[cig-commits] [commit] Hiro_latest: Change MPI communicator in solver_SRs (778087d)
cig_noreply at geodynamics.org
cig_noreply at geodynamics.org
Mon Nov 18 16:20:37 PST 2013
Repository : ssh://geoshell/calypso
On branch : Hiro_latest
Link : https://github.com/geodynamics/calypso/compare/93e9f8f974c7a247c8f02e54ec18de063f86c8fb...3c548304673360ddedd7d68c8095b3fb74a2b9ce
>---------------------------------------------------------------
commit 778087dfcaa363a75962672b9b0b7a291734a81c
Author: Hiroaki Matsui <h_kemono at mac.com>
Date: Fri Sep 27 21:34:08 2013 -0700
Change MPI communicator in solver_SRs
>---------------------------------------------------------------
778087dfcaa363a75962672b9b0b7a291734a81c
.../PARALLEL_src/COMM_src/hdf5_file_IO.F90 | 1 +
.../PARALLEL_src/COMM_src/m_merged_ucd_data.f90 | 5 ++-
.../PARALLEL_src/COMM_src/m_work_time.f90 | 1 +
.../COMM_src/nodal_vector_send_recv.f90 | 8 ++--
.../PARALLEL_src/COMM_src/solver_SR.f90 | 40 +++++++-----------
.../PARALLEL_src/COMM_src/solver_SR_3.f90 | 49 +++++++++-------------
.../PARALLEL_src/COMM_src/solver_SR_6.f90 | 25 ++++-------
.../PARALLEL_src/COMM_src/solver_SR_N.f90 | 30 ++++---------
.../PARALLEL_src/COMM_src/solver_SR_int.f90 | 21 +++-------
.../SPH_SHELL_src/sum_sph_rms_data.f90 | 1 +
10 files changed, 65 insertions(+), 116 deletions(-)
diff --git a/src/Fortran_libraries/PARALLEL_src/COMM_src/hdf5_file_IO.F90 b/src/Fortran_libraries/PARALLEL_src/COMM_src/hdf5_file_IO.F90
index 4d69b14..b6d4ad2 100644
--- a/src/Fortran_libraries/PARALLEL_src/COMM_src/hdf5_file_IO.F90
+++ b/src/Fortran_libraries/PARALLEL_src/COMM_src/hdf5_file_IO.F90
@@ -23,6 +23,7 @@
module hdf5_file_IO
!
use m_precision
+ use calypso_mpi
use m_constants
use m_ucd_data
use m_parallel_var_dof
diff --git a/src/Fortran_libraries/PARALLEL_src/COMM_src/m_merged_ucd_data.f90 b/src/Fortran_libraries/PARALLEL_src/COMM_src/m_merged_ucd_data.f90
index 9c35aba..2dbeeb8 100644
--- a/src/Fortran_libraries/PARALLEL_src/COMM_src/m_merged_ucd_data.f90
+++ b/src/Fortran_libraries/PARALLEL_src/COMM_src/m_merged_ucd_data.f90
@@ -17,6 +17,7 @@
!
use m_precision
use m_constants
+ use calypso_mpi
use m_parallel_var_dof
!
implicit none
@@ -197,10 +198,10 @@
!
call solver_send_recv_i(nnod_ucd_local, NEIBPETOT, NEIBPE, &
& STACK_IMPORT, NOD_IMPORT, STACK_EXPORT, NOD_EXPORT, &
- & inod_local_ucd, SOLVER_COMM, my_rank)
+ & inod_local_ucd)
call solver_send_recv_i(nnod_ucd_local, NEIBPETOT, NEIBPE, &
& STACK_IMPORT, NOD_IMPORT, STACK_EXPORT, NOD_EXPORT, &
- & ihome_pe_ucd, SOLVER_COMM, my_rank)
+ & ihome_pe_ucd)
!
end subroutine set_node_double_address
!
diff --git a/src/Fortran_libraries/PARALLEL_src/COMM_src/m_work_time.f90 b/src/Fortran_libraries/PARALLEL_src/COMM_src/m_work_time.f90
index a5d3bda..bf16d94 100644
--- a/src/Fortran_libraries/PARALLEL_src/COMM_src/m_work_time.f90
+++ b/src/Fortran_libraries/PARALLEL_src/COMM_src/m_work_time.f90
@@ -120,6 +120,7 @@
!
subroutine output_elapsed_times
!
