zgeqrs_param.c 11.5 KB
Newer Older
1 2 3 4 5
/**
 *
 * @copyright (c) 2009-2014 The University of Tennessee and The University
 *                          of Tennessee Research Foundation.
 *                          All rights reserved.
BOUCHERIE Raphael's avatar
BOUCHERIE Raphael committed
6
 * @copyright (c) 2012-2017 Bordeaux INP, CNRS (LaBRI UMR 5800), Inria, Univ. Bordeaux. All rights reserved.
7 8 9 10 11 12 13 14 15 16 17
 *
 **/

/**
 *
 * @file zgeqrs_param.c
 *
 *  MORSE computational routines
 *  MORSE is a software package provided by Univ. of Tennessee,
 *  Univ. of California Berkeley and Univ. of Colorado Denver
 *
BOUCHERIE Raphael's avatar
BOUCHERIE Raphael committed
18
 * @version 1.0.0
19
 * @author Mathieu Faverge
BOUCHERIE Raphael's avatar
BOUCHERIE Raphael committed
20 21
 * @author Raphael Boucherie
 * @date 2017-05-17
22 23 24 25
 * @precisions normal z -> s d c
 *
 **/
#include "control/common.h"
Mathieu Faverge's avatar
Mathieu Faverge committed
26
#include <stdlib.h>
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 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218

/**
 *******************************************************************************
 *
 * @ingroup MORSE_Complex64_t
 *
 *  MORSE_zgeqrs_param - Compute a minimum-norm solution min || A*X - B || using the RQ factorization
 *  A = R*Q computed by MORSE_zgeqrf.
 *
 *******************************************************************************
 *
 * @param[in] M
 *          The number of rows of the matrix A. M >= 0.
 *
 * @param[in] N
 *          The number of columns of the matrix A. N >= M >= 0.
 *
 * @param[in] NRHS
 *          The number of columns of B. NRHS >= 0.
 *
 * @param[in,out] A
 *          Details of the QR factorization of the original matrix A as returned by MORSE_zgeqrf.
 *
 * @param[in] LDA
 *          The leading dimension of the array A. LDA >= M.
 *
 * @param[in] descT
 *          Auxiliary factorization data, computed by MORSE_zgeqrf.
 *
 * @param[in,out] B
 *          On entry, the m-by-nrhs right hand side matrix B.
 *          On exit, the n-by-nrhs solution matrix X.
 *
 * @param[in] LDB
 *          The leading dimension of the array B. LDB >= max(1,N).
 *
 *******************************************************************************
 *
 * @return
 *          \retval MORSE_SUCCESS successful exit
 *          \retval <0 if -i, the i-th argument had an illegal value
 *
 *******************************************************************************
 *
 * @sa MORSE_zgeqrs_param_Tile
 * @sa MORSE_zgeqrs_param_Tile_Async
 * @sa MORSE_cgeqrs
 * @sa MORSE_dgeqrs
 * @sa MORSE_sgeqrs
 * @sa MORSE_zgeqrf
 *
 *******************************************************************************/
int MORSE_zgeqrs_param(const libhqr_tree_t *qrtree, int M, int N, int NRHS,
                       MORSE_Complex64_t *A, int LDA,
                       MORSE_desc_t *descTS, MORSE_desc_t *descTT,
                       MORSE_Complex64_t *B, int LDB)
{
    int NB;
    int status;
    MORSE_context_t *morse;
    MORSE_sequence_t *sequence = NULL;
    MORSE_request_t request = MORSE_REQUEST_INITIALIZER;
    MORSE_desc_t descA, descB;

    morse = morse_context_self();
    if (morse == NULL) {
        morse_fatal_error("MORSE_zgeqrs_param", "MORSE not initialized");
        return MORSE_ERR_NOT_INITIALIZED;
    }

    /* Check input arguments */
    if (M < 0) {
        morse_error("MORSE_zgeqrs_param", "illegal value of M");
        return -1;
    }
    if (N < 0 || N > M) {
        morse_error("MORSE_zgeqrs_param", "illegal value of N");
        return -2;
    }
    if (NRHS < 0) {
        morse_error("MORSE_zgeqrs_param", "illegal value of N");
        return -3;
    }
    if (LDA < chameleon_max(1, M)) {
        morse_error("MORSE_zgeqrs_param", "illegal value of LDA");
        return -5;
    }
    if (LDB < chameleon_max(1, chameleon_max(1, M))) {
        morse_error("MORSE_zgeqrs_param", "illegal value of LDB");
        return -8;
    }
    /* Quick return */
    if (chameleon_min(M, chameleon_min(N, NRHS)) == 0) {
        return MORSE_SUCCESS;
    }

    /* Tune NB & IB depending on M, N & NRHS; Set NBNBSIZE */
    status = morse_tune(MORSE_FUNC_ZGELS, M, N, NRHS);
    if (status != MORSE_SUCCESS) {
        morse_error("MORSE_zgeqrs_param", "morse_tune() failed");
        return status;
    }

    /* Set NT */
    NB = MORSE_NB;

    morse_sequence_create(morse, &sequence);

