elpa2_compute.F90 361 KB
Newer Older
Andreas Marek's avatar
Andreas Marek committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
!    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 Naturwissenschaftrn,
!      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".


#include "config-f90.h"

module ELPA2_compute

! Version 1.1.2, 2011-02-21

69
  use ELPA_utilities
Andreas Marek's avatar
Andreas Marek committed
70
71
72
73
  USE ELPA1_compute
  use elpa1, only : elpa_print_times, time_evp_back, time_evp_fwd, time_evp_solve
  use elpa2_utilities
  use elpa_pdgeqrf
74
  use precision
Andreas Marek's avatar
Andreas Marek committed
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92

  implicit none

  PRIVATE ! By default, all routines contained are private

  public :: bandred_real
  public :: tridiag_band_real
  public :: trans_ev_tridi_to_band_real
  public :: trans_ev_band_to_full_real

  public :: bandred_complex
  public :: tridiag_band_complex
  public :: trans_ev_tridi_to_band_complex
  public :: trans_ev_band_to_full_complex

  public :: band_band_real
  public :: divide_band

93
  integer(kind=ik), public :: which_qr_decomposition = 1     ! defines, which QR-decomposition algorithm will be used
Andreas Marek's avatar
Andreas Marek committed
94
95
96
97
98
99
                                                    ! 0 for unblocked
                                                    ! 1 for blocked (maxrank: nblk)
  include 'mpif.h'

  contains

100
    subroutine bandred_real(na, a, lda, nblk, nbw, matrixCols, numBlocks, mpi_comm_rows, mpi_comm_cols, &
101
                            tmat, wantDebug, useGPU, success, useQR)
Andreas Marek's avatar
Andreas Marek committed
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131

  !-------------------------------------------------------------------------------
  !  bandred_real: 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
  !
  !-------------------------------------------------------------------------------

132
133
      use cuda_functions
      use iso_c_binding
Andreas Marek's avatar
Andreas Marek committed
134
135

#ifdef HAVE_DETAILED_TIMINGS
136
      use timings
Andreas Marek's avatar
Andreas Marek committed
137
#endif
138
139
140
#ifdef WITH_OPENMP
      use omp_lib
#endif
Andreas Marek's avatar
Andreas Marek committed
141
      use precision
142
      implicit none
Andreas Marek's avatar
Andreas Marek committed
143

Andreas Marek's avatar
Andreas Marek committed
144
      integer(kind=ik)           :: na, lda, nblk, nbw, matrixCols, numBlocks, mpi_comm_rows, mpi_comm_cols
Andreas Marek's avatar
Andreas Marek committed
145
146
147
#ifdef DESPERATELY_WANT_ASSUMED_SIZE
      real(kind=rk)              :: a(lda,*), tmat(nbw,nbw,*)
#else
Andreas Marek's avatar
Andreas Marek committed
148
      real(kind=rk)              :: a(lda,matrixCols), tmat(nbw,nbw,numBlocks)
Andreas Marek's avatar
Andreas Marek committed
149
#endif
150
151
      real(kind=rk)              :: eps
      logical, intent(in)        :: useGPU
152

Andreas Marek's avatar
Andreas Marek committed
153
      integer(kind=ik)           :: my_prow, my_pcol, np_rows, np_cols, mpierr
154
      integer(kind=ik)           :: l_cols, l_rows, vmrCols
Andreas Marek's avatar
Andreas Marek committed
155
156
157
      integer(kind=ik)           :: i, j, lcs, lce, lrs, lre, lc, lr, cur_pcol, n_cols, nrow
      integer(kind=ik)           :: istep, ncol, lch, lcx, nlc, mynlc
      integer(kind=ik)           :: tile_size, l_rows_tile, l_cols_tile
158

Andreas Marek's avatar
Andreas Marek committed
159
      real(kind=rk)              :: vnorm2, xf, aux1(nbw), aux2(nbw), vrl, tau, vav(nbw,nbw)
160

161
162
163
      real(kind=rk), allocatable :: tmpCUDA(:),  vmrCUDA(:),  umcCUDA(:)
      real(kind=rk), allocatable :: tmpCPU(:,:), vmrCPU(:,:), umcCPU(:,:)
      real(kind=rk), allocatable :: vr(:)
164
      ! needed for blocked QR decomposition
Andreas Marek's avatar
Andreas Marek committed
165
166
167
      integer(kind=ik)           :: PQRPARAM(11), work_size
      real(kind=rk)              :: dwork_size(1)
      real(kind=rk), allocatable :: work_blocked(:), tauvector(:), blockheuristic(:)
Andreas Marek's avatar
Andreas Marek committed
168

169
170
171
172
173
174
175
      integer(kind=C_intptr_T)   :: a_dev, vmr_dev, umc_dev, tmat_dev, vav_dev
      integer(kind=ik), external :: numroc
      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
      integer(kind=ik)           :: lr_end
      integer(kind=ik)           :: na_rows, na_cols
Andreas Marek's avatar
Andreas Marek committed
176

Andreas Marek's avatar
Andreas Marek committed
177
178
      logical, intent(in)        :: wantDebug
      logical, intent(out)       :: success
179
180
181
      logical                    :: successCUDA
      integer(kind=ik)           :: istat
      character(200)             :: errorMessage
Andreas Marek's avatar
Andreas Marek committed
182

Andreas Marek's avatar
Andreas Marek committed
183
      logical, intent(in)        :: useQR
Andreas Marek's avatar
Andreas Marek committed
184