+ use calypso_mpi
use m_parallel_var_dof
!
integer(kind = kint) :: i
diff --git a/src/Fortran_libraries/PARALLEL_src/COMM_src/nodal_vector_send_recv.f90 b/src/Fortran_libraries/PARALLEL_src/COMM_src/nodal_vector_send_recv.f90
index 82d3194..ebd5aed 100644
--- a/src/Fortran_libraries/PARALLEL_src/COMM_src/nodal_vector_send_recv.f90
+++ b/src/Fortran_libraries/PARALLEL_src/COMM_src/nodal_vector_send_recv.f90
@@ -60,8 +60,7 @@
START_TIME= MPI_WTIME()
call SOLVER_SEND_RECV(numnod, num_neib, id_neib, &
& istack_import, item_import, &
- & istack_export, item_export, &
- & x_vec(1), SOLVER_COMM, my_rank )
+ & istack_export, item_export, x_vec(1) )
END_TIME= MPI_WTIME()
COMMtime = COMMtime + END_TIME - START_TIME
!
@@ -94,8 +93,7 @@
START_TIME= MPI_WTIME()
call SOLVER_SEND_RECV_3(numnod, num_neib, id_neib, &
& istack_import, item_import, &
- & istack_export, item_export, &
- & x_vec(1), SOLVER_COMM, my_rank )
+ & istack_export, item_export, x_vec(1) )
END_TIME= MPI_WTIME()
COMMtime = COMMtime + END_TIME - START_TIME
!
@@ -134,7 +132,7 @@
call SOLVER_SEND_RECV_6(numnod, num_neib, id_neib, &
& istack_import, item_import, &
& istack_export, item_export, &
- & x_vec(1), SOLVER_COMM, my_rank )
+ & x_vec(1))
END_TIME= MPI_WTIME()
COMMtime = COMMtime + END_TIME - START_TIME
!
diff --git a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR.f90 b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR.f90
index 726a3a2..3fe0f69 100644
--- a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR.f90
+++ b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR.f90
@@ -4,19 +4,19 @@
!!@author coded by K.Nakajima (RIST)
!!@date coded by K.Nakajima (RIST) on jul. 1999 (ver 1.0)
!!@n modified by H. Matsui (U. of Chicago) on july 2007 (ver 1.1)
+!!@n modified by H. Matsui (UC Davis) on Sep. 2013 (ver 1.2)
!
!>@brief MPI SEND and RECEIVE routine for scalar fields
!! in overlapped partitioning
!!
!!@verbatim
!! subroutine SOLVER_SEND_RECV &
-!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT,&
-!! & STACK_EXPORT, NOD_EXPORT,&
-!! & X, SOLVER_COMM,my_rank)
+!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+!! & STACK_EXPORT, NOD_EXPORT, X)
!! subroutine SOLVER_SEND_RECVx3 &
-!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT,&
-!! & STACK_EXPORT, NOD_EXPORT,&
-!! & X1, X2, X3, SOLVER_COMM,my_rank)
+!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+!! & STACK_EXPORT, NOD_EXPORT, &
+!! & X1, X2, X3)
!!@endverbatim
!!
!!@n @param N Number of data points
@@ -36,14 +36,12 @@
!!@n @param X1(N) 1st scalar field data
!!@n @param X2(N) 2nd scalar field data
!!@n @param X3(N) 3rd scalar field data
-!!
-!!@n @param SOLVER_COMM MPI communicator
-!!@n @param my_rank own process rank
!
module solver_SR
!
use m_precision
use m_constants
+ use calypso_mpi
!
implicit none
!
@@ -56,11 +54,9 @@
!C*** SOLVER_SEND_RECV
!C
subroutine SOLVER_SEND_RECV &
- & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
- & STACK_EXPORT, NOD_EXPORT, &
- & X, SOLVER_COMM,my_rank)
+ & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+ & STACK_EXPORT, NOD_EXPORT, X)
- use calypso_mpi
use m_solver_SR
!
!> number of nodes
@@ -81,11 +77,6 @@
& :: NOD_EXPORT
!> communicated result vector
real (kind=kreal), dimension(N) , intent(inout):: X
-!> communicator for mpi
- integer , intent(in) ::SOLVER_COMM
-!> Own process
- integer , intent(in) :: my_rank
-!>
!
integer (kind = kint) :: neib, istart, inum, iend, ierr, k
!
