elpa2_bandred_template.X90

#if 0
!    This file is part of ELPA.
!
!    The ELPA library was originally created by the ELPA consortium,
!    consisting of the following organizations:
!
!    - Max Planck Computing and Data Facility (MPCDF), fomerly known as
!      Rechenzentrum Garching der Max-Planck-Gesellschaft (RZG),
!    - Bergische Universität Wuppertal, Lehrstuhl für angewandte
!      Informatik,
!    - Technische Universität München, Lehrstuhl für Informatik mit
!      Schwerpunkt Wissenschaftliches Rechnen ,
!    - Fritz-Haber-Institut, Berlin, Abt. Theorie,
!    - Max-Plack-Institut für Mathematik in den Naturwissenschaften,
!      Leipzig, Abt. Komplexe Strukutren in Biologie und Kognition,
!      and
!    - IBM Deutschland GmbH
!
!    This particular source code file contains additions, changes and
!    enhancements authored by Intel Corporation which is not part of
!    the ELPA consortium.
!
!    More information can be found here:
!    http://elpa.mpcdf.mpg.de/
!
!    ELPA is free software: you can redistribute it and/or modify
!    it under the terms of the version 3 of the license of the
!    GNU Lesser General Public License as published by the Free
!    Software Foundation.
!
!    ELPA is distributed in the hope that it will be useful,
!    but WITHOUT ANY WARRANTY; without even the implied warranty of
!    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
!    GNU Lesser General Public License for more details.
!
!    You should have received a copy of the GNU Lesser General Public License
!    along with ELPA.  If not, see <http://www.gnu.org/licenses/>
!
!    ELPA reflects a substantial effort on the part of the original
!    ELPA consortium, and we ask you to respect the spirit of the
!    license that we chose: i.e., please contribute any changes you
!    may have back to the original ELPA library distribution, and keep
!    any derivatives of ELPA under the same license that we chose for
!    the original distribution, the GNU Lesser General Public License.
!
!
! ELPA1 -- Faster replacements for ScaLAPACK symmetric eigenvalue routines
!
! Copyright of the original code rests with the authors inside the ELPA
! consortium. The copyright of any additional modifications shall rest
! with their original authors, but shall adhere to the licensing terms
! distributed along with the original code in the file "COPYING".



! ELPA2 -- 2-stage solver for ELPA
!
! Copyright of the original code rests with the authors inside the ELPA
! consortium. The copyright of any additional modifications shall rest
! with their original authors, but shall adhere to the licensing terms
! distributed along with the original code in the file "COPYING".
#endif
    subroutine bandred_&
    &MATH_DATATYPE&
    &_&
    &PRECISION &
    (na, a, &
#if REALCASE == 1
     a_dev, &
#endif
     lda, nblk, nbw, matrixCols, numBlocks, mpi_comm_rows, mpi_comm_cols, tmat, &
#if REALCASE == 1
     tmat_dev, &
#endif
     wantDebug, useGPU, success &
#if REALCASE == 1
     , useQR)
#endif
#if COMPLEXCASE == 1
     )
#endif

  !-------------------------------------------------------------------------------
  !  bandred_real/complex: Reduces a distributed symmetric matrix to band form
  !
  !  Parameters
  !
  !  na          Order of matrix
  !
  !  a(lda,matrixCols)    Distributed matrix which should be reduced.
  !              Distribution is like in Scalapack.
  !              Opposed to Scalapack, a(:,:) must be set completely (upper and lower half)
  !              a(:,:) is overwritten on exit with the band and the Householder vectors
  !              in the upper half.
  !
  !  lda         Leading dimension of a
  !  matrixCols  local columns of matrix a
  !
  !  nblk        blocksize of cyclic distribution, must be the same in both directions!
  !
  !  nbw         semi bandwith of output matrix
  !
  !  mpi_comm_rows
  !  mpi_comm_cols
  !              MPI-Communicators for rows/columns
  !
  !  tmat(nbw,nbw,numBlocks)    where numBlocks = (na-1)/nbw + 1
  !              Factors for the Householder vectors (returned), needed for back transformation
  !
  !-------------------------------------------------------------------------------

      use cuda_functions
      use iso_c_binding
      use elpa1_compute
#ifdef HAVE_DETAILED_TIMINGS
      use timings
#else
      use timings_dummy
#endif
#ifdef WITH_OPENMP
      use omp_lib
#endif
      use precision
      implicit none

      integer(kind=ik)                            :: na, lda, nblk, nbw, matrixCols, numBlocks, mpi_comm_rows, mpi_comm_cols

#if REALCASE == 1
#ifdef USE_ASSUMED_SIZE
      real(kind=REAL_DATATYPE)                    :: a(lda,*), tmat(nbw,nbw,*)
#else
      real(kind=REAL_DATATYPE)                    :: a(lda,matrixCols), tmat(nbw,nbw,numBlocks)
#endif
#endif
#if COMPLEXCASE == 1
#ifdef USE_ASSUMED_SIZE
      complex(kind=COMPLEX_DATATYPE)              :: a(lda,*), tmat(nbw,nbw,*)
#else
      complex(kind=COMPLEX_DATATYPE)              :: a(lda,matrixCols), tmat(nbw,nbw,numBlocks)
#endif
#endif /* COMPLEXCASE */

#if REALCASE == 1
#ifdef DOUBLE_PRECISION_REAL
      real(kind=REAL_DATATYPE), parameter         :: ZERO = 0.0_rk8, ONE = 1.0_rk8

#else
      real(kind=REAL_DATATYPE), parameter         :: ZERO = 0.0_rk4, ONE = 1.0_rk4
#endif
#endif

#if COMPLEXCASE == 1
#ifdef DOUBLE_PRECISION_COMPLEX
      complex(kind=COMPLEX_DATATYPE), parameter   :: ZERO = (0.0_rk8, 0.0_rk8), ONE = (1.0_rk8, 0.0_rk8)
#else
      complex(kind=COMPLEX_DATATYPE), parameter   :: ZERO = (0.0_rk4, 0.0_rk4), ONE = (1.0_rk4, 0.0_rk4)
#endif
#endif /* COMPLEXCASE == 1 */

#if REALCASE == 1
      real(kind=REAL_DATATYPE)                    :: eps
#endif
      logical, intent(in)                         :: useGPU

      integer(kind=ik)                            :: my_prow, my_pcol, np_rows, np_cols, mpierr
      integer(kind=ik)                            :: l_cols, l_rows
#if REALCASE == 1
      integer(kind=ik)                            :: vmrCols, mynlc
#endif
      integer(kind=ik)                            :: i, j, lcs, lce, lrs, lre, lc, lr, cur_pcol, n_cols, nrow
      integer(kind=ik)                            :: istep, ncol, lch, lcx, nlc
      integer(kind=ik)                            :: tile_size, l_rows_tile, l_cols_tile

      real(kind=REAL_DATATYPE)                    :: vnorm2
#if REALCASE == 1
      real(kind=REAL_DATATYPE)                    :: xf, aux1(nbw), aux2(nbw), vrl, tau, vav(nbw,nbw)
#endif
#if COMPLEXCASE == 1
      complex(kind=COMPLEX_DATATYPE)              :: xf, aux1(nbw), aux2(nbw), vrl, tau, vav(nbw,nbw)
#endif