Andreas Marek's avatar
Andreas Marek committed
185
      integer(kind=ik)           :: mystart, myend, m_way, n_way, work_per_thread, m_id, n_id, n_threads, ii, pp, transformChunkSize
Andreas Marek's avatar
Andreas Marek committed
186
187

#ifdef HAVE_DETAILED_TIMINGS
188
      call timer%start("bandred_real")
Andreas Marek's avatar
Andreas Marek committed
189
#endif
190
191
192
193
194
      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)
      success = .true.
Andreas Marek's avatar
Andreas Marek committed
195
196


197
198
199
200
201
202
203
204
205
206
207
      ! 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_real: ERROR: nbw=',nbw,', nblk=',nblk
            write(error_unit,*) 'ELPA2_bandred_real: ELPA2 works only for nbw==n*nblk'
          endif
          success = .false.
          return
        endif
      endif
Andreas Marek's avatar
Andreas Marek committed
208

209
210
211
212
      if (useGPU) then
        na_rows = numroc(na, nblk, my_prow, 0, np_rows)
        na_cols = numroc(na, nblk, my_pcol, 0, np_cols)
      endif
Andreas Marek's avatar
Andreas Marek committed
213

214
      ! Matrix is split into tiles; work is done only for tiles on the diagonal or above
Andreas Marek's avatar
Andreas Marek committed
215

216
217
      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
Andreas Marek's avatar
Andreas Marek committed
218

219
220
      l_rows_tile = tile_size/np_rows ! local rows of a tile
      l_cols_tile = tile_size/np_cols ! local cols of a tile
Andreas Marek's avatar
Andreas Marek committed
221

222
      if (useQR) then
Andreas Marek's avatar
Andreas Marek committed
223

224
225
226
227
        if (useGPU) then
          print *,"qr decomposition at the moment not supported with GPU"
          stop
        endif
Andreas Marek's avatar
Andreas Marek committed
228

229
        if (which_qr_decomposition == 1) then
230
          call qr_pqrparam_init(pqrparam(1:11),    nblk,'M',0,   nblk,'M',0,   nblk,'M',1,'s')
231
232
233
234
235
          allocate(tauvector(na), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when allocating tauvector "//errorMessage
            stop
          endif
Andreas Marek's avatar
Andreas Marek committed
236

237
238
239
240
241
          allocate(blockheuristic(nblk), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when allocating blockheuristic "//errorMessage
            stop
          endif
Andreas Marek's avatar
Andreas Marek committed
242

243
          l_rows = local_index(na, my_prow, np_rows, nblk, -1)
244
245
246
247
248
          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
Andreas Marek's avatar
Andreas Marek committed
249

250
251
          vmrCols = na
#ifdef DESPERATELY_WANT_ASSUMED_SIZE_QR
252
          call qr_pdgeqrf_2dcomm(a, lda, matrixCols, vmrCPU, max(l_rows,1), vmrCols, tauvector(1), na, tmat(1,1,1), &
253
254
255
256
                                 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
257
          call qr_pdgeqrf_2dcomm(a(1:lda,1:matrixCols), matrixCols, lda, vmrCPU(1:max(l_rows,1),1:vmrCols), max(l_rows,1), &
258
259
260
261
                                 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
262
          work_size = dwork_size(1)
263
264
265
266
267
          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
Andreas Marek's avatar
Andreas Marek committed
268

269
          work_blocked = 0.0_rk
270
271
272
273
274
          deallocate(vmrCPU, stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when deallocating vmrCPU "//errorMessage
            stop
          endif
Andreas Marek's avatar
Andreas Marek committed
275

276
        endif ! which_qr_decomposition
Andreas Marek's avatar
Andreas Marek committed
277

278
      endif ! useQr
Andreas Marek's avatar
Andreas Marek committed
279

280
281
282
283
284
285
      if (useGPU) then
        ! Here we convert the regular host array into a pinned host array
        successCUDA = cuda_malloc(a_dev, lda*na_cols*size_of_real_datatype)
        if (.not.(successCUDA)) then
          print *,"bandred_real: error in cudaMalloc"
          stop
286
        endif
Andreas Marek's avatar
Andreas Marek committed
287

288
289
290
291
292
        successCUDA = cuda_malloc(tmat_dev, nbw*nbw*size_of_real_datatype)
        if (.not.(successCUDA)) then
          print *,"bandred_real: error in cudaMalloc"
          stop
        endif
Andreas Marek's avatar
Andreas Marek committed
293

294
295
296
297
298
        successCUDA = cuda_malloc(vav_dev, nbw*nbw*size_of_real_datatype)
        if (.not.(successCUDA)) then
          print *,"bandred_real: error in cudaMalloc"
          stop
        endif
Andreas Marek's avatar
Andreas Marek committed
299

300
301
        cur_l_rows = 0
        cur_l_cols = 0
Andreas Marek's avatar
Andreas Marek committed
302

303
        successCUDA = cuda_memcpy(a_dev, loc(a(1,1)), (lda)*(na_cols)*size_of_real_datatype, cudaMemcpyHostToDevice)
304
305
306
307
308
        if (.not.(successCUDA)) then
          print *,"bandred_real: error in cudaMemcpy"
          stop
        endif
      endif ! useGPU
Andreas Marek's avatar
Andreas Marek committed
309
310


311
      do istep = (na-1)/nbw, 1, -1
Andreas Marek's avatar
Andreas Marek committed
312

313
        n_cols = MIN(na,(istep+1)*nbw) - istep*nbw ! Number of columns in current step
Andreas Marek's avatar
Andreas Marek committed
314

315
316
317
        ! 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)
Andreas Marek's avatar
Andreas Marek committed
318

