tensor.f90

Go to the documentation of this file.

! Copyright (c) 2008-2020, UCHICAGO ARGONNE, LLC.
!
! The UChicago Argonne, LLC as Operator of Argonne National
! Laboratory holds copyright in the Software. The copyright holder
! reserves all rights except those expressly granted to licensees,
! and U.S. Government license rights.
!
! Redistribution and use in source and binary forms, with or without
! modification, are permitted provided that the following conditions
! are met:
!
! 1. Redistributions of source code must retain the above copyright
! notice, this list of conditions and the disclaimer below.
!
! 2. Redistributions in binary form must reproduce the above copyright
! notice, this list of conditions and the disclaimer (as noted below)
! in the documentation and/or other materials provided with the
! distribution.
!
! 3. Neither the name of ANL nor the names of its contributors
! may be used to endorse or promote products derived from this software
! without specific prior written permission.
!
! THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
! "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
! LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
! FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL
! UCHICAGO ARGONNE, LLC, THE U.S. DEPARTMENT OF
! ENERGY OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
! SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
! TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
! DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
! THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
! (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
! OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
!
! Additional BSD Notice
! ---------------------
! 1. This notice is required to be provided under our contract with
! the U.S. Department of Energy (DOE). This work was produced at
! Argonne National Laboratory under Contract
! No. DE-AC02-06CH11357 with the DOE.
!
! 2. Neither the United States Government nor UCHICAGO ARGONNE,
! LLC nor any of their employees, makes any warranty,
! express or implied, or assumes any liability or responsibility for the
! accuracy, completeness, or usefulness of any information, apparatus,
! product, or process disclosed, or represents that its use would not
! infringe privately-owned rights.
!
! 3. Also, reference herein to any specific commercial products, process,
! or services by trade name, trademark, manufacturer or otherwise does
! not necessarily constitute or imply its endorsement, recommendation,
! or favoring by the United States Government or UCHICAGO ARGONNE LLC.
! The views and opinions of authors expressed
! herein do not necessarily state or reflect those of the United States
! Government or UCHICAGO ARGONNE, LLC, and shall
! not be used for advertising or product endorsement purposes.
!
module tensor
  use tensor_xsmm
  use tensor_cpu
  use tensor_sx
  use tensor_device
  use num_types, only : rp
  use mxm_wrapper
  use neko_config
  use device
  use, intrinsic :: iso_c_binding, only : c_ptr
  implicit none
  private
 
  interface transpose
     module procedure trsp, trsp1
  end interface transpose
 
  interface triple_tensor_product
     module procedure triple_tensor_product_scalar, triple_tensor_product_vector
  end interface triple_tensor_product
 
  public :: tensr3, transpose, trsp, trsp1, &
     tnsr2d_el, tnsr3d_el, tnsr3d, tnsr1_3d, addtnsr, &
     triple_tensor_product, tnsr3d_el_list
 
 
contains
 
  subroutine tensr3(v, nv, u, nu, A, Bt, Ct, w)
    integer :: nv
    integer :: nu
    real(kind=rp), intent(inout) :: v(nv, nv, nv)
    real(kind=rp), intent(inout) :: u(nu, nu, nu)
    real(kind=rp), intent(inout) :: w(nu*nu*nv)
    real(kind=rp), intent(inout) :: a(nv, nu)
    real(kind=rp), intent(inout) :: bt(nu, nv)
    real(kind=rp), intent(inout) :: ct(nu, nv)
    integer :: j, k, l, nunu, nvnv, nunv
 
    nunu = nu**2
    nvnv = nv**2
    nunv = nu*nv
 
 
    call mxm(a, nv, u, nu, v, nunu)
    k = 1
    l = 1
    do j = 1, nu
       call mxm(v(k, 1, 1), nv, bt, nu, w(l), nv)
       k = k + nunv
       l = l + nvnv
    end do
    call mxm(w, nvnv, ct, nu, v, nv)
 
  end subroutine tensr3
 
  subroutine trsp(a, lda, b, ldb)
    integer, intent(in) :: lda
    integer, intent(in) :: ldb
    real(kind=rp), intent(inout) :: a(lda, ldb)
    real(kind=rp), intent(in) :: b(ldb, lda)
    integer :: i, j
 