#if COMPLEXCASE == 1
      complex(kind=COMPLEX_DATATYPE), allocatable :: tmpCUDA(:,:), vmrCUDA(:,:), umcCUDA(:,:) ! note the different dimension in real case
      complex(kind=COMPLEX_DATATYPE), allocatable :: tmpCPU(:,:), vmrCPU(:,:), umcCPU(:,:)
      complex(kind=COMPLEX_DATATYPE), allocatable :: vr(:)
#endif

#if REALCASE == 1
      real(kind=REAL_DATATYPE), allocatable       :: tmpCUDA(:),  vmrCUDA(:),  umcCUDA(:)
      real(kind=REAL_DATATYPE), allocatable       :: tmpCPU(:,:), vmrCPU(:,:), umcCPU(:,:)
      real(kind=REAL_DATATYPE), allocatable       :: vr(:)
#endif

#if REALCASE == 1
      ! needed for blocked QR decomposition
      integer(kind=ik)                            :: PQRPARAM(11), work_size
      real(kind=REAL_DATATYPE)                    :: dwork_size(1)
      real(kind=REAL_DATATYPE), allocatable       :: work_blocked(:), tauvector(:), blockheuristic(:)
#endif
      ! a_dev is passed from bandred_real to trans_ev_band
      integer(kind=C_intptr_T)                    :: a_dev, vmr_dev, umc_dev, tmat_dev, vav_dev
#ifdef WITH_MPI
      integer(kind=ik), external                  :: numroc
#endif
      integer(kind=ik)                            :: ierr
      integer(kind=ik)                            :: cur_l_rows, cur_l_cols, vmr_size, umc_size
      integer(kind=c_size_t)                      :: lc_start, lc_end
#if COMPLEXCASE == 1
      integer(kind=c_size_t)                      :: lce_1, lcs_1, lre_1
#endif
      integer(kind=ik)                            :: lr_end
      integer(kind=ik)                            :: na_cols
#if COMPLEXCASE == 1
      integer(kind=ik)                            :: na_rows
#endif

      logical, intent(in)                         :: wantDebug
      logical, intent(out)                        :: success
      logical                                     :: successCUDA
      integer(kind=ik)                            :: istat
      character(200)                              :: errorMessage

#if REALCASE == 1
      logical, intent(in)                         :: useQR
#endif
#if REALCASE == 1
      integer(kind=ik)                            :: mystart, myend, m_way, n_way, work_per_thread, m_id, n_id, n_threads, &
                                                    ii, pp, transformChunkSize
#endif

      call timer%start("bandred_&
      &MATH_DATATYPE&
      &" // &
      &PRECISION_SUFFIX &
      )
      call timer%start("mpi_communication")

      call mpi_comm_rank(mpi_comm_rows,my_prow,mpierr)
      call mpi_comm_size(mpi_comm_rows,np_rows,mpierr)
      call mpi_comm_rank(mpi_comm_cols,my_pcol,mpierr)
      call mpi_comm_size(mpi_comm_cols,np_cols,mpierr)

      call timer%stop("mpi_communication")
      success = .true.


      ! Semibandwith nbw must be a multiple of blocksize nblk
      if (mod(nbw,nblk)/=0) then
        if (my_prow==0 .and. my_pcol==0) then
          if (wantDebug) then
            write(error_unit,*) 'ELPA2_bandred_&
	    &MATH_DATATYPE&
	    &: ERROR: nbw=',nbw,', nblk=',nblk
            write(error_unit,*) 'ELPA2_bandred_&
	    &MATH_DATATYPE&
	    &: ELPA2 works only for nbw==n*nblk'
          endif
          success = .false.
          return
        endif
      endif

! na_rows in used nowhere; only na_cols
      if (useGPU) then
#ifdef WITH_MPI
#if COMPLEXCASE == 1
        na_rows = numroc(na, nblk, my_prow, 0, np_rows)
#endif
        na_cols = numroc(na, nblk, my_pcol, 0, np_cols)
#else
#if COMPLEXCASE == 1
         na_rows = na
#endif
        na_cols = na
#endif /* WITH_MPI */

        ! Here we convert the regular host array into a pinned host array
        successCUDA = cuda_malloc(a_dev, lda*na_cols*  &
#if REALCASE == 1
	                          size_of_PRECISION_real)
#endif
#if COMPLEXCASE == 1
                                  size_of_PRECISION_complex)
#endif
        if (.not.(successCUDA)) then
          print *,"bandred_&
	  &MATH_DATATYPE&
	  &: error in cudaMalloc"
          stop
        endif

        successCUDA = cuda_malloc(tmat_dev, nbw*nbw*   &
#if REALCASE == 1
	                          size_of_PRECISION_real)
#endif
#if COMPLEXCASE == 1
                                  size_of_PRECISION_complex)
#endif
        if (.not.(successCUDA)) then
          print *,"bandred_&
	  &MATH_DATATYPE&
	  &: error in cudaMalloc"
          stop
        endif

        successCUDA = cuda_malloc(vav_dev, nbw*nbw*   &
#if REALCASE == 1
	                          size_of_PRECISION_real)
#endif
#if COMPLEXCASE == 1
                                  size_of_PRECISION_complex)
#endif
        if (.not.(successCUDA)) then
          print *,"bandred_&
	  &MATH_DATATYPE&
	  &: error in cudaMalloc"
          stop
        endif
      endif ! useGPU

      ! Matrix is split into tiles; work is done only for tiles on the diagonal or above

      tile_size = nblk*least_common_multiple(np_rows,np_cols) ! minimum global tile size
      tile_size = ((128*max(np_rows,np_cols)-1)/tile_size+1)*tile_size ! make local tiles at least 128 wide

      l_rows_tile = tile_size/np_rows ! local rows of a tile
      l_cols_tile = tile_size/np_cols ! local cols of a tile

#if REALCASE == 1
      if (useQR) then

        if (useGPU) then
          print *,"qr decomposition at the moment not supported with GPU"
          stop
        endif

        if (which_qr_decomposition == 1) then
          call qr_pqrparam_init(pqrparam(1:11),    nblk,'M',0,   nblk,'M',0,   nblk,'M',1,'s')
          allocate(tauvector(na), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when allocating tauvector "//errorMessage
            stop
          endif

          allocate(blockheuristic(nblk), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when allocating blockheuristic "//errorMessage
            stop
          endif

          l_rows = local_index(na, my_prow, np_rows, nblk, -1)
          allocate(vmrCPU(max(l_rows,1),na), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when allocating vmrCPU "//errorMessage
            stop
          endif

          vmrCols = na

#ifdef USE_ASSUMED_SIZE_QR
          call qr_pdgeqrf_2dcomm_&
	  &PRECISION&
	  &(a, lda, matrixCols, vmrCPU, max(l_rows,1), vmrCols, tauvector(1), na, tmat(1,1,1), &
                                 nbw, nbw, dwork_size, 1, -1, na, nbw, nblk, nblk, na, na, 1, 0, PQRPARAM(1:11), &
                                 mpi_comm_rows, mpi_comm_cols, blockheuristic)