319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
        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_real: error when deallocating vr "//errorMessage
                stop
              endif
            endif
            allocate(vr(l_rows + 1), stat=istat, errmsg=errorMessage)
            if (istat .ne. 0) then
              print *,"bandred_real: error when allocating vr "//errorMessage
              stop
            endif
Andreas Marek's avatar
Andreas Marek committed
341

342
          endif
Andreas Marek's avatar
Andreas Marek committed
343

344
345
346
347
348
349
350
          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_real: error when allocating vmrCUDA "//errorMessage
                stop
              endif
Andreas Marek's avatar
Andreas Marek committed
351

352
353
354
355
356
357
              successCUDA = cuda_free(vmr_dev)
              if (.not.(successCUDA)) then
                print *,"bandred_real: error in cuda_free"
                stop
              endif
            endif
Andreas Marek's avatar
Andreas Marek committed
358

359
360
361
362
363
            allocate(vmrCUDA(vmr_size), stat=istat, errmsg=errorMessage)
            if (istat .ne. 0) then
              print *,"bandred_real: error when allocating vmrCUDA "//errorMessage
              stop
            endif
Andreas Marek's avatar
Andreas Marek committed
364

365
366
367
368
369
            successCUDA = cuda_malloc(vmr_dev, vmr_size*size_of_real_datatype)
            if (.not.(successCUDA)) then
              print *,"bandred_real: error in cudaMalloc"
              stop
            endif
Andreas Marek's avatar
Andreas Marek committed
370

371
          endif
Andreas Marek's avatar
Andreas Marek committed
372

373
374
375
376
377
378
379
          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_real: error when deallocating umcCUDA "//errorMessage
                stop
              endif
Andreas Marek's avatar
Andreas Marek committed
380

381
382
383
384
385
              successCUDA = cuda_free(umc_dev)
              if (.not.(successCUDA)) then
                 print *,"bandred_real: error in cudaFree"
                 stop
              endif
Andreas Marek's avatar
Andreas Marek committed
386

387
            endif
Andreas Marek's avatar
Andreas Marek committed
388

389
390
391
392
393
            allocate(umcCUDA(umc_size), stat=istat, errmsg=errorMessage)
            if (istat .ne. 0) then
              print *,"bandred_real: error when deallocating umcCUDA "//errorMessage
              stop
            endif
Andreas Marek's avatar
Andreas Marek committed
394

395
396
397
398
399
            successCUDA = cuda_malloc(umc_dev, umc_size*size_of_real_datatype)
            if (.not.(successCUDA)) then
              print *,"bandred_real: error in cudaMalloc"
              stop
            endif
Andreas Marek's avatar
Andreas Marek committed
400

401
402
403
          endif
        else ! GPU not used
          ! Allocate vmr and umc to their exact sizes so that they can be used in bcasts and reduces
Andreas Marek's avatar
Andreas Marek committed
404

405
406
407
408
409
          allocate(vmrCPU(max(l_rows,1),2*n_cols), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when allocating vmrCPU "//errorMessage
            stop
          endif
Andreas Marek's avatar
Andreas Marek committed
410

411
412
413
414
415
          allocate(umcCPU(max(l_cols,1),2*n_cols), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when allocating umcCPU "//errorMessage
            stop
          endif
Andreas Marek's avatar
Andreas Marek committed
416

417
418
419
420
421
422
          allocate(vr(l_rows+1), stat=istat, errmsg=errorMessage)
          if (istat .ne. 0) then
            print *,"bandred_real: error when allocating vr "//errorMessage
            stop
          endif
        endif ! use GPU
Andreas Marek's avatar
Andreas Marek committed
423

424
        if (useGPU) then
425
          vmrCUDA(1 : cur_l_rows * n_cols) = 0._rk
426
        else
427
          vmrCPU(1:l_rows,1:n_cols) = 0._rk
428
        endif
Andreas Marek's avatar
Andreas Marek committed
429

430
431
        vr(:) = 0
        tmat(:,:,istep) = 0
Andreas Marek's avatar
Andreas Marek committed
432

433
        if (useGPU) then
434
          umcCUDA(1 : umc_size) = 0._rk
Andreas Marek's avatar
Andreas Marek committed
435

436
437
438
          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)
Andreas Marek's avatar
Andreas Marek committed
439

440
          if(lc_start .le. 0) lc_start = 1
Andreas Marek's avatar
Andreas Marek committed
441

442
443
          ! Here we assume that the processor grid and the block grid are aligned
          cur_pcol = pcol(istep*nbw+1, nblk, np_cols)
Andreas Marek's avatar
Andreas Marek committed
444

445
          if(my_pcol == cur_pcol) then
Andreas Marek's avatar
Andreas Marek committed
446

447
448
449
450
451
452
453
454
            successCUDA = cuda_memcpy2d(loc(a(1, lc_start)), lda*size_of_real_datatype,         &
                                       (a_dev + ((lc_start-1) * lda*size_of_real_datatype)),    &
                                       lda*size_of_real_datatype, lr_end*size_of_real_datatype, &
                                       (lc_end - lc_start+1), cudaMemcpyDeviceToHost)
            if (.not.(successCUDA)) then
              print *,"bandred_real: error in cudaMemcpy2d"
              stop
            endif
Andreas Marek's avatar
Andreas Marek committed
455

456
457
          endif
        endif ! useGPU
Andreas Marek's avatar
Andreas Marek committed
458

459
        ! Reduce current block to lower triangular form
Andreas Marek's avatar
Andreas Marek committed
460

