
 
-          PROVIDE global_selection_buffer 
+          PROVIDE global_selection_buffer pt2_N_teeth pt2_F N_det_generators
+          PROVIDE psi_bilinear_matrix_columns_loc psi_det_alpha_unique psi_det_beta_unique
+          PROVIDE psi_bilinear_matrix_rows psi_det_sorted_order psi_bilinear_matrix_order
+          PROVIDE psi_bilinear_matrix_transp_rows_loc psi_bilinear_matrix_transp_columns
+          PROVIDE psi_bilinear_matrix_transp_order psi_selectors_coef_transp psi_det_sorted
+          PROVIDE psi_det_hii selection_weight pseudo_sym pt2_min_parallel_tasks
+
           if (mpi_master) then
             print *,  'Running PT2'
           endif

src/csf/tree_utils.h

@@ -22,7 +22,15 @@ struct bin_tree {
     int NBF;
 };
 
-#include <cblas.h>
+typedef enum {CblasRowMajor=101, CblasColMajor=102} CBLAS_LAYOUT;
+typedef enum {CblasNoTrans=111, CblasTrans=112, CblasConjTrans=113} CBLAS_TRANSPOSE;
+typedef CBLAS_LAYOUT CBLAS_ORDER;
+void cblas_dgemm(CBLAS_LAYOUT layout, CBLAS_TRANSPOSE TransA,
+                 CBLAS_TRANSPOSE TransB, const int M, const int N,
+                 const int K, const double alpha, const double *A,
+                 const int lda, const double *B, const int ldb,
+                 const double beta, double *C, const int ldc);
+
 
 #define MAX_SOMO 32
 

src/fci/fci.irp.f

@@ -46,7 +46,7 @@ program fci
     endif
 
   else
-    PROVIDE mo_two_e_integrals_in_map
+    PROVIDE mo_two_e_integrals_in_map pt2_min_parallel_tasks
 
     call run_slave_cipsi