diff options
Diffstat (limited to 'src/Dissipation_4_2.F90')
-rw-r--r-- | src/Dissipation_4_2.F90 | 283 |
1 files changed, 163 insertions, 120 deletions
diff --git a/src/Dissipation_4_2.F90 b/src/Dissipation_4_2.F90 index d2c6d89..3ca4316 100644 --- a/src/Dissipation_4_2.F90 +++ b/src/Dissipation_4_2.F90 @@ -24,6 +24,7 @@ subroutine dissipation_4_2 (var, ni, nj, nk, bb, gsize, offset, delta, epsilon, CCTK_REAL :: idel CCTK_INT :: il, ir, jl, jr, kl, kr, ol, or + CCTK_INT :: i, j, k call set_coeff ( a ) @@ -37,22 +38,26 @@ subroutine dissipation_4_2 (var, ni, nj, nk, bb, gsize, offset, delta, epsilon, il = 1 + gsize(1) else ol = offset(1) -!$omp parallel workshare - rhs(1+ol,:,:) = rhs(1+ol,:,:) + & - ( a(1,1) * var(1+ol,:,:) + a(2,1) * var(2+ol,:,:) + & - a(3,1) * var(3+ol,:,:) ) * idel - rhs(2+ol,:,:) = rhs(2+ol,:,:) + & - ( a(1,2) * var(1+ol,:,:) + a(2,2) * var(2+ol,:,:) + & - a(3,2) * var(3+ol,:,:) + a(4,2) * var(4+ol,:,:) ) * idel - rhs(3+ol,:,:) = rhs(3+ol,:,:) + & - ( a(1,3) * var(1+ol,:,:) + a(2,3) * var(2+ol,:,:) + & - a(3,3) * var(3+ol,:,:) + a(4,3) * var(4+ol,:,:) + & - a(5,3) * var(5+ol,:,:) ) * idel - rhs(4+ol,:,:) = rhs(4+ol,:,:) + & - ( a(2,4) * var(2+ol,:,:) + a(3,4) * var(3+ol,:,:) + & - a(4,4) * var(4+ol,:,:) + a(5,4) * var(5+ol,:,:) + & - a(6,4) * var(6+ol,:,:) ) * idel -!$omp end parallel workshare +!$omp parallel do private(j,k) + do k=1,nk + do j=1,nj + rhs(1+ol,j,k) = rhs(1+ol,j,k) + & + ( a(1,1) * var(1+ol,j,k) + a(2,1) * var(2+ol,j,k) + & + a(3,1) * var(3+ol,j,k) ) * idel + rhs(2+ol,j,k) = rhs(2+ol,j,k) + & + ( a(1,2) * var(1+ol,j,k) + a(2,2) * var(2+ol,j,k) + & + a(3,2) * var(3+ol,j,k) + a(4,2) * var(4+ol,j,k) ) * idel + rhs(3+ol,j,k) = rhs(3+ol,j,k) + & + ( a(1,3) * var(1+ol,j,k) + a(2,3) * var(2+ol,j,k) + & + a(3,3) * var(3+ol,j,k) + a(4,3) * var(4+ol,j,k) + & + a(5,3) * var(5+ol,j,k) ) * idel + rhs(4+ol,j,k) = rhs(4+ol,j,k) + & + ( a(2,4) * var(2+ol,j,k) + a(3,4) * var(3+ol,j,k) + & + a(4,4) * var(4+ol,j,k) + a(5,4) * var(5+ol,j,k) + & + a(6,4) * var(6+ol,j,k) ) * idel + end do + end do +!$omp end parallel do il = 5 + ol end if @@ -60,33 +65,43 @@ subroutine dissipation_4_2 (var, ni, nj, nk, bb, gsize, offset, delta, epsilon, ir = ni - gsize(1) else or = ni - offset(2) -!