461
462
        if (useQR) then
          if (which_qr_decomposition == 1) then
463
464
            vmrCols = 2*n_cols
#ifdef DESPERATELY_WANT_ASSUMED_SIZE_QR
465
            call qr_pdgeqrf_2dcomm(a, lda, matrixCols, vmrCPU, max(l_rows,1), vmrCols, tauvector(1), &
466
                                   na, tmat(1,1,istep), nbw, nbw, work_blocked, work_size,        &
467
468
                                     work_size, na, n_cols, nblk, nblk,        &
                                     istep*nbw+n_cols-nbw, istep*nbw+n_cols, 1,&
469
                                     0, PQRPARAM(1:11), mpi_comm_rows, mpi_comm_cols,&
470
                                     blockheuristic)
471
472

#else
473
            call qr_pdgeqrf_2dcomm(a(1:lda,1:matrixCols), lda, matrixCols, vmrCPU(1:max(l_rows,1),1:vmrCols) ,   &
474
475
476
477
478
479
480
                                    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
481
          endif
482
       else !useQR
Andreas Marek's avatar
Andreas Marek committed
483

484
         do lc = n_cols, 1, -1
Andreas Marek's avatar
Andreas Marek committed
485

486
487
           ncol = istep*nbw + lc ! absolute column number of householder vector
           nrow = ncol - nbw ! Absolute number of pivot row
Andreas Marek's avatar
Andreas Marek committed
488

489
490
           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
Andreas Marek's avatar
Andreas Marek committed
491

492
           tau = 0
Andreas Marek's avatar
Andreas Marek committed
493

494
           if (nrow == 1) exit ! Nothing to do
Andreas Marek's avatar
Andreas Marek committed
495

496
           cur_pcol = pcol(ncol, nblk, np_cols) ! Processor column owning current block
Andreas Marek's avatar
Andreas Marek committed
497

498
           if (my_pcol==cur_pcol) then
Andreas Marek's avatar
Andreas Marek committed
499

500
501
             ! Get vector to be transformed; distribute last element and norm of
             ! remaining elements to all procs in current column
Andreas Marek's avatar
Andreas Marek committed
502

503
             vr(1:lr) = a(1:lr,lch) ! vector to be transformed
Andreas Marek's avatar
Andreas Marek committed
504

505
506
507
508
509
             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))
510
               aux1(2) = 0._rk
511
             endif
Andreas Marek's avatar
Andreas Marek committed
512

513
514
515
516
517
518
#ifdef DOUBLE_PRECISION_REAL
             call mpi_allreduce(aux1, aux2, 2, MPI_REAL8, MPI_SUM, mpi_comm_rows, mpierr)
#else
             call mpi_allreduce(aux1, aux2, 2, MPI_REAL4, MPI_SUM, mpi_comm_rows, mpierr)
#endif

Andreas Marek's avatar
Andreas Marek committed
519

520
521
             vnorm2 = aux2(1)
             vrl    = aux2(2)
Andreas Marek's avatar
Andreas Marek committed
522

523
             ! Householder transformation
Andreas Marek's avatar
Andreas Marek committed
524

525
             call hh_transform_real(vrl, vnorm2, xf, tau)
Andreas Marek's avatar
Andreas Marek committed
526

527
             ! Scale vr and store Householder vector for back transformation
Andreas Marek's avatar
Andreas Marek committed
528

529
530
531
532
             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
533
               vr(lr) = 1._rk
534
535
536
             else
               a(1:lr,lch) = vr(1:lr)
             endif
537

538
           endif
539

540
           ! Broadcast Householder vector and tau along columns
541

542
           vr(lr+1) = tau
543
544
545
546
547
#ifdef DOUBLE_PRECISION_REAL
           call MPI_Bcast(vr, lr+1, MPI_REAL8, cur_pcol, mpi_comm_cols, mpierr)
#else
           call MPI_Bcast(vr, lr+1, MPI_REAL4, cur_pcol, mpi_comm_cols, mpierr)
#endif
548

549
550
           if (useGPU) then
             vmrCUDA(cur_l_rows * (lc - 1) + 1 : cur_l_rows * (lc - 1) + lr) = vr(1:lr)
Andreas Marek's avatar
Andreas Marek committed
551
           else
552
             vmrCPU(1:lr,lc) = vr(1:lr)
Andreas Marek's avatar
Andreas Marek committed
553
554
           endif

555
556
           tau = vr(lr+1)
           tmat(lc,lc,istep) = tau ! Store tau in diagonal of tmat
557

558
559
           ! Transform remaining columns in current block with Householder vector
           ! Local dot product
560

561
           aux1 = 0
562

563
564
565
566
#ifdef WITH_OPENMP
           !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
567

568
569
           !$omp parallel private(mynlc, j, lcx, ii, pp ) shared(aux1)
           mynlc = 0 ! number of local columns
570

571
572
573
574
575
576
577
578
579
580
581
582
583
584
           !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
585

586
587
588
           ! Get global dot products
           !$omp barrier
           !$omp single
589
590
591
592
593
#ifdef DOUBLE_PRECISION_REAL
           if (mynlc>0) call mpi_allreduce(aux1, aux2, mynlc, MPI_REAL8, MPI_SUM, mpi_comm_rows, mpierr)
#else
           if (mynlc>0) call mpi_allreduce(aux1, aux2, mynlc, MPI_REAL8, MPI_SUM, mpi_comm_rows, mpierr)
#endif
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
           !$omp end single
           !$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
#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
624