    do j = 1, ldb
       do i = 1, lda
          a(i, j) = b(j, i)
       end do
    end do
 
  end subroutine trsp
 
  subroutine trsp1(a, n)
    integer, intent(in) :: n
    real(kind=rp), intent(inout) :: a(n, n)
    real(kind=rp) :: tmp
    integer :: i, j
 
    do j = 1, n
       do i = j + 1, n
          tmp = a(i, j)
          a(i, j) = a(j, i)
          a(j, i) = tmp
       end do
    end do
 
  end subroutine trsp1
 
  subroutine tnsr2d_el(v, nv, u, nu, A, Bt)
    integer, intent(in) :: nv, nu
    real(kind=rp), intent(inout) :: v(nv*nv), u(nu*nu)
    real(kind=rp), intent(inout) :: a(nv,nu), bt(nu,nv)
 
    if (neko_bcknd_sx .eq. 1) then
       call tnsr2d_el_sx(v, nv, u, nu, a, bt)
    else if (neko_bcknd_xsmm .eq. 1) then
       call tnsr2d_el_xsmm(v, nv, u, nu, a, bt)
    else
       call tnsr2d_el_cpu(v, nv, u, nu, a, bt)
    end if
 
  end subroutine tnsr2d_el
 
  subroutine tnsr3d_el(v, nv, u, nu, A, Bt, Ct)
    integer, intent(in) :: nv, nu
    real(kind=rp), intent(inout) :: v(nv*nv*nv), u(nu*nu*nu)
    real(kind=rp), intent(inout) :: a(nv,nu),bt(nu, nv),ct(nu,nv)
 
    if (neko_bcknd_sx .eq. 1) then
       call tnsr3d_el_sx(v, nv, u, nu, a, bt, ct)
    else if (neko_bcknd_xsmm .eq. 1) then
       call tnsr3d_el_xsmm(v, nv, u, nu, a, bt, ct)
    else
       call tnsr3d_el_cpu(v, nv, u, nu, a, bt, ct)
    end if
 
  end subroutine tnsr3d_el
 
  subroutine tnsr3d_el_list(v, nv, u, nu, A, Bt, Ct, el_list, n_pt)
    integer, intent(in) :: nv, nu, n_pt, el_list(n_pt)
    real(kind=rp), intent(inout) :: v(nv*nv*nv, n_pt), u(nu*nu*nu,1)
    real(kind=rp), intent(inout) :: a(nv,nu,n_pt),bt(nu, nv,n_pt),ct(nu,nv,n_pt)
    type(c_ptr) :: v_d, u_d, a_d, bt_d, ct_d, el_list_d
    integer :: i
 
    if (neko_bcknd_sx .eq. 1) then
       do i = 1, n_pt
          call tnsr3d_el_sx(v(1,i), nv, u(1,el_list(i)), nu, a(1,1,i), bt(1,1,i), ct(1,1,i))
       end do
    else if (neko_bcknd_xsmm .eq. 1) then
       do i = 1, n_pt
          call tnsr3d_el_xsmm(v(1,i), nv, u(1,el_list(i)), nu, a(1,1,i), bt(1,1,i), ct(1,1,i))
       end do
    else if (neko_bcknd_device .eq. 1) then
       v_d = device_get_ptr(v)
       u_d = device_get_ptr(u)
       a_d = device_get_ptr(a)
       bt_d = device_get_ptr(bt)
       ct_d = device_get_ptr(ct)
       el_list_d = device_get_ptr(el_list)
       call tnsr3d_el_list_device(v_d, nv, u_d, nu, a_d, bt_d, ct_d, el_list_d, n_pt)
    else
       do i = 1, n_pt
          !       Note the use of el_list(i) + 1, because of the gslib C interface
          call tnsr3d_el_cpu(v(1,i), nv, u(1,el_list(i)+1), nu, a(1,1,i), bt(1,1,i), ct(1,1,i))
       end do
    end if
 
  end subroutine tnsr3d_el_list
 
 
  subroutine tnsr3d(v, nv, u, nu, A, Bt, Ct, nelv)
    integer, intent(inout) :: nv, nu, nelv
    real(kind=rp), intent(inout) :: v(nv*nv*nv,nelv), u(nu*nu*nu,nelv)
    real(kind=rp), intent(inout) :: a(nv,nu), bt(nu, nv), ct(nu,nv)
    type(c_ptr) :: v_d, u_d, a_d, bt_d, ct_d
 