#else
          call qr_pdgeqrf_2dcomm_&
	  &PRECISION&
	  &(a(1:lda,1:matrixCols), matrixCols, lda, vmrCPU(1:max(l_rows,1),1:vmrCols), max(l_rows,1), &
                                 vmrCols, tauvector(1:na), na, tmat(1:nbw,1:nbw,1), nbw, &
                                 nbw, dwork_size(1:1), 1, -1, na, nbw, nblk, nblk, na, na, 1, 0, PQRPARAM(1:11), &
                                 mpi_comm_rows, mpi_comm_cols, blockheuristic)
#endif

          work_size = dwork_size(1)
          allocate(work_blocked(work_size), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when allocating work_blocked "//errorMessage
            stop
          endif
          work_blocked = CONST_0_0
          deallocate(vmrCPU, stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when deallocating vmrCPU "//errorMessage
            stop
          endif

        endif ! which_qr_decomposition

      endif ! useQr
#endif /* REALCASE */

      if (useGPU) then

        cur_l_rows = 0
        cur_l_cols = 0

        successCUDA = cuda_memcpy(a_dev, loc(a(1,1)), (lda)*(na_cols)*   &
#if REALCASE == 1
	                          size_of_PRECISION_real,    &
#endif
#if COMPLEXCASE == 1
                                  size_of_PRECISION_complex, &
#endif
				  cudaMemcpyHostToDevice)
        if (.not.(successCUDA)) then
          print *,"bandred_&
	  &MATH_DATATYPE&
	  &: error in cudaMemcpy"
          stop
        endif
      endif ! useGPU


      do istep = (na-1)/nbw, 1, -1

        n_cols = MIN(na,(istep+1)*nbw) - istep*nbw ! Number of columns in current step

        ! Number of local columns/rows of remaining matrix
        l_cols = local_index(istep*nbw, my_pcol, np_cols, nblk, -1)
        l_rows = local_index(istep*nbw, my_prow, np_rows, nblk, -1)

        ! Allocate vmr and umc to their exact sizes so that they can be used in bcasts and reduces

        if (useGPU) then
          cur_l_rows = max(l_rows, 1)
          cur_l_cols = max(l_cols, 1)

          vmr_size = cur_l_rows * 2 * n_cols
          umc_size = cur_l_cols * 2 * n_cols

          ! Allocate vmr and umc only if the inew size exceeds their current capacity
          ! Added for FORTRAN CALLS
          if ((.not. allocated(vr)) .or. (l_rows + 1 .gt. ubound(vr, dim=1))) then
            if (allocated(vr)) then
              deallocate(vr, stat=istat, errmsg=errorMessage)
              if (istat .ne. 0) then
                print *,"bandred_&
		&MATH_DATATYPE&
		&: error when deallocating vr "//errorMessage
                stop
              endif
            endif
            allocate(vr(l_rows + 1), stat=istat, errmsg=errorMessage)
            if (istat .ne. 0) then
              print *,"bandred_&
	      &MATH_DATATYPE&
	      &: error when allocating vr "//errorMessage
              stop
            endif

          endif

          if ((.not. allocated(vmrCUDA)) .or. (vmr_size .gt. ubound(vmrCUDA, dim=1))) then
            if (allocated(vmrCUDA)) then
              deallocate(vmrCUDA, stat=istat, errmsg=errorMessage)
              if (istat .ne. 0) then
                print *,"bandred_&
		&MATH_DATATYPE&
		&: error when allocating vmrCUDA "//errorMessage
                stop
              endif

              successCUDA = cuda_free(vmr_dev)
              if (.not.(successCUDA)) then
                print *,"bandred_&
		&MATH_DATATYPE&: error in cuda_free"
                stop
              endif
            endif

#if REALCASE == 1
            allocate(vmrCUDA(vmr_size), stat=istat, errmsg=errorMessage)
#endif
#if COMPLEXCASE == 1
            allocate(vmrCUDA(max(l_rows,1),2*n_cols), stat=istat, errmsg=errorMessage)
#endif
            if (istat .ne. 0) then
              print *,"bandred_&
	      &MATH_DATATYPE&
	      &: error when allocating vmrCUDA "//errorMessage
              stop
            endif
            successCUDA = cuda_malloc(vmr_dev, vmr_size*    &
#if REALCASE == 1
	                              size_of_PRECISION_real)
#endif
#if COMPLEXCASE == 1
                                      size_of_PRECISION_complex)
#endif
            if (.not.(successCUDA)) then
              print *,"bandred_&
	      &MATH_DATATYPE&
	      &: error in cudaMalloc: vmr_dev"
              stop
            endif

          endif

          if ((.not. allocated(umcCUDA)) .or. (umc_size .gt. ubound(umcCUDA, dim=1))) then
            if (allocated(umcCUDA)) then
              deallocate(umcCUDA, stat=istat, errmsg=errorMessage)
              if (istat .ne. 0) then
                print *,"bandred_&
		&MATH_DATATYPE&
		&: error when deallocating umcCUDA "//errorMessage
                stop
              endif

              successCUDA = cuda_free(umc_dev)
              if (.not.(successCUDA)) then
                 print *,"bandred_&
		 &MATH_DATATYPE&
		 &: error in cudaFree umc_dev"
                 stop
              endif

            endif

#if REALCASE == 1
            allocate(umcCUDA(umc_size), stat=istat, errmsg=errorMessage)
#endif
#if COMPLEXCASE == 1
            allocate(umcCUDA(max(l_cols,1),2*n_cols), stat=istat, errmsg=errorMessage)
#endif
            if (istat .ne. 0) then
              print *,"bandred_&
	      &MATH_DATATYPE&
	      &: error when deallocating umcCUDA "//errorMessage
              stop
            endif

            successCUDA = cuda_malloc(umc_dev, umc_size*     &
#if REALCASE == 1
	                              size_of_PRECISION_real)
#endif
#if COMPLEXCASE == 1
                                      size_of_PRECISION_complex)
#endif
            if (.not.(successCUDA)) then
              print *,"bandred_&
	      &MATH_DATATYPE&
	      &: error in cudaMalloc umc_dev"
              stop
            endif

          endif

        else ! GPU not used

          ! unify the the name vmr and vmrCPU, as well as vmrGPU
          ! the same for umcCPU and umcGPU
          ! Allocate vmr and umcCPU to their exact sizes so that they can be used in bcasts and reduces

          allocate(vmrCPU(max(l_rows,1),2*n_cols), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_&
	    &MATH_DATATYPE&
	    &: error when allocating vmrCPU "//errorMessage
            stop
          endif

          allocate(umcCPU(max(l_cols,1),2*n_cols), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_&
	    &MATH_DATATYPE&
	    &: error when allocating umcCPU "//errorMessage
            stop
          endif

          allocate(vr(l_rows+1), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_&
	    &MATH_DATATYPE&
	    &: error when allocating vr "//errorMessage
            stop
          endif

        endif ! use GPU

        if (useGPU) then
#if REALCASE == 1
          vmrCUDA(1 : cur_l_rows * n_cols) = CONST_0_0
#endif
#if COMPLEXCASE == 1
          vmrCUDA(1:l_rows,1:n_cols) = CONST_COMPLEX_0_0
#endif
        else
#if REALCASE == 1
          vmrCPU(1:l_rows,1:n_cols) = CONST_0_0
#endif
#if COMPLEXCASE == 1
          vmrCPU(1:l_rows,1:n_cols) = CONST_COMPLEX_0_0
#endif
        endif ! useGPU