625
           ! Get global dot products
626
627
628
629
630
#ifdef DOUBLE_PRECISION_REAL
           if (nlc>0) call mpi_allreduce(aux1, aux2, nlc, MPI_REAL8, MPI_SUM, mpi_comm_rows, mpierr)
#else
           if (nlc>0) call mpi_allreduce(aux1, aux2, nlc, MPI_REAL4, MPI_SUM, mpi_comm_rows, mpierr)
#endif
631
           ! Transform
Andreas Marek's avatar
Andreas Marek committed
632

633
634
635
636
637
638
639
640
641
642
           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) - tau*aux2(nlc)*vr(1:lr)
             endif
           enddo
#endif /* WITH_OPENMP */
         enddo ! lc
Andreas Marek's avatar
Andreas Marek committed
643
644

         if (useGPU) then
645
646
647
648
649
650
651
652
653
654
655
           ! 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+((lc_start-1)*lda*size_of_real_datatype)),          &
                                          lda*size_of_real_datatype, loc(a(1, lc_start)),           &
                                          lda*size_of_real_datatype,  lr_end*size_of_real_datatype, &
                                          (lc_end - lc_start+1),cudaMemcpyHostToDevice)
             if (.not.(successCUDA)) then
               print *,"bandred_real: error in cudaMemcpy2d"
               stop
             endif
656

657
           endif
Andreas Marek's avatar
Andreas Marek committed
658
659
         endif

660
661
         ! Calculate scalar products of stored Householder vectors.
         ! This can be done in different ways, we use dsyrk
Andreas Marek's avatar
Andreas Marek committed
662

663
         vav = 0
Andreas Marek's avatar
Andreas Marek committed
664

665
#ifdef DOUBLE_PRECISION_REAL
666
667
         if (useGPU) then
           if (l_rows>0) &
668
             call dsyrk('U', 'T', n_cols, l_rows, 1.0_rk, vmrCUDA, cur_l_rows, 0.0_rk, vav, ubound(vav,dim=1))
669
670
         else
           if (l_rows>0) &
671
             call dsyrk('U', 'T', n_cols, l_rows, 1.0_rk, vmrCPU, ubound(vmrCPU,dim=1), 0.0_rk, vav, ubound(vav,dim=1))
672
         endif
673
674
675
676
677
678
679
680
681
682
#else
         if (useGPU) then
           if (l_rows>0) &
             call ssyrk('U', 'T', n_cols, l_rows, 1.0_rk, vmrCUDA, cur_l_rows, 0.0_rk, vav, ubound(vav,dim=1))
         else
           if (l_rows>0) &
             call ssyrk('U', 'T', n_cols, l_rows, 1.0_rk, vmrCPU, ubound(vmrCPU,dim=1), 0.0_rk, vav, ubound(vav,dim=1))
         endif
#endif

683
         call symm_matrix_allreduce(n_cols,vav, nbw, nbw,mpi_comm_rows)
Andreas Marek's avatar
Andreas Marek committed
684

685
         ! Calculate triangular matrix T for block Householder Transformation
Andreas Marek's avatar
Andreas Marek committed
686

687
688
689
         do lc=n_cols,1,-1
           tau = tmat(lc,lc,istep)
           if (lc<n_cols) then
690
691
692
693
694
#ifdef DOUBLE_PRECISION_REAL
             call dtrmv('U', 'T', 'N', n_cols-lc, tmat(lc+1,lc+1,istep), ubound(tmat,dim=1), vav(lc+1,lc), 1)
#else
             call strmv('U', 'T', 'N', n_cols-lc, tmat(lc+1,lc+1,istep), ubound(tmat,dim=1), vav(lc+1,lc), 1)
#endif
695
             tmat(lc,lc+1:n_cols,istep) = -tau * vav(lc+1:n_cols,lc)
Andreas Marek's avatar
Andreas Marek committed
696
697
           endif
         enddo
698
       endif
Andreas Marek's avatar
Andreas Marek committed
699

700
       ! Transpose vmr -> vmc (stored in umc, second half)
Andreas Marek's avatar
Andreas Marek committed
701
702

       if (useGPU) then
703
704
705
706
707
708
709
         call elpa_transpose_vectors_real  (vmrCUDA, cur_l_rows, mpi_comm_rows, &
                                            umcCUDA(cur_l_cols * n_cols + 1), cur_l_cols, mpi_comm_cols, &
                                            1, istep*nbw, n_cols, nblk)
       else
         call elpa_transpose_vectors_real  (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)
Andreas Marek's avatar
Andreas Marek committed
710
711
       endif

712
713
714
715
716
717
718
       ! 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

       ! 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
Andreas Marek's avatar
Andreas Marek committed
719
       if (useGPU) then
720
         umcCUDA(1 : l_cols * n_cols) = 0.0_rk
721
         vmrCUDA(cur_l_rows * n_cols + 1 : cur_l_rows * n_cols * 2) = 0
Andreas Marek's avatar
Andreas Marek committed
722

723
724
725
726
727
728
         if (l_cols>0 .and. l_rows>0) then
           successCUDA = cuda_memcpy(vmr_dev, loc(vmrCUDA(1)), vmr_size*size_of_real_datatype,cudaMemcpyHostToDevice)
           if (.not.(successCUDA)) then
             print *,"bandred_real: error in cudaMemcpy"
             stop
           endif
Andreas Marek's avatar
Andreas Marek committed
729

730
731
732
733
734
           successCUDA = cuda_memcpy(umc_dev, loc(umcCUDA(1)), umc_size*size_of_real_datatype,cudaMemcpyHostToDevice)
           if (.not.(successCUDA)) then
             print *,"bandred_real: error in cudaMemcpy"
             stop
           endif
Andreas Marek's avatar
Andreas Marek committed
735