 
    if (neko_bcknd_sx .eq. 1) then
       call tnsr3d_sx(v, nv, u, nu, a, bt, ct, nelv)
    else if (neko_bcknd_xsmm .eq. 1) then
       call tnsr3d_xsmm(v, nv, u, nu, a, bt, ct, nelv)
    else if (neko_bcknd_device .eq. 1) then
       v_d = device_get_ptr(v)
       u_d = device_get_ptr(u)
       a_d = device_get_ptr(a)
       bt_d = device_get_ptr(bt)
       ct_d = device_get_ptr(ct)
       call tnsr3d_device(v_d, nv, u_d, nu, a_d, bt_d, ct_d, nelv)
    else
       call tnsr3d_cpu(v, nv, u, nu, a, bt, ct, nelv)
    end if
 
  end subroutine tnsr3d
 
  subroutine tnsr1_3d(v, nv, nu, A, Bt, Ct, nelv)
    integer, intent(inout) :: nv, nu, nelv
    real(kind=rp), intent(inout) :: v(nv*nv*nv*nelv)
    real(kind=rp), intent(inout) :: a(nv,nu), bt(nu, nv), ct(nu,nv)
 
    if (neko_bcknd_sx .eq. 1) then
       call tnsr1_3d_sx(v, nv, nu, a, bt, ct, nelv)
    else if (neko_bcknd_xsmm .eq. 1) then
       call tnsr1_3d_xsmm(v, nv, nu, a, bt, ct, nelv)
    else
       call tnsr1_3d_cpu(v, nv, nu, a, bt, ct, nelv)
    end if
 
  end subroutine tnsr1_3d
 
  subroutine addtnsr(s, h1, h2, h3, nx, ny, nz)
 
    integer, intent(in) :: nx, ny, nz
    real(kind=rp), intent(in) :: h1(nx), h2(ny), h3(nz)
    real(kind=rp), intent(inout) ::  s(nx, ny, nz)
    real(kind=rp) :: hh
    integer :: ix, iy, iz
 
    do iz = 1,nz
       do iy = 1,ny
          hh = h2(iy)*h3(iz)
          do ix = 1,nx
             s(ix,iy,iz) = s(ix,iy,iz)+hh*h1(ix)
          end do
       end do
    end do
 
  end subroutine addtnsr
 
  subroutine triple_tensor_product_scalar(v, u, nu, Hr, Hs, Ht)
    real(kind=rp), intent(inout) :: v
    integer, intent(in) :: nu
    real(kind=rp), intent(inout) :: u(nu,nu,nu)
    real(kind=rp), intent(inout) :: hr(nu)
    real(kind=rp), intent(inout) :: hs(nu)
    real(kind=rp), intent(inout) :: ht(nu)
 
    ! Artificially reshape v into a 1-dimensional array
    ! since this is what tnsr3d_el needs as input argument
    real(kind=rp) :: vv(1)
    ! vv(1) = v
 
    call tnsr3d_el(vv,1,u,nu,hr,hs,ht)
 
    v = vv(1)
 
  end subroutine triple_tensor_product_scalar
 
  subroutine triple_tensor_product_vector(v, u1, u2, u3, nu, Hr, Hs, Ht)
    real(kind=rp), intent(inout) :: v(3)
    integer, intent(in) :: nu
    real(kind=rp), intent(inout) :: u1(nu,nu,nu)
    real(kind=rp), intent(inout) :: u2(nu,nu,nu)
    real(kind=rp), intent(inout) :: u3(nu,nu,nu)
    real(kind=rp), intent(inout) :: hr(nu)
    real(kind=rp), intent(inout) :: hs(nu)
    real(kind=rp), intent(inout) :: ht(nu)
 
    call triple_tensor_product_scalar(v(1), u1, nu, hr, hs, ht)
    call triple_tensor_product_scalar(v(2), u2, nu, hr, hs, ht)
    call triple_tensor_product_scalar(v(3), u3, nu, hr, hs, ht)
 
  end subroutine triple_tensor_product_vector
 
end module tensor