/*    if ( MORSE_TRANSLATION == MORSE_OUTOFPLACE ) {*/
        morse_zooplap2tile( descA, A, NB, NB, LDA, N,    0, 0, M, N,    sequence, &request,
                             morse_desc_mat_free(&(descA)) );
        morse_zooplap2tile( descB, B, NB, NB, LDB, NRHS, 0, 0, M, NRHS, sequence, &request,
                             morse_desc_mat_free(&(descA)); morse_desc_mat_free(&(descB)));
/*    } else {*/
/*        morse_ziplap2tile( descA, A, NB, NB, LDA, N,    0, 0, M, N,   */
/*                            sequence, &request);*/
/*        morse_ziplap2tile( descB, B, NB, NB, LDB, NRHS, 0, 0, M, NRHS,*/
/*                            sequence, &request);*/
/*    }*/

    /* Call the tile interface */
    MORSE_zgeqrs_param_Tile_Async(qrtree, &descA, descTS, descTT, &descB, sequence, &request);

/*    if ( MORSE_TRANSLATION == MORSE_OUTOFPLACE ) {*/
        morse_zooptile2lap(descA, A, NB, NB, LDA, N,     sequence, &request);
        morse_zooptile2lap(descB, B, NB, NB, LDB, NRHS,  sequence, &request);
        morse_sequence_wait(morse, sequence);
        morse_desc_mat_free(&descA);
        morse_desc_mat_free(&descB);
/*    } else {*/
/*        morse_ziptile2lap( descA, A, NB, NB, LDA, N,     sequence, &request);*/
/*        morse_ziptile2lap( descB, B, NB, NB, LDB, NRHS,  sequence, &request);*/
/*        morse_sequence_wait(morse, sequence);*/
/*    }*/

    status = sequence->status;
    morse_sequence_destroy(morse, sequence);
    return status;
}

/**
 *******************************************************************************
 *
 * @ingroup MORSE_Complex64_t_Tile
 *
 *  MORSE_zgeqrs_param_Tile - Computes a minimum-norm solution using the tile QR factorization.
 *  Tile equivalent of MORSE_zgeqrf().
 *  Operates on matrices stored by tiles.
 *  All matrices are passed through descriptors.
 *  All dimensions are taken from the descriptors.
 *
 *******************************************************************************
 *
 * @param[in,out] A
 *          Details of the QR factorization of the original matrix A as returned by MORSE_zgeqrf.
 *
 * @param[in] T
 *          Auxiliary factorization data, computed by MORSE_zgeqrf.
 *
 * @param[in,out] B
 *          On entry, the m-by-nrhs right hand side matrix B.
 *          On exit, the n-by-nrhs solution matrix X.
 *
 *******************************************************************************
 *
 * @return
 *          \retval MORSE_SUCCESS successful exit
 *
 *******************************************************************************
 *
 * @sa MORSE_zgeqrs_param
 * @sa MORSE_zgeqrs_param_Tile_Async
 * @sa MORSE_cgeqrs_Tile
 * @sa MORSE_dgeqrs_Tile
 * @sa MORSE_sgeqrs_Tile
 * @sa MORSE_zgeqrf_Tile
 *
 ******************************************************************************/
int MORSE_zgeqrs_param_Tile(const libhqr_tree_t *qrtree, MORSE_desc_t *A, MORSE_desc_t *TS, MORSE_desc_t *TT, MORSE_desc_t *B)
{
    MORSE_context_t *morse;
    MORSE_sequence_t *sequence = NULL;
    MORSE_request_t request = MORSE_REQUEST_INITIALIZER;
    int status;

    morse = morse_context_self();
    if (morse == NULL) {
        morse_fatal_error("MORSE_zgeqrs_param_Tile", "MORSE not initialized");
        return MORSE_ERR_NOT_INITIALIZED;
    }
    morse_sequence_create(morse, &sequence);
    MORSE_zgeqrs_param_Tile_Async(qrtree, A, TS, TT, B, sequence, &request);
219 220 221 222
    RUNTIME_desc_getoncpu_async( A, sequence );
    RUNTIME_desc_getoncpu_async( TS, sequence );
    RUNTIME_desc_getoncpu_async( TT, sequence );
    RUNTIME_desc_getoncpu_async( B, sequence );
Mathieu Faverge's avatar
Mathieu Faverge committed
223
    morse_sequence_wait(morse, sequence);
224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266

    status = sequence->status;
    morse_sequence_destroy(morse, sequence);
    return status;
}