$omp parallel workshare - rhs(or-3,:,:) = rhs(or-3,:,:) + & - ( a(2,4) * var(or-1,:,:) + a(3,4) * var(or-2,:,:) + & - a(4,4) * var(or-3,:,:) + a(5,4) * var(or-4,:,:) + & - a(6,4) * var(or-5,:,:) ) * idel - rhs(or-2,:,:) = rhs(or-2,:,:) + & - ( a(1,3) * var(or,:,:) + a(2,3) * var(or-1,:,:) + & - a(3,3) * var(or-2,:,:) + a(4,3) * var(or-3,:,:) + & - a(5,3) * var(or-4,:,:) ) * idel - rhs(or-1,:,:) = rhs(or-1,:,:) + & - ( a(1,2) * var(or,:,:) + a(2,2) * var(or-1,:,:) + & - a(3,2) * var(or-2,:,:) + a(4,2) * var(or-3,:,:) ) * idel - rhs(or,:,:) = rhs(or,:,:) + & - ( a(1,1) * var(or,:,:) + a(2,1) * var(or-1,:,:) + & - a(3,1) * var(or-2,:,:) ) * idel -!$omp end parallel workshare +!$omp parallel do private(j,k) + do k=1,nk + do j=1,nj + rhs(or-3,j,k) = rhs(or-3,j,k) + & + ( a(2,4) * var(or-1,j,k) + a(3,4) * var(or-2,j,k) + & + a(4,4) * var(or-3,j,k) + a(5,4) * var(or-4,j,k) + & + a(6,4) * var(or-5,j,k) ) * idel + rhs(or-2,j,k) = rhs(or-2,j,k) + & + ( a(1,3) * var(or,j,k) + a(2,3) * var(or-1,j,k) + & + a(3,3) * var(or-2,j,k) + a(4,3) * var(or-3,j,k) + & + a(5,3) * var(or-4,j,k) ) * idel + rhs(or-1,j,k) = rhs(or-1,j,k) + & + ( a(1,2) * var(or,j,k) + a(2,2) * var(or-1,j,k) + & + a(3,2) * var(or-2,j,k) + a(4,2) * var(or-3,j,k) ) * idel + rhs(or,j,k) = rhs(or,j,k) + & + ( a(1,1) * var(or,j,k) + a(2,1) * var(or-1,j,k) + & + a(3,1) * var(or-2,j,k) ) * idel + end do + end do +!$omp end parallel do ir = or - 4 end if -!$omp parallel workshare - rhs(il:ir,:,:) = rhs(il:ir,:,:) + & - ( -6.0_wp * var(il:ir,:,:) + & - 4.0_wp * ( var(il-1:ir-1,:,:) + & - var(il+1:ir+1,:,:) ) - & - ( var(il-2:ir-2,:,:) + & - var(il+2:ir+2,:,:) ) ) * idel -!$omp end parallel workshare +!$omp parallel do private(i,j,k) + do k=1,nk + do j=1,nj + do i=il,ir + rhs(i,j,k) = rhs(i,j,k) + & + ( -6.0_wp * var(i,j,k) + & + 4.0_wp * ( var(i-1,j,k) + & + var(i+1,j,k) ) - & + ( var(i-2,j,k) + & + var(i+2,j,k) ) ) * idel + end do + end do + end do +!$omp end parallel do if ( zero_derivs_y == 0 ) then call set_coeff ( a ) @@ -101,22 +116,26 @@ subroutine dissipation_4_2 (var, ni, nj, nk, bb, gsize, offset, delta, epsilon, jl = 1 + gsize(2) else ol = offset(3) -!