736
           do i=0,(istep*nbw-1)/tile_size
Andreas Marek's avatar
Andreas Marek committed
737

738
739
740
             lcs = i*l_cols_tile+1
             lce = min(l_cols,(i+1)*l_cols_tile)
             if (lce<lcs) cycle
Andreas Marek's avatar
Andreas Marek committed
741

742
             lre = min(l_rows,(i+1)*l_rows_tile)
743
744
745
746
747
748
749
750
751
#ifdef DOUBLE_PRECISION_REAL
             call cublas_dgemm('T', 'N', lce-lcs+1, n_cols, lre, &
                               1.0_rk, (a_dev + ((lcs-1)*lda*size_of_real_datatype)), lda, vmr_dev,cur_l_rows, &
                               1.0_rk, (umc_dev+ (lcs-1)*size_of_real_datatype), cur_l_cols)
#else
             call cublas_sgemm('T', 'N', lce-lcs+1, n_cols, lre, &
                               1.0_rk, (a_dev + ((lcs-1)*lda*size_of_real_datatype)), lda, vmr_dev,cur_l_rows, &
                               1.0_rk, (umc_dev+ (lcs-1)*size_of_real_datatype), cur_l_cols)
#endif
752
753
             if(i==0) cycle
             lre = min(l_rows,i*l_rows_tile)
754
755
756
757
758
759
760
761
#ifdef DOUBLE_PRECISION_REAL
             call cublas_dgemm('N', 'N', lre,n_cols, lce-lcs+1,&
                               1.0_rk, (a_dev+ ((lcs-1)*lda*size_of_real_datatype)), lda,                  &
                               (umc_dev+(cur_l_cols * n_cols+lcs-1)*size_of_real_datatype), cur_l_cols, &
                               1.0_rk, (vmr_dev+(cur_l_rows * n_cols)*size_of_real_datatype), cur_l_rows)
#else
             call cublas_sgemm('N', 'N', lre,n_cols, lce-lcs+1,&
                               1.0_rk, (a_dev+ ((lcs-1)*lda*size_of_real_datatype)), lda,                  &
762
                               (umc_dev+(cur_l_cols * n_cols+lcs-1)*size_of_real_datatype), cur_l_cols, &
763
764
                               1.0_rk, (vmr_dev+(cur_l_rows * n_cols)*size_of_real_datatype), cur_l_rows)
#endif
765
           enddo
Andreas Marek's avatar
Andreas Marek committed
766

767
768
769
770
771
           successCUDA = cuda_memcpy(loc(vmrCUDA(1)), vmr_dev,vmr_size*size_of_real_datatype,cudaMemcpyDeviceToHost)
           if (.not.(successCUDA)) then
             print *,"bandred_real: error in cudaMemcpy"
             stop
           endif
Andreas Marek's avatar
Andreas Marek committed
772

773
774
775
776
777
           successCUDA = cuda_memcpy(loc(umcCUDA(1)), umc_dev, umc_size*size_of_real_datatype,cudaMemcpyDeviceToHost)
           if (.not.(successCUDA)) then
             print *,"bandred_real: error in cudaMemcpy"
             stop
           endif
Andreas Marek's avatar
Andreas Marek committed
778

779
         endif ! l_cols>0 .and. l_rows>0
Andreas Marek's avatar
Andreas Marek committed
780

781
782
       else ! do not useGPU version
         !Code for Algorithm 4
Andreas Marek's avatar
Andreas Marek committed
783

784
785
786
787
788
789
790
791
792
793
794
795
         n_way = 1
#ifdef WITH_OPENMP
         n_way = omp_get_max_threads()
#endif
         !umc(1:l_cols,1:n_cols) = 0.d0
         !vmr(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)
796
             umcCPU(i,1:n_cols) = 0.0_rk
797
798
799
           enddo
           !$omp do
           do i=1,l_rows
800
             vmrCPU(i,n_cols+1:2*n_cols) = 0.0_rk
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
           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
830
831
832
833
834
835
836
837
#ifdef DOUBLE_PRECISION_REAL
                 call DGEMM('N', 'N', lre-lrs+1, n_cols, l_cols-lcs+1,          &
                            1.0_rk, a(lrs,lcs), ubound(a,dim=1),                 &
                                  umcCPU(lcs,n_cols+1), ubound(umcCPU,dim=1),  &
                            0.0_rk, vmrCPU(lrs,n_cols+1), ubound(vmrCPU,dim=1))
#else
                 call SGEMM('N', 'N', lre-lrs+1, n_cols, l_cols-lcs+1,          &
                            1.0_rk, a(lrs,lcs), ubound(a,dim=1),                 &
838
                                  umcCPU(lcs,n_cols+1), ubound(umcCPU,dim=1),  &
839
840
                            0.0_rk, vmrCPU(lrs,n_cols+1), ubound(vmrCPU,dim=1))
#endif
841
               endif
Andreas Marek's avatar
Andreas Marek committed
842