@@ -104,7 +95,7 @@
istart= STACK_EXPORT(neib-1) + 1
inum = STACK_EXPORT(neib ) - STACK_EXPORT(neib-1)
call MPI_ISEND (WS(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, &
& req1(neib), ierr)
enddo
@@ -115,7 +106,7 @@
istart= STACK_IMPORT(neib-1) + 1
inum = STACK_IMPORT(neib ) - STACK_IMPORT(neib-1)
call MPI_IRECV (WR(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, &
& req2(neib), ierr)
enddo
@@ -138,9 +129,8 @@
subroutine SOLVER_SEND_RECVx3 &
& ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
& STACK_EXPORT, NOD_EXPORT, &
- & X1, X2, X3, SOLVER_COMM,my_rank)
+ & X1, X2, X3)
- use calypso_mpi
use m_solver_SR
!
integer(kind=kint ) , intent(in) :: N
@@ -155,8 +145,6 @@
real (kind=kreal), dimension(N) , intent(inout):: X1
real (kind=kreal), dimension(N) , intent(inout):: X2
real (kind=kreal), dimension(N) , intent(inout):: X3
- integer , intent(in) ::SOLVER_COMM
- integer , intent(in) :: my_rank
!C
!
integer (kind = kint) :: neib, istart, inum, iend, ierr, k
@@ -179,7 +167,7 @@
istart= 3 * STACK_EXPORT(neib-1) + 1
inum = 3 * ( STACK_EXPORT(neib ) - STACK_EXPORT(neib-1) )
call MPI_ISEND (WS(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, &
& req1(neib), ierr)
enddo
@@ -190,7 +178,7 @@
istart= 3 * STACK_IMPORT(neib-1) + 1
inum = 3 * ( STACK_IMPORT(neib ) - STACK_IMPORT(neib-1) )
call MPI_IRECV (WR(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, &
& req2(neib), ierr)
enddo
diff --git a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_3.f90 b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_3.f90
index 90634cc..48aedca 100644
--- a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_3.f90
+++ b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_3.f90
@@ -10,13 +10,12 @@
!!
!!@verbatim
!! subroutine SOLVER_SEND_RECV_3 &
-!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT,&
-!! & STACK_EXPORT, NOD_EXPORT,&
-!! & X, SOLVER_COMM,my_rank)
+!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+!! & STACK_EXPORT, NOD_EXPORT, X)
!! subroutine solver_send_recv_3x3 &
-!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT,&
-!! & STACK_EXPORT, NOD_EXPORT,&
-!! & X1, X2, X3, SOLVER_COMM,my_rank)
+!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+!! & STACK_EXPORT, NOD_EXPORT, &
+!! & X1, X2, X3)
!!@endverbatim
!!
!!@n @param N Number of data points
@@ -36,9 +35,6 @@
!!@n @param X1(3*N) 1st vector field data
!!@n @param X2(3*N) 2nd vector field data
!!@n @param X3(3*N) 3rd vector field data
-!!
-!!@n @param SOLVER_COMM MPI communicator
-!!@n @param my_rank own process rank
!
module solver_SR_3
!
@@ -54,12 +50,11 @@
! ----------------------------------------------------------------------
!
!C
-!C*** SOLVER_SEND_RECV
+!C*** SOLVER_SEND_RECV_3
!C
subroutine SOLVER_SEND_RECV_3 &
- & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
- & STACK_EXPORT, NOD_EXPORT, &
- & X, SOLVER_COMM,my_rank)
+ & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+ & STACK_EXPORT, NOD_EXPORT, X)
use calypso_mpi
use m_solver_SR
@@ -82,10 +77,6 @@
& :: NOD_EXPORT
!> communicated result vector
real (kind=kreal), dimension(3*N), intent(inout):: X
-!< communicator for mpi
- integer , intent(in) ::SOLVER_COMM
-!> Own process
- integer , intent(in) :: my_rank
!
integer (kind = kint) :: neib, istart, inum, iend, ierr, k, ii
!