$omp parallel workshare - rhs(:,1+ol,:) = rhs(:,1+ol,:) + & - ( a(1,1) * var(:,1+ol,:) + a(2,1) * var(:,2+ol,:) + & - a(3,1) * var(:,3+ol,:) ) * idel - rhs(:,2+ol,:) = rhs(:,2+ol,:) + & - ( a(1,2) * var(:,1+ol,:) + a(2,2) * var(:,2+ol,:) + & - a(3,2) * var(:,3+ol,:) + a(4,2) * var(:,4+ol,:) ) * idel - rhs(:,3+ol,:) = rhs(:,3+ol,:) + & - ( a(1,3) * var(:,1+ol,:) + a(2,3) * var(:,2+ol,:) + & - a(3,3) * var(:,3+ol,:) + a(4,3) * var(:,4+ol,:) + & - a(5,3) * var(:,5+ol,:) ) * idel - rhs(:,4+ol,:) = rhs(:,4+ol,:) + & - ( a(2,4) * var(:,2+ol,:) + a(3,4) * var(:,3+ol,:) + & - a(4,4) * var(:,4+ol,:) + a(5,4) * var(:,5+ol,:) + & - a(6,4) * var(:,6+ol,:) ) * idel -!$omp end parallel workshare +!$omp parallel do private(i,k) + do k=1,nk + do i=1,ni + rhs(i,1+ol,k) = rhs(i,1+ol,k) + & + ( a(1,1) * var(i,1+ol,k) + a(2,1) * var(i,2+ol,k) + & + a(3,1) * var(i,3+ol,k) ) * idel + rhs(i,2+ol,k) = rhs(i,2+ol,k) + & + ( a(1,2) * var(i,1+ol,k) + a(2,2) * var(i,2+ol,k) + & + a(3,2) * var(i,3+ol,k) + a(4,2) * var(i,4+ol,k) ) * idel + rhs(i,3+ol,k) = rhs(i,3+ol,k) + & + ( a(1,3) * var(i,1+ol,k) + a(2,3) * var(i,2+ol,k) + & + a(3,3) * var(i,3+ol,k) + a(4,3) * var(i,4+ol,k) + & + a(5,3) * var(i,5+ol,k) ) * idel + rhs(i,4+ol,k) = rhs(i,4+ol,k) + & + ( a(2,4) * var(i,2+ol,k) + a(3,4) * var(i,3+ol,k) + & + a(4,4) * var(i,4+ol,k) + a(5,4) * var(i,5+ol,k) + & + a(6,4) * var(i,6+ol,k) ) * idel + end do + end do +!$omp end parallel do jl = 5 + ol end if @@ -124,33 +143,43 @@ subroutine dissipation_4_2 (var, ni, nj, nk, bb, gsize, offset, delta, epsilon, jr = nj - gsize(2) else or = nj - offset(4) -!$omp parallel workshare - rhs(:,or-3,:) = rhs(:,or-3,:) + & - ( a(2,4) * var(:,or-1,:) + a(3,4) * var(:,or-2,:) + & - a(4,4) * var(:,or-3,:) + a(5,4) * var(:,or-4,:) + & - a(6,4) * var(:,or-5,:) ) * idel - rhs(:,or-2,:) = rhs(:,or-2,:) + & - ( a(1,3) * var(:,or,:) + a(2,3) * var(:,or-1,:) + & - a(3,3) * var(:,or-2,:) + a(4,3) * var(:,or-3,:) + & - a(5,3) * var(:,or-4,:) ) * idel - rhs(:,or-1,:) = rhs(:,or-1,:) + & - ( a(1,2) * var(:,or,:) + a(2,2) * var(:,or-1,:) + & - a(3,2) * var(:,or-2,:) + a(4,2) * var(:,or-3,:) ) * idel - rhs(:,or,:) = rhs(:,or,:) + & - ( a(1,1) * var(:,or,:) + a(2,1) * var(:,or-1,:) + & - a(3,1) * var(:,or-2,:) ) * idel -!$omp end parallel workshare +!