#if REALCASE == 1
        vr(:) = CONST_0_0
        tmat(:,:,istep) = CONST_0_0
#endif
#if COMPLEXCASE == 1
        vr(:) = CONST_COMPLEX_0_0
        tmat(:,:,istep) = CONST_COMPLEX_0_0
#endif
        if (useGPU) then
#if REALCASE == 1
          umcCUDA(1 : umc_size) = CONST_0_0
#endif
          lc_start = local_index(istep*nbw+1, my_pcol, np_cols, nblk, -1)
          lc_end   = local_index(istep*nbw+n_cols, my_pcol, np_cols, nblk, -1)
          lr_end   = local_index((istep-1)*nbw + n_cols, my_prow, np_rows, nblk, -1)

          if (lc_start .le. 0) lc_start = 1

          ! Here we assume that the processor grid and the block grid are aligned
          cur_pcol = pcol(istep*nbw+1, nblk, np_cols)

          if (my_pcol == cur_pcol) then
            successCUDA = cuda_memcpy2d(loc(a(1, lc_start)), &
#if REALCASE == 1
                                            lda*size_of_PRECISION_real,         &
#endif
#if COMPLEXCASE == 1
                                        int(lda*size_of_PRECISION_complex,kind=c_size_t), &
#endif
#if REALCASE == 1
                                        (a_dev + ((lc_start-1) * lda*size_of_PRECISION_real)),    &
#endif
#if COMPLEXCASE == 1
                                        (a_dev + int( ( (lc_start-1) * lda*size_of_PRECISION_complex),kind=c_size_t )),      &
#endif
#if REALCASE == 1
                                        lda*size_of_PRECISION_real, lr_end*size_of_PRECISION_real, &
#endif
#if COMPLEXCASE == 1
                                         int(lda*size_of_PRECISION_complex,kind=c_size_t),              &
                                         int(lr_end*size_of_PRECISION_complex,kind=c_size_t),           &
#endif
#if REALCASE == 1
                                         (lc_end - lc_start+1), cudaMemcpyDeviceToHost)
#endif
#if COMPLEXCASE == 1
                                         int((lc_end - lc_start+1),kind=c_size_t),int(cudaMemcpyDeviceToHost,kind=c_int))
#endif

            if (.not.(successCUDA)) then
              print *,"bandred_&
	      &MATH_DATATYPE&
	      &: error in cudaMemcpy2d"
              stop
            endif

          endif
        endif ! useGPU

        ! Reduce current block to lower triangular form
#if REALCASE == 1
        if (useQR) then
          if (which_qr_decomposition == 1) then
            vmrCols = 2*n_cols
#ifdef USE_ASSUMED_SIZE_QR
            call qr_pdgeqrf_2dcomm_&
	    &PRECISION&
	    &(a, lda, matrixCols, vmrCPU, max(l_rows,1), vmrCols, tauvector(1), &
                                   na, tmat(1,1,istep), nbw, nbw, work_blocked, work_size,        &
                                     work_size, na, n_cols, nblk, nblk,        &
                                     istep*nbw+n_cols-nbw, istep*nbw+n_cols, 1,&
                                     0, PQRPARAM(1:11), mpi_comm_rows, mpi_comm_cols,&
                                     blockheuristic)

#else
            call qr_pdgeqrf_2dcomm_&
	    &PRECISION&
	    &(a(1:lda,1:matrixCols), lda, matrixCols, vmrCPU(1:max(l_rows,1),1:vmrCols) ,   &
                                    max(l_rows,1), vmrCols, tauvector(1:na), na, &
                                     tmat(1:nbw,1:nbw,istep), nbw, nbw, work_blocked(1:work_size), work_size, &
                                     work_size, na, n_cols, nblk, nblk,        &
                                     istep*nbw+n_cols-nbw, istep*nbw+n_cols, 1,&
                                     0, PQRPARAM(1:11), mpi_comm_rows, mpi_comm_cols,&
                                     blockheuristic)
#endif
          endif

       else !useQR
#endif /* REALCASE == 1 */
         do lc = n_cols, 1, -1

           ncol = istep*nbw + lc ! absolute column number of householder vector
           nrow = ncol - nbw ! Absolute number of pivot row

           lr  = local_index(nrow, my_prow, np_rows, nblk, -1) ! current row length
           lch = local_index(ncol, my_pcol, np_cols, nblk, -1) ! HV local column number

           tau = 0

           if (nrow == 1) exit ! Nothing to do

           cur_pcol = pcol(ncol, nblk, np_cols) ! Processor column owning current block

           if (my_pcol==cur_pcol) then

             ! Get vector to be transformed; distribute last element and norm of
             ! remaining elements to all procs in current column

             vr(1:lr) = a(1:lr,lch) ! vector to be transformed

             if (my_prow==prow(nrow, nblk, np_rows)) then
               aux1(1) = dot_product(vr(1:lr-1),vr(1:lr-1))
               aux1(2) = vr(lr)
             else
               aux1(1) = dot_product(vr(1:lr),vr(1:lr))
#if REALCASE == 1
               aux1(2) = CONST_0_0
#endif
#if COMPLEXCASE == 1
               aux1(2) = CONST_COMPLEX_0_0
#endif
             endif

#ifdef WITH_MPI
             call timer%start("mpi_communication")
             call mpi_allreduce(aux1, aux2, 2, &
#if REALCASE == 1
                                MPI_REAL_PRECISION, &
#endif
#if COMPLEXCASE == 1
                                MPI_COMPLEX_PRECISION, &
#endif
                                MPI_SUM, mpi_comm_rows, mpierr)
             call timer%stop("mpi_communication")

#else /* WITH_MPI */
              aux2 = aux1 ! this should be optimized
#endif

             vnorm2 = aux2(1)
             vrl    = aux2(2)

             ! Householder transformation
#if REALCASE == 1
	     call hh_transform_real_&
#endif
#if COMPLEXCASE == 1
	     call hh_transform_complex_&
#endif
             &PRECISION &
                              (vrl, vnorm2, xf, tau)
             ! Scale vr and store Householder vector for back transformation

             vr(1:lr) = vr(1:lr) * xf
             if (my_prow==prow(nrow, nblk, np_rows)) then
               a(1:lr-1,lch) = vr(1:lr-1)
               a(lr,lch) = vrl
#if REALCASE == 1
               vr(lr) = CONST_1_0
#endif
#if COMPLEXCASE == 1
               vr(lr) = CONST_COMPLEX_1_0
#endif
             else
               a(1:lr,lch) = vr(1:lr)
             endif

           endif

           ! Broadcast Householder vector and tau along columns

           vr(lr+1) = tau
#ifdef WITH_MPI
           call timer%start("mpi_communication")
	   call MPI_Bcast(vr, lr+1, &
#if REALCASE == 1
                          MPI_REAL_PRECISION, &
#endif
#if COMPLEXCASE == 1
                          MPI_COMPLEX_PRECISION, &
#endif
                          cur_pcol, mpi_comm_cols, mpierr)
           call timer%stop("mpi_communication")

#endif /* WITH_MPI */

           if (useGPU) then
#if REALCASE == 1
             vmrCUDA(cur_l_rows * (lc - 1) + 1 : cur_l_rows * (lc - 1) + lr) = vr(1:lr)
#endif
#if COMPLEXCASE == 1
             vmrCUDA(1:lr,lc) = vr(1:lr)
#endif
           else
             vmrCPU(1:lr,lc) = vr(1:lr)
           endif
           tau = vr(lr+1)

#if REALCASE == 1
           tmat(lc,lc,istep) = tau ! Store tau in diagonal of tmat
#endif
#if COMPLEXCASE == 1
           tmat(lc,lc,istep) = conjg(tau) ! Store tau in diagonal of tmat
#endif
           ! Transform remaining columns in current block with Householder vector
           ! Local dot product