843
844
               ! C1 += A10' B0
               if ( lce > lcs .and. i > 0 ) then
845
846
847
#ifdef DOUBLE_PRECISION_REAL
                 call DGEMM('T', 'N', lce-lcs+1, n_cols, lrs-1,           &
                            1.0_rk, a(1,lcs),   ubound(a,dim=1),           &
848
                                  vmrCPU(1,1),   ubound(vmrCPU,dim=1),   &
849
850
851
852
853
854
855
                            0.0_rk, umcCPU(lcs,1), ubound(umcCPU,dim=1))
#else
                 call SGEMM('T', 'N', lce-lcs+1, n_cols, lrs-1,           &
                            1.0_rk, a(1,lcs),   ubound(a,dim=1),           &
                                  vmrCPU(1,1),   ubound(vmrCPU,dim=1),   &
                            0.0_rk, umcCPU(lcs,1), ubound(umcCPU,dim=1))
#endif
856
857
858
859
               endif
             enddo
           endif ! l_cols>0 .and. l_rows>0
         else ! n_way > 1
860
           umcCPU(1:l_cols,1:n_cols) = 0.0_rk
861
862
863
864
865
866
867
868
869
           vmrCPU(1:l_rows,n_cols+1:2*n_cols) = 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

               lre = min(l_rows,(i+1)*l_rows_tile)
870
871
872
#ifdef DOUBLE_PRECISION_REAL
               call DGEMM('T', 'N', lce-lcs+1, n_cols, lre, 1.0_rk, a(1,lcs), ubound(a,dim=1), &
                            vmrCPU, ubound(vmrCPU,dim=1), 1.0_rk, umcCPU(lcs,1), ubound(umcCPU,dim=1))
873
874
875

               if (i==0) cycle
                 lre = min(l_rows,i*l_rows_tile)
876
877
878
879
880
881
882
883
884
885
886
887
888
                 call DGEMM('N', 'N', lre, n_cols, lce-lcs+1, 1.0_rk, a(1,lcs), lda, &
                            umcCPU(lcs,n_cols+1), ubound(umcCPU,dim=1), 1.0_rk, vmrCPU(1,n_cols+1), ubound(vmrCPU,dim=1))
#else
               call SGEMM('T', 'N', lce-lcs+1, n_cols, lre, 1.0_rk, a(1,lcs), ubound(a,dim=1), &
                            vmrCPU, ubound(vmrCPU,dim=1), 1.0_rk, umcCPU(lcs,1), ubound(umcCPU,dim=1))

               if (i==0) cycle
                 lre = min(l_rows,i*l_rows_tile)
                 call SGEMM('N', 'N', lre, n_cols, lce-lcs+1, 1.0_rk, a(1,lcs), lda, &
                            umcCPU(lcs,n_cols+1), ubound(umcCPU,dim=1), 1.0_rk, vmrCPU(1,n_cols+1), ubound(vmrCPU,dim=1))
#endif


889
890
891
892
893
             enddo
           endif
         endif ! n_way > 1
#ifdef WITH_OPENMP
        !$omp end parallel
894
#endif
895
       endif ! do not useGPU version
Andreas Marek's avatar
Andreas Marek committed
896

897
898
899
900
       ! 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
Andreas Marek's avatar
Andreas Marek committed
901

902
903
       if (useGPU) then
         ! here the GPU version and CPU version divereged due to the same reasons as above
Andreas Marek's avatar
Andreas Marek committed
904

905
906
907
908
909
         if (tile_size < istep*nbw) then
           call elpa_reduce_add_vectors_real  (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
Andreas Marek's avatar
Andreas Marek committed
910

911
912
913
914
915
916
         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
Andreas Marek's avatar
Andreas Marek committed
917

918
919
920
921
922
#ifdef DOUBLE_PRECISION_REAL
           call mpi_allreduce(umcCUDA, tmpCUDA, l_cols*n_cols, MPI_REAL8, MPI_SUM, mpi_comm_rows, ierr)
#else
           call mpi_allreduce(umcCUDA, tmpCUDA, l_cols*n_cols, MPI_REAL4, MPI_SUM, mpi_comm_rows, ierr)
#endif
923
           umcCUDA(1 : l_cols * n_cols) = tmpCUDA(1 : l_cols * n_cols)
Andreas Marek's avatar
Andreas Marek committed
924

925
926
927
928
929
930
931
932
           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
Andreas Marek's avatar
Andreas Marek committed
933

934
935
936
937
938
939
         ! U = U * Tmat**T
         successCUDA = cuda_memcpy(umc_dev, loc(umcCUDA(1)), umc_size*size_of_real_datatype, cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
Andreas Marek's avatar
Andreas Marek committed
940

941
942
943
944
945
         successCUDA = cuda_memcpy(tmat_dev,loc(tmat(1,1,istep)),nbw*nbw*size_of_real_datatype,cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
946
947
948
949
950
951
952
#ifdef DOUBLE_PRECISION_REAL
         call cublas_dtrmm('Right', 'Upper', 'Trans', 'Nonunit', l_cols, n_cols, &
                           1.0_rk, tmat_dev, nbw, umc_dev, cur_l_cols)
#else
         call cublas_strmm('Right', 'Upper', 'Trans', 'Nonunit', l_cols, n_cols, &
                           1.0_rk, tmat_dev, nbw, umc_dev, cur_l_cols)
#endif
953
         ! VAV = Tmat * V**T * A * V * Tmat**T = (U*Tmat**T)**T * V * Tmat**T
Andreas Marek's avatar
Andreas Marek committed
954

955
956
957
958
959
         successCUDA = cuda_memcpy(vav_dev,loc(vav(1,1)), nbw*nbw*size_of_real_datatype,cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
960
961
962
963
#ifdef DOUBLE_PRECISION_REAL
         call cublas_dgemm('T', 'N', n_cols, n_cols, l_cols, &
                           1.0_rk, umc_dev, cur_l_cols, (umc_dev+(cur_l_cols * n_cols )*size_of_real_datatype),cur_l_cols, &
                           0.0_rk, vav_dev, nbw)
Andreas Marek's avatar
Andreas Marek committed
964