$omp parallel do private(i,k) + do k=1,nk + do i=1,ni + rhs(i,or-3,k) = rhs(i,or-3,k) + & + ( a(2,4) * var(i,or-1,k) + a(3,4) * var(i,or-2,k) + & + a(4,4) * var(i,or-3,k) + a(5,4) * var(i,or-4,k) + & + a(6,4) * var(i,or-5,k) ) * idel + rhs(i,or-2,k) = rhs(i,or-2,k) + & + ( a(1,3) * var(i,or,k) + a(2,3) * var(i,or-1,k) + & + a(3,3) * var(i,or-2,k) + a(4,3) * var(i,or-3,k) + & + a(5,3) * var(i,or-4,k) ) * idel + rhs(i,or-1,k) = rhs(i,or-1,k) + & + ( a(1,2) * var(i,or,k) + a(2,2) * var(i,or-1,k) + & + a(3,2) * var(i,or-2,k) + a(4,2) * var(i,or-3,k) ) * idel + rhs(i,or,k) = rhs(i,or,k) + & + ( a(1,1) * var(i,or,k) + a(2,1) * var(i,or-1,k) + & + a(3,1) * var(i,or-2,k) ) * idel + end do + end do +!$omp end parallel do jr = or - 4 end if -!$omp parallel workshare - rhs(:,jl:jr,:) = rhs(:,jl:jr,:) + & - ( -6.0_wp * var(:,jl:jr,:) + & - 4.0_wp * ( var(:,jl-1:jr-1,:) + & - var(:,jl+1:jr+1,:) ) - & - ( var(:,jl-2:jr-2,:) + & - var(:,jl+2:jr+2,:) ) ) * idel -!$omp end parallel workshare +!$omp parallel do private(i,j,k) + do k=1,nk + do j=jl,jr + do i=1,ni + rhs(i,j,k) = rhs(i,j,k) + & + ( -6.0_wp * var(i,j,k) + & + 4.0_wp * ( var(i,j-1,k) + & + var(i,j+1,k) ) - & + ( var(i,j-2,k) + & + var(i,j+2,k) ) ) * idel + end do + end do + end do +!$omp end parallel do end if if ( zero_derivs_z == 0 ) then @@ -166,22 +195,26 @@ subroutine dissipation_4_2 (var, ni, nj, nk, bb, gsize, offset, delta, epsilon, kl = 1 + gsize(3) else ol = offset(5) -!$omp parallel workshare - rhs(:,:,1+ol) = rhs(:,:,1+ol) + & - ( a(1,1) * var(:,:,1+ol) + a(2,1) * var(:,:,2+ol) + & - a(3,1) * var(:,:,3+ol) ) * idel - rhs(:,:,2+ol) = rhs(:,:,2+ol) + & - ( a(1,2) * var(:,:,1+ol) + a(2,2) * var(:,:,2+ol) + & - a(3,2) * var(:,:,3+ol) + a(4,2) * var(:,:,4+ol) ) * idel - rhs(:,:,3+ol) = rhs(:,:,3+ol) + & - ( a(1,3) * var(:,:,1+ol) + a(2,3) * var(:,:,2+ol) + & - a(3,3) * var(:,:,3+ol) + a(4,3) * var(:,:,4+ol) + & - a(5,3) * var(:,:,5+ol) ) * idel - rhs(:,:,4+ol) = rhs(:,:,4+ol) + & - ( a(2,4) * var(:,:,2+ol) + a(3,4) * var(:,:,3+ol) + & - a(4,4) * var(:,:,4+ol) + a(5,4) * var(:,:,5+ol) + & - a(6,4) * var(:,:,6+ol) ) * idel -!$omp end parallel workshare +!$omp parallel do private(i,j) + do j=1,nj + do i=1,ni + rhs(i,j,1+ol) = rhs(i,j,1+ol) + & + ( a(1,1) * var(i,j,1+ol) + a(2,1) * var(i,j,2+ol) + & + a(3,1) * var(i,j,3+ol) ) * idel + rhs(i,j,2+ol) = rhs(i,j,2+ol) + & + ( a(1,2) * var(i,j,1+ol) + a(2,2) * var(i,j,2+ol) + & + a(3,2) * var(i,j,3+ol) + a(4,2) * var(i,j,4+ol) ) * idel + rhs(i,j,3+ol) = rhs(i,j,3+ol) + & + ( a(1,3) * var(i,j,1+ol) + a(2,3) * var(i,j,2+ol) + & + a(3,3) * var(i,j,3+ol) + a(4,3) * var(i,j,4+ol) + & + a(5,3) * var(i,j,5+ol) ) * idel + rhs(i,j,4+ol) = rhs(i,j,4+ol) + & + ( a(2,4) * var(i,j,2+ol) + a(3,4) * var(i,j,3+ol) + & + a(4,4) * var(i,j,4+ol) + a(5,4) * var(i,j,5+ol) + & + a(6,4) * var(i,j,6+ol) ) * idel + end do + end do +!