Skip to content

Implement a draft for faster reordering #131

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
semi-h wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

Implement a draft for faster reordering #131

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
99 changes: 83 additions & 16 deletions src/omp/backend.f90
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,8 @@ module m_omp_backend
use m_allocator, only: allocator_t, field_t
use m_base_backend, only: base_backend_t
use m_ordering, only: get_index_reordering
use m_common, only: dp, get_dirs_from_rdr, VERT, DIR_X, DIR_Y, DIR_Z, DIR_C
use m_common, only: dp, get_dirs_from_rdr, VERT, DIR_X, DIR_Y, DIR_Z, &
DIR_C, RDR_X2Y, RDR_X2Z
use m_tdsops, only: dirps_t, tdsops_t, get_tds_n
use m_omp_exec_dist, only: exec_dist_tds_compact, exec_dist_transeq_compact
use m_omp_sendrecv, only: sendrecv_fields
Expand Down Expand Up @@ -331,24 +332,55 @@ subroutine reorder_omp(self, u_, u, direction)
class(field_t), intent(in) :: u
integer, intent(in) :: direction
integer, dimension(3) :: dims
integer :: i, j, k
integer :: i, j, k, ka, kb
integer :: out_i, out_j, out_k
integer :: dir_from, dir_to

dims = self%mesh%get_padded_dims(u)
call get_dirs_from_rdr(dir_from, dir_to, direction)

!$omp parallel do private(out_i, out_j, out_k) collapse(2)
do k = 1, dims(3)
do j = 1, dims(2)
do i = 1, dims(1)
call get_index_reordering( &
out_i, out_j, out_k, i, j, k, dir_from, dir_to, self%mesh)
u_%data(out_i, out_j, out_k) = u%data(i, j, k)
select case (direction)
case (RDR_X2Y)
dims = self%mesh%get_padded_dims(DIR_C)
!$omp parallel do collapse(2)
do kb = 1, dims(2)/SZ
do ka = 1, dims(3)
do j = 1, dims(1)
do i = 1, SZ
u_%data(mod(j-1, SZ)+1, (kb-1)*SZ+i, ka+dims(3)*((j-1)/SZ)) = &
u%data(i, j, ka+dims(3)*(kb-1))
end do
end do
end do
end do
end do
!$omp end parallel do
!$omp end parallel do
case (RDR_X2Z)
dims = self%mesh%get_padded_dims(DIR_C)
!$omp parallel do collapse(2)
do kb = 1, dims(2)/SZ
do ka = 1, dims(1)
do j = 1, dims(3)
do i = 1, SZ
u_%data(i, j, ka + (kb - 1)*dims(1)) = &
u%data(i, ka, (kb - 1)*dims(3) + mod(j - 1, dims(3)) + 1)
end do
end do
end do
end do
!$omp end parallel do
case default
dims = self%mesh%get_padded_dims(u)
call get_dirs_from_rdr(dir_from, dir_to, direction)

!$omp parallel do private(out_i, out_j, out_k) collapse(2)
do k = 1, dims(3)
do j = 1, dims(2)
do i = 1, dims(1)
call get_index_reordering( &
out_i, out_j, out_k, i, j, k, dir_from, dir_to, self%mesh)
u_%data(out_i, out_j, out_k) = u%data(i, j, k)
end do
end do
end do
!$omp end parallel do
end select

end subroutine reorder_omp

Expand All @@ -359,7 +391,25 @@ subroutine sum_yintox_omp(self, u, u_)
class(field_t), intent(inout) :: u
class(field_t), intent(in) :: u_

call sum_intox_omp(self, u, u_, DIR_Y)
integer :: i, j, ka, kb, dims(3)

!call sum_intox_omp(self, u, u_, DIR_Y)

dims = self%mesh%get_padded_dims(DIR_C)
!$omp parallel do collapse(2)
do kb = 1, dims(2)/SZ
do ka = 1, dims(3)
do j = 1, dims(1)
do i = 1, SZ
u%data(i, j, ka + dims(3)*(kb - 1)) = &
u%data(i, j, ka + dims(3)*(kb - 1)) &
+ u_%data(mod(j - 1, SZ) + 1, (kb - 1)*SZ + i, &
ka + dims(3)*((j - 1)/SZ))
end do
end do
end do
end do
!$omp end parallel do

end subroutine sum_yintox_omp

Expand All @@ -370,7 +420,24 @@ subroutine sum_zintox_omp(self, u, u_)
class(field_t), intent(inout) :: u
class(field_t), intent(in) :: u_

call sum_intox_omp(self, u, u_, DIR_Z)
integer :: i, j, ka, kb, dims(3)

!call sum_intox_omp(self, u, u_, DIR_Z)

dims = self%mesh%get_padded_dims(DIR_C)
!$omp parallel do collapse(2)
do kb = 1, dims(2)/SZ
do ka = 1, dims(1)
do j = 1, dims(3)
do i = 1, SZ
u%data(i, ka, (kb - 1)*dims(3) + mod(j - 1, dims(3)) + 1) = &
u%data(i, ka, (kb - 1)*dims(3) + mod(j - 1, dims(3)) + 1) &
+ u_%data(i, j, ka + (kb - 1)*dims(1))
end do
end do
end do
end do
!$omp end parallel do

end subroutine sum_zintox_omp

Expand Down