/**
 *******************************************************************************
 *
 * @ingroup MORSE_Complex64_t_Tile_Async
 *
 *  MORSE_zgeqrs_param_Tile_Async - Computes a minimum-norm solution using the tile
 *  QR factorization.
 *  Non-blocking equivalent of MORSE_zgeqrs_param_Tile().
 *  May return before the computation is finished.
 *  Allows for pipelining of operations at runtime.
 *
 *******************************************************************************
 *
 * @param[in] sequence
 *          Identifies the sequence of function calls that this call belongs to
 *          (for completion checks and exception handling purposes).
 *
 * @param[out] request
 *          Identifies this function call (for exception handling purposes).
 *
 *******************************************************************************
 *
 * @sa MORSE_zgeqrs_param
 * @sa MORSE_zgeqrs_param_Tile
 * @sa MORSE_cgeqrs_Tile_Async
 * @sa MORSE_dgeqrs_Tile_Async
 * @sa MORSE_sgeqrs_Tile_Async
 * @sa MORSE_zgeqrf_Tile_Async
 *
 ******************************************************************************/
int MORSE_zgeqrs_param_Tile_Async(const libhqr_tree_t *qrtree,
                                  MORSE_desc_t *A, MORSE_desc_t *TS, MORSE_desc_t *TT, MORSE_desc_t *B,
                                  MORSE_sequence_t *sequence, MORSE_request_t *request)
{
    MORSE_desc_t *subA;
    MORSE_desc_t *subB;
    MORSE_context_t *morse;
Mathieu Faverge's avatar
Mathieu Faverge committed
267
    MORSE_desc_t D, *Dptr = NULL;
BOUCHERIE Raphael's avatar
BOUCHERIE Raphael committed
268

269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
    morse = morse_context_self();
    if (morse == NULL) {
        morse_fatal_error("MORSE_zgeqrs_param_Tile", "MORSE not initialized");
        return MORSE_ERR_NOT_INITIALIZED;
    }
    if (sequence == NULL) {
        morse_fatal_error("MORSE_zgeqrs_param_Tile", "NULL sequence");
        return MORSE_ERR_UNALLOCATED;
    }
    if (request == NULL) {
        morse_fatal_error("MORSE_zgeqrs_param_Tile", "NULL request");
        return MORSE_ERR_UNALLOCATED;
    }
    /* Check sequence status */
    if (sequence->status == MORSE_SUCCESS)
        request->status = MORSE_SUCCESS;
    else
        return morse_request_fail(sequence, request, MORSE_ERR_SEQUENCE_FLUSHED);

    /* Check descriptors for correctness */
    if (morse_desc_check(A) != MORSE_SUCCESS) {
        morse_error("MORSE_zgeqrs_param_Tile", "invalid first descriptor");
        return morse_request_fail(sequence, request, MORSE_ERR_ILLEGAL_VALUE);
    }
    if (morse_desc_check(TS) != MORSE_SUCCESS) {
        morse_error("MORSE_zgeqrs_param_Tile", "invalid second descriptor");
        return morse_request_fail(sequence, request, MORSE_ERR_ILLEGAL_VALUE);
    }
    if (morse_desc_check(TT) != MORSE_SUCCESS) {
        morse_error("MORSE_zgeqrs_param_Tile", "invalid third descriptor");
        return morse_request_fail(sequence, request, MORSE_ERR_ILLEGAL_VALUE);
    }
    if (morse_desc_check(B) != MORSE_SUCCESS) {
        morse_error("MORSE_zgeqrs_param_Tile", "invalid fourth descriptor");
        return morse_request_fail(sequence, request, MORSE_ERR_ILLEGAL_VALUE);
    }
    /* Check input arguments */
    if (A->nb != A->mb || B->nb != B->mb) {
        morse_error("MORSE_zgeqrs_param_Tile", "only square tiles supported");
        return morse_request_fail(sequence, request, MORSE_ERR_ILLEGAL_VALUE);
    }
    /* Quick return */
    /*
     if (chameleon_min(M, chameleon_min(N, NRHS)) == 0) {
     return MORSE_SUCCESS;
     }
     */
BOUCHERIE Raphael's avatar
BOUCHERIE Raphael committed
316
#if defined(CHAMELEON_COPY_DIAG)
Mathieu Faverge's avatar
Mathieu Faverge committed
317 318 319 320 321
    {
        int n = chameleon_min(A->mt, A->nt) * A->nb;
        morse_zdesc_alloc(D, A->mb, A->nb, A->m, n, 0, 0, A->m, n, );
        Dptr = &D;
    }
BOUCHERIE Raphael's avatar
BOUCHERIE Raphael committed
322
#endif
323 324 325

    subB = morse_desc_submatrix(B, 0, 0, A->n, B->n);
    subA = morse_desc_submatrix(A, 0, 0, A->n, A->n);
Mathieu Faverge's avatar
Mathieu Faverge committed
326 327

    morse_pzunmqr_param(qrtree, MorseLeft, MorseConjTrans, A, B, TS, TT, Dptr, sequence, request);
328
    morse_pztrsm(MorseLeft, MorseUpper, MorseNoTrans, MorseNonUnit, 1.0, subA, subB, sequence, request);
Mathieu Faverge's avatar
Mathieu Faverge committed
329

330 331
    free(subA);
    free(subB);
Mathieu Faverge's avatar
Mathieu Faverge committed
332 333 334 335

    if (Dptr != NULL) {
        morse_desc_mat_free(Dptr);
    }
336
    (void)D;
337 338
    return MORSE_SUCCESS;
}