965
966
967
968
969
970
971
972
973
974
         call cublas_dtrmm('Right', 'Upper', 'Trans', 'Nonunit', n_cols, n_cols, &
                           1.0_rk, tmat_dev, nbw, vav_dev, nbw)
#else
         call cublas_sgemm('T', 'N', n_cols, n_cols, l_cols, &
                           1.0_rk, umc_dev, cur_l_cols, (umc_dev+(cur_l_cols * n_cols )*size_of_real_datatype),cur_l_cols, &
                           0.0_rk, vav_dev, nbw)

         call cublas_strmm('Right', 'Upper', 'Trans', 'Nonunit', n_cols, n_cols, &
                           1.0_rk, tmat_dev, nbw, vav_dev, nbw)
#endif
Andreas Marek's avatar
Andreas Marek committed
975
976
977



978
979
980
981
982
         successCUDA = cuda_memcpy(loc(vav(1,1)), vav_dev, nbw*nbw*size_of_real_datatype, cudaMemcpyDeviceToHost)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
Andreas Marek's avatar
Andreas Marek committed
983

984
         call symm_matrix_allreduce(n_cols,vav, nbw,nbw,mpi_comm_cols)
Andreas Marek's avatar
Andreas Marek committed
985

986
987
988
989
990
         successCUDA = cuda_memcpy(vav_dev, loc(vav(1,1)), nbw*nbw*size_of_real_datatype,cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
Andreas Marek's avatar
Andreas Marek committed
991

992
         ! U = U - 0.5 * V * VAV
993
994
995
996
997
998
999
1000
1001
#ifdef DOUBLE_PRECISION_REAL
         call cublas_dgemm('N', 'N', l_cols, n_cols, n_cols,&
                           -0.5_rk, (umc_dev+(cur_l_cols * n_cols )*size_of_real_datatype),cur_l_cols, vav_dev,nbw,&
                           1.0_rk, umc_dev, cur_l_cols)
#else
         call cublas_sgemm('N', 'N', l_cols, n_cols, n_cols,&
                           -0.5_rk, (umc_dev+(cur_l_cols * n_cols )*size_of_real_datatype),cur_l_cols, vav_dev,nbw,&
                           1.0_rk, umc_dev, cur_l_cols)
#endif
1002
1003
1004
1005
1006
         successCUDA = cuda_memcpy(loc(umcCUDA(1)), umc_dev, umc_size*size_of_real_datatype, cudaMemcpyDeviceToHost)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
Andreas Marek's avatar
Andreas Marek committed
1007

1008
         ! Transpose umc -> umr (stored in vmr, second half)
Andreas Marek's avatar
Andreas Marek committed
1009

1010
1011
1012
1013
1014
1015
1016
1017
         call elpa_transpose_vectors_real  (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_real_datatype, cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
Andreas Marek's avatar
Andreas Marek committed
1018

1019
1020
1021
1022
1023
         successCUDA = cuda_memcpy(umc_dev, loc(umcCUDA(1)), umc_size*size_of_real_datatype, cudaMemcpyHostToDevice)
         if (.not.(successCUDA)) then
           print *,"bandred_real: error in cudaMemcpy"
           stop
         endif
Andreas Marek's avatar
Andreas Marek committed
1024

1025
1026
1027
1028
1029
1030
         ! A = A - V*U**T - U*V**T
         do i=0,(istep*nbw-1)/tile_size
           lcs = i*l_cols_tile+1
           lce = min(l_cols,(i+1)*l_cols_tile)
           lre = min(l_rows,(i+1)*l_rows_tile)
           if (lce<lcs .or. lre<1) cycle
1031
1032
1033
1034
1035
1036
1037
1038
1039
#ifdef DOUBLE_PRECISION_REAL
           call cublas_dgemm('N', 'T', lre, lce-lcs+1, 2*n_cols, -1.0_rk, &
                             vmr_dev, cur_l_rows, (umc_dev +(lcs-1)*size_of_real_datatype), cur_l_cols, &
                             1.0_rk, (a_dev+(lcs-1)*lda*size_of_real_datatype), lda)
#else
           call cublas_sgemm('N', 'T', lre, lce-lcs+1, 2*n_cols, -1.0_rk, &
                             vmr_dev, cur_l_rows, (umc_dev +(lcs-1)*size_of_real_datatype), cur_l_cols, &
                             1.0_rk, (a_dev+(lcs-1)*lda*size_of_real_datatype), lda)
#endif
1040
1041
1042
1043
1044
1045
1046
1047
         enddo
       else ! do not useGPU
         ! Or if we used the Algorithm 4
         if (tile_size < istep*nbw .or. n_way > 1) then
         call elpa_reduce_add_vectors_real  (vmrCPU(1,n_cols+1),ubound(vmrCPU,dim=1),mpi_comm_rows, &
                                             umcCPU, ubound(umcCPU,dim=1), mpi_comm_cols, &
                                             istep*nbw, n_cols, nblk)
         endif
Andreas Marek's avatar
Andreas Marek committed
1048

1049
1050
1051
1052
1053
1054
         if (l_cols>0) then
           allocate(tmpCPU(l_cols,n_cols), stat=istat, errmsg=errorMessage)
           if (istat .ne. 0) then
             print *,"bandred_real: error when allocating tmpCPU "//errorMessage
             stop
           endif
1055
1056
1057
1058
1059
#ifdef DOUBLE_PRECISION_REAL
           call mpi_allreduce(umcCPU, tmpCPU, l_cols*n_cols, MPI_REAL8, MPI_SUM, mpi_comm_rows, mpierr)
#else
           call mpi_allreduce(umcCPU, tmpCPU, l_cols*n_cols, MPI_REAL4, MPI_SUM, mpi_comm_rows, mpierr)
#endif
1060