$omp end parallel do kl = 5 + ol end if @@ -189,33 +222,43 @@ subroutine dissipation_4_2 (var, ni, nj, nk, bb, gsize, offset, delta, epsilon, kr = nk - gsize(3) else or = nk - offset(6) -!$omp parallel workshare - rhs(:,:,or-3) = rhs(:,:,or-3) + & - ( a(2,4) * var(:,:,or-1) + a(3,4) * var(:,:,or-2) + & - a(4,4) * var(:,:,or-3) + a(5,4) * var(:,:,or-4) + & - a(6,4) * var(:,:,or-5) ) * idel - rhs(:,:,or-2) = rhs(:,:,or-2) + & - ( a(1,3) * var(:,:,or) + a(2,3) * var(:,:,or-1) + & - a(3,3) * var(:,:,or-2) + a(4,3) * var(:,:,or-3) + & - a(5,3) * var(:,:,or-4) ) * idel - rhs(:,:,or-1) = rhs(:,:,or-1) + & - ( a(1,2) * var(:,:,or) + a(2,2) * var(:,:,or-1) + & - a(3,2) * var(:,:,or-2) + a(4,2) * var(:,:,or-3) ) * idel - rhs(:,:,or) = rhs(:,:,or) + & - ( a(1,1) * var(:,:,or) + a(2,1) * var(:,:,or-1) + & - a(3,1) * var(:,:,or-2) ) * idel -!$omp end parallel workshare +!$omp parallel do private(i,j) + do j=1,nj + do i=1,ni + rhs(i,j,or-3) = rhs(i,j,or-3) + & + ( a(2,4) * var(i,j,or-1) + a(3,4) * var(i,j,or-2) + & + a(4,4) * var(i,j,or-3) + a(5,4) * var(i,j,or-4) + & + a(6,4) * var(i,j,or-5) ) * idel + rhs(i,j,or-2) = rhs(i,j,or-2) + & + ( a(1,3) * var(i,j,or) + a(2,3) * var(i,j,or-1) + & + a(3,3) * var(i,j,or-2) + a(4,3) * var(i,j,or-3) + & + a(5,3) * var(i,j,or-4) ) * idel + rhs(i,j,or-1) = rhs(i,j,or-1) + & + ( a(1,2) * var(i,j,or) + a(2,2) * var(i,j,or-1) + & + a(3,2) * var(i,j,or-2) + a(4,2) * var(i,j,or-3) ) * idel + rhs(i,j,or) = rhs(i,j,or) + & + ( a(1,1) * var(i,j,or) + a(2,1) * var(i,j,or-1) + & + a(3,1) * var(i,j,or-2) ) * idel + end do + end do +!$omp end parallel do kr = or - 4 end if -!$omp parallel workshare - rhs(:,:,kl:kr) = rhs(:,:,kl:kr) + & - ( -6.0_wp * var(:,:,kl:kr) + & - 4.0_wp * ( var(:,:,kl-1:kr-1) + & - var(:,:,kl+1:kr+1) ) - & - ( var(:,:,kl-2:kr-2) + & - var(:,:,kl+2:kr+2) ) ) * idel -!$omp end parallel workshare +!$omp parallel do private(i,j,k) + do k=kl,kr + do j=1,nj + do i=1,ni + rhs(i,j,k) = rhs(i,j,k) + & + ( -6.0_wp * var(i,j,k) + & + 4.0_wp * ( var(i,j,k-1) + & + var(i,j,k+1) ) - & + ( var(i,j,k-2) + & + var(i,j,k+2) ) ) * idel + end do + end do + end do +!$omp end parallel do end if contains |