#if REALCASE == 1
           aux1 = 0
#endif
#if COMPLEXCASE == 1
          aux1 = CONST_COMPLEX_0_0
#endif

#ifdef WITH_OPENMP

#if REALCASE == 1
           !Open up one omp region to avoid paying openmp overhead.
           !This does not help performance due to the addition of two openmp barriers around the MPI call,
           !But in the future this may be beneficial if these barriers are replaced with a faster implementation

           !$omp parallel private(mynlc, j, lcx, ii, pp ) shared(aux1)
           mynlc = 0 ! number of local columns

           !This loop does not have independent iterations,
           !'mynlc' is incremented each iteration, and it is difficult to remove this dependency
           !Thus each thread executes every iteration of the loop, except it only does the work if it 'owns' that iteration
           !That is, a thread only executes the work associated with an iteration if its thread id is congruent to
           !the iteration number modulo the number of threads
           do j=1,lc-1
             lcx = local_index(istep*nbw+j, my_pcol, np_cols, nblk, 0)
             if (lcx>0 ) then
               mynlc = mynlc+1
               if ( mod((j-1), omp_get_num_threads()) .eq. omp_get_thread_num() ) then
                   if (lr>0) aux1(mynlc) = dot_product(vr(1:lr),a(1:lr,lcx))
               endif
             endif
           enddo

           ! Get global dot products

           !$omp barrier
           !$omp single
#ifdef WITH_MPI
           call timer%start("mpi_communication")
           if (mynlc>0) call mpi_allreduce(aux1, aux2, mynlc, MPI_REAL_PRECISION, MPI_SUM, mpi_comm_rows, mpierr)
           call timer%stop("mpi_communication")
#else /* WITH_MPI */
           if (mynlc>0) aux2 = aux1
#endif /* WITH_MPI */
           !$omp end single
           !$omp barrier

           ! Transform
           transformChunkSize=32
           mynlc = 0
           do j=1,lc-1
             lcx = local_index(istep*nbw+j, my_pcol, np_cols, nblk, 0)
             if (lcx>0) then
               mynlc = mynlc+1
               !This loop could be parallelized with an openmp pragma with static scheduling and chunk size 32
               !However, for some reason this is slower than doing it manually, so it is parallelized as below.
               do ii=omp_get_thread_num()*transformChunkSize,lr,omp_get_num_threads()*transformChunkSize
                  do pp = 1,transformChunkSize
                      if (pp + ii > lr) exit
                          a(ii+pp,lcx) = a(ii+pp,lcx) - tau*aux2(mynlc)*vr(ii+pp)
                  enddo
               enddo
             endif
           enddo
           !$omp end parallel
#endif /* REALCASE == 1 */

#if COMPLEXCASE == 1
           nlc = 0 ! number of local columns
           do j=1,lc-1
             lcx = local_index(istep*nbw+j, my_pcol, np_cols, nblk, 0)
             if (lcx>0) then
               nlc = nlc+1
               aux1(nlc) = dot_product(vr(1:lr),a(1:lr,lcx))
             endif
           enddo

           ! Get global dot products
#ifdef WITH_MPI
           call timer%start("mpi_communication")
           if (nlc>0) call mpi_allreduce(aux1, aux2, nlc, MPI_COMPLEX_PRECISION, MPI_SUM, mpi_comm_rows, mpierr)

           ! Transform

           nlc = 0
           do j=1,lc-1
             lcx = local_index(istep*nbw+j, my_pcol, np_cols, nblk, 0)
             if (lcx>0) then
               nlc = nlc+1
               a(1:lr,lcx) = a(1:lr,lcx) - conjg(tau)*aux2(nlc)*vr(1:lr)

             endif
           enddo

           call timer%stop("mpi_communication")

#else /* WITH_MPI */
!          if (nlc>0) aux2=aux1

           ! Transform

           nlc = 0
           do j=1,lc-1
             lcx = local_index(istep*nbw+j, my_pcol, np_cols, nblk, 0)
             if (lcx>0) then
               nlc = nlc+1
               a(1:lr,lcx) = a(1:lr,lcx) - conjg(tau)*aux1(nlc)*vr(1:lr)
             endif
           enddo

#endif /* WITH_MPI */
!
!           ! Transform
!
!           nlc = 0
!           do j=1,lc-1
!             lcx = local_index(istep*nbw+j, my_pcol, np_cols, nblk, 0)
!             if (lcx>0) then
!               nlc = nlc+1
!               a(1:lr,lcx) = a(1:lr,lcx) - conjg(tau)*aux2(nlc)*vr(1:lr)
!             endif
!           enddo

#endif /* COMPLEXCASE */

#else /* WITH_OPENMP */

           nlc = 0 ! number of local columns
           do j=1,lc-1
             lcx = local_index(istep*nbw+j, my_pcol, np_cols, nblk, 0)
             if (lcx>0) then
               nlc = nlc+1
               if (lr>0) aux1(nlc) = dot_product(vr(1:lr),a(1:lr,lcx))
             endif
           enddo

           ! Get global dot products
#ifdef WITH_MPI
           call timer%start("mpi_communication")
           if (nlc>0) call mpi_allreduce(aux1, aux2, nlc,      &
#if REALCASE == 1
	                                 MPI_REAL_PRECISION,   &
#endif
#if COMPLEXCASE == 1
                                         MPI_COMPLEX_PRECISION,&
#endif
					 MPI_SUM, mpi_comm_rows, mpierr)
           call timer%stop("mpi_communication")
#else /* WITH_MPI */
           if (nlc>0) aux2=aux1
#endif /* WITH_MPI */
           ! Transform

           nlc = 0
           do j=1,lc-1
             lcx = local_index(istep*nbw+j, my_pcol, np_cols, nblk, 0)
             if (lcx>0) then
               nlc = nlc+1
#if REALCASE == 1
               a(1:lr,lcx) = a(1:lr,lcx) - tau*aux2(nlc)*vr(1:lr)
#endif
#if COMPLEXCASE == 1
               a(1:lr,lcx) = a(1:lr,lcx) - conjg(tau)*aux2(nlc)*vr(1:lr)
#endif
             endif
           enddo
#endif /* WITH_OPENMP */
         enddo ! lc

         if (useGPU) then
           ! store column tiles back to GPU
           cur_pcol = pcol(istep*nbw+1, nblk, np_cols)
           if (my_pcol == cur_pcol) then
             successCUDA = cuda_memcpy2d((a_dev+        &
#if REALCASE == 1
                                         ((lc_start-1)*lda*size_of_PRECISION_real)),          &
#endif
#if COMPLEXCASE == 1
                                         int(((lc_start-1)*lda*size_of_PRECISION_complex),kind=c_size_t)),    &
#endif
#if REALCASE == 1
                                         lda*size_of_PRECISION_real, loc(a(1, lc_start)),           &
#endif
#if COMPLEXCASE == 1
                                         int(lda*size_of_PRECISION_complex,kind=c_size_t), loc(a(1,lc_start)), &
#endif
#if REALCASE == 1
                                         lda*size_of_PRECISION_real,  lr_end*size_of_PRECISION_real, &
#endif
#if COMPLEXCASE == 1
                                         int(lda*size_of_PRECISION_complex,kind=c_size_t),           &
                                         int(lr_end*size_of_PRECISION_complex,kind=c_size_t),        &
#endif
#if REALCASE == 1
                                         (lc_end - lc_start+1),cudaMemcpyHostToDevice)
#endif
#if COMPLEXCASE == 1
                                         int((lc_end - lc_start+1),kind=c_size_t), &
                                         int(cudaMemcpyHostToDevice,kind=c_int))
#endif
             if (.not.(successCUDA)) then
               print *, "bandred_&
	       &MATH_DATATYPE&
	       &: cuda memcpy a_dev  failed ", istat
               stop
             endif
           endif
         endif