@@ -108,8 +99,8 @@
enddo
istart= 3 * STACK_EXPORT(neib-1) + 1
inum = 3 * ( STACK_EXPORT(neib ) - STACK_EXPORT(neib-1) )
- call MPI_ISEND (WS(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req1(neib), ierr)
+ call MPI_ISEND(WS(istart), inum, MPI_DOUBLE_PRECISION, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, req1(neib), ierr)
enddo
!C
@@ -117,8 +108,8 @@
do neib= 1, NEIBPETOT
istart= 3 * STACK_IMPORT(neib-1) + 1
inum = 3 * ( STACK_IMPORT(neib ) - STACK_IMPORT(neib-1) )
- call MPI_IRECV (WR(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req2(neib), ierr)
+ call MPI_IRECV(WR(istart), inum, MPI_DOUBLE_PRECISION, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, req2(neib), ierr)
enddo
call MPI_WAITALL (NEIBPETOT, req2(1), sta2(1,1), ierr)
@@ -141,9 +132,9 @@
! ---------------------------------------------------------------------
!
subroutine solver_send_recv_3x3 &
- & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
- & STACK_EXPORT, NOD_EXPORT, &
- & X1, X2, X3, SOLVER_COMM,my_rank)
+ & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+ & STACK_EXPORT, NOD_EXPORT, &
+ & X1, X2, X3)
use calypso_mpi
use m_solver_SR
@@ -160,8 +151,6 @@
real (kind=kreal), dimension(3*N), intent(inout):: X1
real (kind=kreal), dimension(3*N), intent(inout):: X2
real (kind=kreal), dimension(3*N), intent(inout):: X3
- integer , intent(in) ::SOLVER_COMM
- integer , intent(in) :: my_rank
!
integer (kind = kint) :: neib, istart, inum, iend, ierr, k, ii
!
@@ -189,8 +178,8 @@
enddo
istart= 9 * STACK_EXPORT(neib-1) + 1
inum = 9 * (STACK_EXPORT(neib ) - STACK_EXPORT(neib-1))
- call MPI_ISEND (WS(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req1(neib), ierr)
+ call MPI_ISEND(WS(istart), inum, MPI_DOUBLE_PRECISION, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, req1(neib), ierr)
enddo
!C
@@ -198,8 +187,8 @@
do neib= 1, NEIBPETOT
istart= 9 * STACK_IMPORT(neib-1) + 1
inum = 9 * ( STACK_IMPORT(neib ) - STACK_IMPORT(neib-1) )
- call MPI_IRECV (WR(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req2(neib), ierr)
+ call MPI_IRECV(WR(istart), inum, MPI_DOUBLE_PRECISION, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, req2(neib), ierr)
enddo
call MPI_WAITALL (NEIBPETOT, req2(1), sta2(1,1), ierr)
diff --git a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_6.f90 b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_6.f90
index faef9db..bc9cea1 100644
--- a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_6.f90
+++ b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_6.f90
@@ -10,9 +10,8 @@
!!
!!@verbatim
!! subroutine SOLVER_SEND_RECV_6 &
-!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT,&
-!! & STACK_EXPORT, NOD_EXPORT,&
-!! & X, SOLVER_COMM,my_rank)
+!! & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+!! & STACK_EXPORT, NOD_EXPORT, X)
!!@endverbatim
!!
!!@n @param N Number of data points
@@ -29,9 +28,6 @@
!! local node ID to copy in export buffer
!!
!!@n @param X(6*N) field data with 6 components
-!!
-!!@n @param SOLVER_COMM MPI communicator
-!!@n @param my_rank own process rank
!
module solver_SR_6
!
@@ -47,9 +43,8 @@
! ----------------------------------------------------------------------
!C
subroutine SOLVER_SEND_RECV_6 &
- & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
- & STACK_EXPORT, NOD_EXPORT, &
- & X, SOLVER_COMM,my_rank)
+ & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+ & STACK_EXPORT, NOD_EXPORT, X)
!
use calypso_mpi
use m_solver_SR
@@ -72,10 +67,6 @@
& :: NOD_EXPORT
!> communicated result vector
real (kind=kreal), dimension(6*N), intent(inout):: X
-!> communicator for mpi
- integer , intent(in) ::SOLVER_COMM
-!> Process ID
- integer , intent(in) :: my_rank
!
integer (kind = kint) :: neib, istart, inum, ierr, k, ii
!