         ! Calculate scalar products of stored Householder vectors.
         ! This can be done in different ways, we use dsyrk

         vav = 0
	 call timer%start("blas")
         if (useGPU) then
           if (l_rows>0) &
#if REALCASE == 1
             call PRECISION_SYRK('U', 'T',            &
#endif
#if COMPLEXCASE == 1
             call PRECISION_HERK('U', 'C',            &
#endif
	                         n_cols, l_rows, ONE, &
#if REALCASE == 1
				 vmrCUDA, cur_l_rows, &
#endif
#if COMPLEXCASE == 1
                                 vmrCUDA, ubound(vmrCUDA,dim=1), &
#endif
				 ZERO, vav, ubound(vav,dim=1))
         else ! useGPU
           if (l_rows>0) &
#if REALCASE == 1
             call PRECISION_SYRK('U', 'T',           &
#endif
#if COMPLEXCASE == 1
             call PRECISION_HERK('U', 'C',           &
#endif
	                         n_cols, l_rows, ONE, vmrCPU, ubound(vmrCPU,dim=1), ZERO, vav, ubound(vav,dim=1))
         endif
	 call timer%stop("blas")
#if REALCASE == 1
	 call symm_matrix_allreduce_&
#endif
#if COMPLEXCASE == 1
         call herm_matrix_allreduce_&
#endif
         &PRECISION &
                         (n_cols,vav, nbw, nbw,mpi_comm_rows)
         ! Calculate triangular matrix T for block Householder Transformation
	 call timer%start("blas")
         do lc=n_cols,1,-1
           tau = tmat(lc,lc,istep)
           if (lc<n_cols) then
#if REALCASE == 1
             call PRECISION_TRMV('U', 'T', 'N',          &
#endif
#if COMPLEXCASE == 1
             call PRECISION_TRMV('U', 'C', 'N',          &
#endif
                                 n_cols-lc, tmat(lc+1,lc+1,istep), ubound(tmat,dim=1), vav(lc+1,lc), 1)

#if REALCASE == 1
             tmat(lc,lc+1:n_cols,istep) = -tau * vav(lc+1:n_cols,lc)
#endif
#if COMPLEXCASE == 1
             tmat(lc,lc+1:n_cols,istep) = -tau * conjg(vav(lc+1:n_cols,lc))
#endif
           endif
         enddo
 	 call timer%stop("blas")
#if REALCASE == 1
       endif !useQR
#endif
       ! Transpose vmr -> vmc (stored in umc, second half)
       if (useGPU) then
         call elpa_transpose_vectors_&
              &MATH_DATATYPE&
              &_&
              &PRECISION &
#if REALCASE == 1
	                                   (vmrCUDA, cur_l_rows,  &
#endif
#if COMPLEXCASE == 1
                                           (vmrCUDA, ubound(vmrCUDA,dim=1), &
#endif
					    mpi_comm_rows, &
#if REALCASE == 1
                                            umcCUDA(cur_l_cols * n_cols + 1), cur_l_cols, &
#endif
#if COMPLEXCASE == 1
                                            umcCUDA(1,n_cols+1), ubound(umcCUDA,dim=1),     &
#endif
					    mpi_comm_cols, 1, istep*nbw, n_cols, nblk)
       else ! useGPU
         call elpa_transpose_vectors_&
              &MATH_DATATYPE&
              &_&
              &PRECISION &
                                           (vmrCPU, ubound(vmrCPU,dim=1), mpi_comm_rows, &
                                            umcCPU(1,n_cols+1), ubound(umcCPU,dim=1), mpi_comm_cols, &
                                            1, istep*nbw, n_cols, nblk)
       endif

       ! Calculate umc = A**T * vmr
       ! Note that the distributed A has to be transposed
       ! Opposed to direct tridiagonalization there is no need to use the cache locality
       ! of the tiles, so we can use strips of the matrix
#if REALCASE == 1
       ! here the GPU version and CPU version diverged substantially, due to the newest
       ! optimizations due to Intel. The GPU version has to be re-written
       if (useGPU) then
         umcCUDA(1 : l_cols * n_cols) = CONST_0_0
         vmrCUDA(cur_l_rows * n_cols + 1 : cur_l_rows * n_cols * 2) = CONST_0_0

         if (l_cols>0 .and. l_rows>0) then
           successCUDA = cuda_memcpy(vmr_dev, loc(vmrCUDA(1)), vmr_size*size_of_PRECISION_real,cudaMemcpyHostToDevice)
           if (.not.(successCUDA)) then
             print *,"bandred_real: error in cudaMemcpy"
             stop
           endif
           successCUDA = cuda_memcpy(umc_dev, loc(umcCUDA(1)), umc_size*size_of_PRECISION_real,cudaMemcpyHostToDevice)
           if (.not.(successCUDA)) then
             print *,"bandred_real: error in cudaMemcpy"
             stop
           endif

           do i=0,(istep*nbw-1)/tile_size

             lcs = i*l_cols_tile+1
             lce = min(l_cols,(i+1)*l_cols_tile)
             if (lce<lcs) cycle
             call timer%start("cublas")
             lre = min(l_rows,(i+1)*l_rows_tile)
             call cublas_PRECISION_GEMM('T', 'N', lce-lcs+1, n_cols, lre, &
                               CONST_1_0, (a_dev + ((lcs-1)*lda*size_of_PRECISION_real)), lda, vmr_dev,cur_l_rows, &
                               CONST_1_0, (umc_dev+ (lcs-1)*size_of_PRECISION_real), cur_l_cols)
             if(i==0) cycle
             lre = min(l_rows,i*l_rows_tile)
             call cublas_PRECISION_GEMM('N', 'N', lre,n_cols, lce-lcs+1,&
                               CONST_1_0, (a_dev+ ((lcs-1)*lda*size_of_PRECISION_real)), lda,                  &
                               (umc_dev+(cur_l_cols * n_cols+lcs-1)*size_of_PRECISION_real), cur_l_cols, &
                               CONST_1_0, (vmr_dev+(cur_l_rows * n_cols)*size_of_PRECISION_real), cur_l_rows)
             call timer%stop("cublas")
           enddo
           successCUDA = cuda_memcpy(loc(vmrCUDA(1)), vmr_dev,vmr_size*size_of_PRECISION_real,cudaMemcpyDeviceToHost)
           if (.not.(successCUDA)) then
             print *,"bandred_real: error in cudaMemcpy"
             stop
           endif
           successCUDA = cuda_memcpy(loc(umcCUDA(1)), umc_dev, umc_size*size_of_PRECISION_real,cudaMemcpyDeviceToHost)
           if (.not.(successCUDA)) then
             print *,"bandred_real: error in cudaMemcpy"
             stop
           endif

         endif ! l_cols>0 .and. l_rows>0

       else ! do not useGPU version

         !Code for Algorithm 4

         n_way = 1
#ifdef WITH_OPENMP
         n_way = omp_get_max_threads()
#endif
         !umcCPU(1:l_cols,1:n_cols) = 0.d0
         !vmrCPU(1:l_rows,n_cols+1:2*n_cols) = 0
#ifdef WITH_OPENMP
         !$omp parallel private( i,lcs,lce,lrs,lre)
#endif
         if (n_way > 1) then
           !$omp do
           do i=1,min(l_cols_tile, l_cols)
             umcCPU(i,1:n_cols) = CONST_0_0
           enddo

           !$omp do
           do i=1,l_rows
             vmrCPU(i,n_cols+1:2*n_cols) = CONST_0_0
           enddo
           if (l_cols>0 .and. l_rows>0) then

             !SYMM variant 4
             !Partitioned Matrix Expression:
             ! Ct = Atl Bt + Atr Bb
             ! Cb = Atr' Bt + Abl Bb
             !
             !Loop invariant:
             ! Ct = Atl Bt + Atr Bb
             !
             !Update:
             ! C1 = A10'B0 + A11B1 + A21 B2
             !
             !This algorithm chosen because in this algoirhtm, the loop around the dgemm calls
             !is easily parallelized, and regardless of choise of algorithm,
             !the startup cost for parallelizing the dgemms inside the loop is too great

             !$omp do schedule(static,1)
             do i=0,(istep*nbw-1)/tile_size
               lcs = i*l_cols_tile+1                   ! local column start
               lce = min(l_cols, (i+1)*l_cols_tile)    ! local column end

               lrs = i*l_rows_tile+1                   ! local row start
               lre = min(l_rows, (i+1)*l_rows_tile)    ! local row end

               !C1 += [A11 A12] [B1
               !                 B2]
               if ( lre > lrs .and. l_cols > lcs ) then
	         call timer%start("blas")
                 call PRECISION_GEMM('N', 'N', lre-lrs+1, n_cols, l_cols-lcs+1,          &
                            CONST_1_0, a(lrs,lcs), ubound(a,dim=1),                 &
                                  umcCPU(lcs,n_cols+1), ubound(umcCPU,dim=1),  &
                            CONST_0_0, vmrCPU(lrs,n_cols+1), ubound(vmrCPU,dim=1))
	         call timer%stop("blas")
               endif

               ! C1 += A10' B0
               if ( lce > lcs .and. i > 0 ) then
	       	 call timer%start("blas")
                 call PRECISION_GEMM('T', 'N', lce-lcs+1, n_cols, lrs-1,           &
                            CONST_1_0, a(1,lcs),   ubound(a,dim=1),           &
                                  vmrCPU(1,1),   ubound(vmrCPU,dim=1),   &
                            CONST_0_0, umcCPU(lcs,1), ubound(umcCPU,dim=1))
	       	 call timer%stop("blas")
               endif
             enddo
           endif ! l_cols>0 .and. l_rows>0
         else ! n_way > 1
           umcCPU(1:l_cols,1:n_cols) = CONST_0_0
           vmrCPU(1:l_rows,n_cols+1:2*n_cols) = CONST_0_0
           if (l_cols>0 .and. l_rows>0) then
             do i=0,(istep*nbw-1)/tile_size

               lcs = i*l_cols_tile+1
               lce = min(l_cols,(i+1)*l_cols_tile)
               if (lce<lcs) cycle
	       call timer%start("blas")
               lre = min(l_rows,(i+1)*l_rows_tile)
               call PRECISION_GEMM('T', 'N', lce-lcs+1, n_cols, lre, CONST_1_0, a(1,lcs), ubound(a,dim=1), &
                            vmrCPU, ubound(vmrCPU,dim=1), CONST_1_0, umcCPU(lcs,1), ubound(umcCPU,dim=1))
	       call timer%stop("blas")
               if (i==0) cycle
                 lre = min(l_rows,i*l_rows_tile)
	       	 call timer%start("blas")
                 call PRECISION_GEMM('N', 'N', lre, n_cols, lce-lcs+1, CONST_1_0, a(1,lcs), lda, &
                            umcCPU(lcs,n_cols+1), ubound(umcCPU,dim=1), CONST_1_0, vmrCPU(1,n_cols+1), ubound(vmrCPU,dim=1))
	       	 call timer%stop("blas")
             enddo
           endif
         endif ! n_way > 1
#ifdef WITH_OPENMP
        !$omp end parallel
#endif
       endif ! do not useGPU version
#endif /* REALCASE == 1 */

#if COMPLEXCASE == 1
        if (useGPU) then
          umcCUDA(1:l_cols,1:n_cols) = CONST_COMPLEX_0_0
          vmrCUDA(1:l_rows,n_cols+1:2*n_cols) = CONST_COMPLEX_0_0
	else
          umcCPU(1:l_cols,1:n_cols) = CONST_COMPLEX_0_0
          vmrCPU(1:l_rows,n_cols+1:2*n_cols) = CONST_COMPLEX_0_0
	endif

        if (l_cols>0 .and. l_rows>0) then
          if (useGPU) then
           ! if (size(vmrCPU,dim=1)*size(vmrCPU,dim=2) .gt. vmr_size) then
           !   print *,"bandred_complex: vmr size 2 :",size(vmrCPU,dim=1)*size(vmrCPU,dim=2),vmr_size
           !   stop
           ! endif
            successCUDA = cuda_memcpy(vmr_dev, loc(vmrCUDA(1,1)),vmr_size*size_of_PRECISION_complex,cudaMemcpyHostToDevice)

            if (.not.(successCUDA)) then
              print *, "bandred_complex:  cuda memcpy vmr_dev failed ", istat
              stop
            endif
            !if (size(umcCPU,dim=1)*size(umcCPU,dim=2) .gt. umc_size) then
            !  print *,"bandred_complex: umc size 2 :",size(umcCPU,dim=1)*size(umcCPU,dim=2),umc_size
            !  stop
            !endif
            successCUDA = cuda_memcpy(umc_dev, loc(umcCUDA(1,1)),umc_size*size_of_PRECISION_complex,cudaMemcpyHostToDevice)
            if (.not.(successCUDA)) then
              print *, "bandred_complex:  cuda memcpy umc_dev failed  ", istat
              stop
            endif
          endif
          do i=0,(istep*nbw-1)/tile_size

            lcs = i*l_cols_tile+1
            lce = min(l_cols,(i+1)*l_cols_tile)
            if (lce<lcs) cycle

            lre = min(l_rows,(i+1)*l_rows_tile)

            if (useGPU) then
              call timer%start("cublas")
              call cublas_PRECISION_GEMM('C', 'N', lce-lcs+1, n_cols, lre, ONE, (a_dev + ((lcs-1)*lda* &
                        size_of_PRECISION_complex)), lda, &
                        vmr_dev, cur_l_rows, ONE, (umc_dev +(lcs-1)*size_of_PRECISION_complex), cur_l_cols)
              call timer%stop("cublas")
            else
              call timer%start("blas")
              call PRECISION_GEMM('C', 'N', lce-lcs+1, n_cols, lre, ONE, a(1,lcs), ubound(a,dim=1), &
                         vmrCPU, ubound(vmrCPU,dim=1), ONE, umcCPU(lcs,1), ubound(umcCPU,dim=1))
              call timer%stop("blas")
            endif