@@ -98,8 +89,8 @@
WS(6*k-1)= X(ii-1)
WS(6*k )= X(ii )
enddo
- call MPI_ISEND (WS(6*istart+1), 6*inum,MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req1(neib), ierr)
+ call MPI_ISEND(WS(6*istart+1), 6*inum,MPI_DOUBLE_PRECISION, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, req1(neib), ierr)
enddo
!C
@@ -107,8 +98,8 @@
do neib= 1, NEIBPETOT
istart= STACK_IMPORT(neib-1)
inum = STACK_IMPORT(neib ) - istart
- call MPI_IRECV (WR(6*istart+1), 6*inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req2(neib), ierr)
+ call MPI_IRECV(WR(6*istart+1), 6*inum, MPI_DOUBLE_PRECISION, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, req2(neib), ierr)
enddo
call MPI_WAITALL (NEIBPETOT, req2(1), sta2(1,1), ierr)
diff --git a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_N.f90 b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_N.f90
index ff31260..744675c 100644
--- a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_N.f90
+++ b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_N.f90
@@ -12,11 +12,11 @@
!! subroutine SOLVER_SEND_RECV_N &
!! & (N, NB, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT,&
!! & STACK_EXPORT, NOD_EXPORT,&
-!! & X, SOLVER_COMM,my_rank)
+!! & X)
!! subroutine SOLVER_SEND_RECV_Nx3 &
!! & (N, NB, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT,&
!! & STACK_EXPORT, NOD_EXPORT,&
-!! & X1, X2, X3, SOLVER_COMM,my_rank)
+!! & X1, X2, X3)
!!@endverbatim
!!
!!@n @param N Number of data points
@@ -37,10 +37,6 @@
!!@n @param X1(NB*N) 1st field data with NB components
!!@n @param X2(NB*N) 2nd field data with NB components
!!@n @param X3(NB*N) 3rd field data with NB components
-!!
-!!@n @param SOLVER_COMM MPI communicator
-!!@n @param my_rank own process rank
-!
!
module solver_SR_N
!
@@ -55,9 +51,8 @@
! ----------------------------------------------------------------------
!
subroutine SOLVER_SEND_RECV_N &
- & ( N, NB, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
- & STACK_EXPORT, NOD_EXPORT, &
- & X, SOLVER_COMM, my_rank)
+ & (N, NB, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+ & STACK_EXPORT, NOD_EXPORT, X)
!
use calypso_mpi
use m_solver_SR
@@ -80,10 +75,6 @@
& :: NOD_EXPORT(STACK_EXPORT(NEIBPETOT))
!> communicated result vector
real (kind=kreal), intent(inout):: X(NB*N)
-!> communicator for mpi
- integer , intent(in) ::SOLVER_COMM
-!> Own process
- integer , intent(in) :: my_rank
!
integer (kind = kint) :: neib, istart, inum, iend
integer (kind = kint) :: ierr, k, ii, ix, nd
@@ -109,7 +100,7 @@
istart= NB * STACK_EXPORT(neib-1) + 1
inum = NB * (STACK_EXPORT(neib ) - STACK_EXPORT(neib-1) )
call MPI_ISEND(WS(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req1(neib), ierr)
+ & NEIBPE(neib), 0, CALYPSO_COMM, req1(neib), ierr)
end do
!C
@@ -118,7 +109,7 @@
istart= NB * STACK_IMPORT(neib-1) + 1
inum = NB * (STACK_IMPORT(neib ) - STACK_IMPORT(neib-1) )
call MPI_IRECV(WR(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req2(neib), ierr)
+ & NEIBPE(neib), 0, CALYPSO_COMM, req2(neib), ierr)
enddo
call MPI_WAITALL (NEIBPETOT, req2(1), sta2(1,1), ierr)
@@ -144,7 +135,7 @@
subroutine SOLVER_SEND_RECV_Nx3 &
& ( N, NB, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
& STACK_EXPORT, NOD_EXPORT, &
- & X1, X2, X3, SOLVER_COMM, my_rank)
+ & X1, X2, X3)
use calypso_mpi
!
@@ -166,9 +157,6 @@
real (kind=kreal), intent(inout):: X2(NB*N)
real (kind=kreal), intent(inout):: X3(NB*N)
!
- integer , intent(in) ::SOLVER_COMM
- integer , intent(in) :: my_rank
-!
integer (kind = kint) :: neib, istart, inum, iend
integer (kind = kint) :: ierr, k, ii, ix, nd, NB3
!
@@ -196,7 +184,7 @@
istart= 3*NB * STACK_EXPORT(neib-1) + 1
inum = 3*NB * (STACK_EXPORT(neib ) - STACK_EXPORT(neib-1) )
call MPI_ISEND(WS(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req1(neib), ierr)
+ & NEIBPE(neib), 0, CALYPSO_COMM, req1(neib), ierr)
end do
!C
@@ -205,7 +193,7 @@
istart= 3*NB * STACK_IMPORT(neib-1) + 1
inum = 3*NB * (STACK_IMPORT(neib ) - STACK_IMPORT(neib-1) )
call MPI_IRECV(WR(istart), inum, MPI_DOUBLE_PRECISION, &
- & NEIBPE(neib), 0, SOLVER_COMM, req2(neib), ierr)
+ & NEIBPE(neib), 0, CALYPSO_COMM, req2(neib), ierr)
enddo
call MPI_WAITALL (NEIBPETOT, req2(1), sta2(1,1), ierr)
diff --git a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_int.f90 b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_int.f90
index ed4f779..732f045 100644
--- a/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_int.f90
+++ b/src/Fortran_libraries/PARALLEL_src/COMM_src/solver_SR_int.f90
@@ -10,9 +10,8 @@
!!
!!@verbatim
!! subroutine solver_send_recv_i &
-!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT,&
-!! & STACK_EXPORT, NOD_EXPORT,&
-!! & ix, SOLVER_COMM,my_rank)
+!! & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+!! & STACK_EXPORT, NOD_EXPORT, ix)
!!@endverbatim
!!
!!@n @param N Number of data points
@@ -29,9 +28,6 @@
!! local node ID to copy in export buffer
!!
!!@n @param ix(N) integer data with NB components
-!!
-!!@n @param SOLVER_COMM MPI communicator
-!!@n @param my_rank own process rank
!
module solver_SR_int
!
@@ -46,9 +42,8 @@
! ----------------------------------------------------------------------
!
subroutine solver_send_recv_i &
- & ( N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
- & STACK_EXPORT, NOD_EXPORT, &
- & ix, SOLVER_COMM,my_rank)
+ & (N, NEIBPETOT, NEIBPE, STACK_IMPORT, NOD_IMPORT, &
+ & STACK_EXPORT, NOD_EXPORT, ix)
!
use calypso_mpi
use m_solver_SR
@@ -71,10 +66,6 @@
& :: NOD_EXPORT
!> communicated result vector
integer (kind=kint), dimension(N) , intent(inout):: iX
-!> communicator for mpi
- integer , intent(in) ::SOLVER_COMM
-!> Own process
- integer , intent(in) :: my_rank
!C
!
integer (kind = kint) :: neib, istart, inum, ierr, k
@@ -93,7 +84,7 @@
iWS(k)= iX(NOD_EXPORT(k))
enddo
call MPI_ISEND (iWS(istart+1), inum, MPI_INTEGER, &
- & NEIBPE(neib), 0, SOLVER_COMM, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, &
& req1(neib), ierr)
enddo
@@ -104,7 +95,7 @@
istart= STACK_IMPORT(neib-1)
inum = STACK_IMPORT(neib ) - istart
call MPI_IRECV (iWR(istart+1), inum, MPI_INTEGER, &
- & NEIBPE(neib), 0, SOLVER_COMM, &
+ & NEIBPE(neib), 0, CALYPSO_COMM, &
& req2(neib), ierr)
enddo
diff --git a/src/Fortran_libraries/PARALLEL_src/SPH_SHELL_src/sum_sph_rms_data.f90 b/src/Fortran_libraries/PARALLEL_src/SPH_SHELL_src/sum_sph_rms_data.f90
index 967b6c6..2ab8cad 100644
--- a/src/Fortran_libraries/PARALLEL_src/SPH_SHELL_src/sum_sph_rms_data.f90
+++ b/src/Fortran_libraries/PARALLEL_src/SPH_SHELL_src/sum_sph_rms_data.f90
@@ -168,6 +168,7 @@
!
subroutine sum_sph_layerd_rms
!
+ use calypso_mpi
use m_parallel_var_dof
use m_spheric_parameter
use m_rms_4_sph_spectr
More information about the CIG-COMMITS
mailing list