            if (i==0) cycle
            lre = min(l_rows,i*l_rows_tile)
            if (useGPU) then
              call timer%start("cublas")
              call cublas_PRECISION_GEMM('N', 'N', lre, n_cols, lce-lcs+1, ONE, (a_dev+((lcs-1)*lda* &
                        size_of_PRECISION_complex)),lda,  &
                        (umc_dev+(cur_l_cols * n_cols+lcs-1)*size_of_PRECISION_complex), cur_l_cols,ONE,  &
                        (vmr_dev+(cur_l_rows * n_cols)*size_of_PRECISION_complex), cur_l_rows)
              call timer%stop("cublas")
            else
              call timer%start("blas")
              call PRECISION_GEMM('N', 'N', lre, n_cols, lce-lcs+1, ONE, a(1,lcs), lda, &
                         umcCPU(lcs,n_cols+1), ubound(umcCPU,dim=1), ONE, vmrCPU(1,n_cols+1), ubound(vmrCPU,dim=1))
              call timer%stop("blas")
            endif
          enddo

          if (useGPU) then
!            if (size(vmrCPU,dim=1)*size(vmrCPU,dim=2) .gt. vmr_size) then
!              print *,"bandred_complex: vmr size 3 :",size(vmrCPU,dim=1)*size(vmrCPU,dim=2),vmr_size
!              stop
!            endif
            successCUDA = cuda_memcpy(loc(vmrCUDA(1,1)),vmr_dev,vmr_size*size_of_PRECISION_complex,cudaMemcpyDeviceToHost)
            if (.not.(successCUDA)) then
              print *, "bandred_complex:  cuad memcpy failed vmrCUDA ", istat
              stop
            endif
 !           if (size(umcCPU,dim=1)*size(umcCPU,dim=2) .gt. umc_size) then
 !             print *,"bandred_complex: umc size 3 :",size(umcCPU,dim=1)*size(umcCPU,dim=2),umc_size
 !             stop
 !           endif
            successCUDA = cuda_memcpy(loc(umcCUDA(1,1)), umc_dev,umc_size*size_of_PRECISION_complex,cudaMemcpyDeviceToHost)
            if (.not.(successCUDA)) then
              print *, "bandred_complex:  cuad memcpy failed umcCUDA ", istat
              stop
            endif
          endif ! useGPU
        endif ! (l_cols>0 .and. l_rows>0)
#endif /* COMPLEXCASE == 1 */

       ! Sum up all ur(:) parts along rows and add them to the uc(:) parts
       ! on the processors containing the diagonal
       ! This is only necessary if ur has been calculated, i.e. if the
       ! global tile size is smaller than the global remaining matrix

#if REALCASE == 1
       if (useGPU) then
         ! here the GPU version and CPU version divereged due to the same reasons as above

         if (tile_size < istep*nbw) then
           call elpa_reduce_add_vectors_&
                &MATH_DATATYPE&
                &_&
                &PRECISION &
                                               (vmrCUDA(cur_l_rows * n_cols + 1),cur_l_rows,mpi_comm_rows, &
                                               umcCUDA, cur_l_cols, mpi_comm_cols, &
                                               istep*nbw, n_cols, nblk)
         endif

         if (l_cols>0) then
           allocate(tmpCUDA(l_cols * n_cols), stat=istat, errmsg=errorMessage)
           if (istat .ne. 0) then
             print *,"bandred_real: error when allocating tmpCUDA "//errorMessage
             stop
           endif

#ifdef WITH_MPI
           call timer%start("mpi_communication")

           call mpi_allreduce(umcCUDA, tmpCUDA, l_cols*n_cols, MPI_REAL_PRECISION, MPI_SUM, mpi_comm_rows, ierr)
           umcCUDA(1 : l_cols * n_cols) = tmpCUDA(1 : l_cols * n_cols)
           call timer%stop("mpi_communication")
#else /* WITH_MPI */

!           tmpCUDA(1 : l_cols * n_cols) = umcCUDA(1 : l_cols * n_cols)

#endif /* WITH_MPI */

           if (allocated(tmpCUDA)) then
             deallocate(tmpCUDA, stat=istat, errmsg=errorMessage)
             if (istat .ne. 0) then
               print *,"bandred_real: error when deallocating tmpCUDA "//errorMessage
               stop
             endif
           endif
         endif ! l_cols

         ! U = U * Tmat**T
         successCUDA = cuda_memcpy(umc_dev, loc(umcCUDA(1)), umc_size*size_of_PRECISION_real, cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
         successCUDA = cuda_memcpy(tmat_dev,loc(tmat(1,1,istep)),nbw*nbw*size_of_PRECISION_real,cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
	 call timer%start("cublas")
         call cublas_PRECISION_TRMM('Right', 'Upper', 'Trans', 'Nonunit', l_cols, n_cols, &
                           CONST_1_0, tmat_dev, nbw, umc_dev, cur_l_cols)
	 call timer%start("cublas")

         ! VAV = Tmat * V**T * A * V * Tmat**T = (U*Tmat**T)**T * V * Tmat**T
         successCUDA = cuda_memcpy(vav_dev,loc(vav(1,1)), nbw*nbw*size_of_PRECISION_real,cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
	 call timer%start("cublas")

         call cublas_PRECISION_GEMM('T', 'N', n_cols, n_cols, l_cols, &
                           CONST_1_0, umc_dev, cur_l_cols, (umc_dev+(cur_l_cols * n_cols )*size_of_PRECISION_real),cur_l_cols, &
                           CONST_0_0, vav_dev, nbw)

         call cublas_PRECISION_TRMM('Right', 'Upper', 'Trans', 'Nonunit', n_cols, n_cols, &
                           CONST_1_0, tmat_dev, nbw, vav_dev, nbw)
	 call timer%stop("cublas")

         successCUDA = cuda_memcpy(loc(vav(1,1)), vav_dev, nbw*nbw*size_of_PRECISION_real, cudaMemcpyDeviceToHost)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif

         call symm_matrix_allreduce_&
         &PRECISION &
	                          (n_cols,vav, nbw,nbw,mpi_comm_cols)

         successCUDA = cuda_memcpy(vav_dev, loc(vav(1,1)), nbw*nbw*size_of_PRECISION_real,cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif

         ! U = U - 0.5 * V * VAV
 	 call timer%start("cublas")

         call cublas_PRECISION_GEMM('N', 'N', l_cols, n_cols, n_cols,&
                           -CONST_0_5, (umc_dev+(cur_l_cols * n_cols )*size_of_PRECISION_real),cur_l_cols, vav_dev,nbw,&
                           CONST_1_0, umc_dev, cur_l_cols)
	 call timer%stop("cublas")

         successCUDA = cuda_memcpy(loc(umcCUDA(1)), umc_dev, umc_size*size_of_PRECISION_real, cudaMemcpyDeviceToHost)

         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif

         ! Transpose umc -> umr (stored in vmr, second half)
         call elpa_transpose_vectors_&
         &MATH_DATATYPE&
         &_&
         &PRECISION &
                                           (umcCUDA, cur_l_cols, mpi_comm_cols, &
                                            vmrCUDA(cur_l_rows * n_cols + 1), cur_l_rows, mpi_comm_rows, &
                                            1, istep*nbw, n_cols, nblk)

         successCUDA = cuda_memcpy(vmr_dev, loc(vmrCUDA(1)), vmr_size*size_of_PRECISION_real, cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif

         successCUDA = cuda_memcpy(umc_dev, loc(umcCUDA(1)), umc_size*size_of_PRECISION_real, cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"