You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.

potrf_U_parallel.c 5.0 kB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130
  1. /*********************************************************************/
  2. /* Copyright 2009, 2010 The University of Texas at Austin. */
  3. /* All rights reserved. */
  4. /* */
  5. /* Redistribution and use in source and binary forms, with or */
  6. /* without modification, are permitted provided that the following */
  7. /* conditions are met: */
  8. /* */
  9. /* 1. Redistributions of source code must retain the above */
  10. /* copyright notice, this list of conditions and the following */
  11. /* disclaimer. */
  12. /* */
  13. /* 2. Redistributions in binary form must reproduce the above */
  14. /* copyright notice, this list of conditions and the following */
  15. /* disclaimer in the documentation and/or other materials */
  16. /* provided with the distribution. */
  17. /* */
  18. /* THIS SOFTWARE IS PROVIDED BY THE UNIVERSITY OF TEXAS AT */
  19. /* AUSTIN ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, */
  20. /* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF */
  21. /* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE */
  22. /* DISCLAIMED. IN NO EVENT SHALL THE UNIVERSITY OF TEXAS AT */
  23. /* AUSTIN OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, */
  24. /* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES */
  25. /* (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE */
  26. /* GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR */
  27. /* BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF */
  28. /* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT */
  29. /* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT */
  30. /* OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE */
  31. /* POSSIBILITY OF SUCH DAMAGE. */
  32. /* */
  33. /* The views and conclusions contained in the software and */
  34. /* documentation are those of the authors and should not be */
  35. /* interpreted as representing official policies, either expressed */
  36. /* or implied, of The University of Texas at Austin. */
  37. /*********************************************************************/
  38. #include <stdio.h>
  39. #include "common.h"
  40. blasint CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, FLOAT *sa, FLOAT *sb, BLASLONG myid) {
  41. BLASLONG n, bk, i, blocking, lda;
  42. BLASLONG info;
  43. int mode;
  44. blas_arg_t newarg;
  45. FLOAT *a;
  46. FLOAT alpha[2] = { -ONE, ZERO};
  47. #ifndef COMPLEX
  48. #ifdef XDOUBLE
  49. mode = BLAS_XDOUBLE | BLAS_REAL;
  50. #elif defined(DOUBLE)
  51. mode = BLAS_DOUBLE | BLAS_REAL;
  52. #else
  53. mode = BLAS_SINGLE | BLAS_REAL;
  54. #endif
  55. #else
  56. #ifdef XDOUBLE
  57. mode = BLAS_XDOUBLE | BLAS_COMPLEX;
  58. #elif defined(DOUBLE)
  59. mode = BLAS_DOUBLE | BLAS_COMPLEX;
  60. #else
  61. mode = BLAS_SINGLE | BLAS_COMPLEX;
  62. #endif
  63. #endif
  64. if (args -> nthreads == 1) {
  65. info = POTRF_U_SINGLE(args, NULL, NULL, sa, sb, 0);
  66. return info;
  67. }
  68. n = args -> n;
  69. a = (FLOAT *)args -> a;
  70. lda = args -> lda;
  71. if (range_n) n = range_n[1] - range_n[0];
  72. if (n <= GEMM_UNROLL_N * 4) {
  73. info = POTRF_U_SINGLE(args, NULL, range_n, sa, sb, 0);
  74. return info;
  75. }
  76. newarg.lda = lda;
  77. newarg.ldb = lda;
  78. newarg.ldc = lda;
  79. newarg.alpha = alpha;
  80. newarg.beta = NULL;
  81. newarg.nthreads = args -> nthreads;
  82. blocking = ((n / 2 + GEMM_UNROLL_N - 1)/GEMM_UNROLL_N) * GEMM_UNROLL_N;
  83. if (blocking > GEMM_Q) blocking = GEMM_Q;
  84. for (i = 0; i < n; i += blocking) {
  85. bk = n - i;
  86. if (bk > blocking) bk = blocking;
  87. newarg.m = bk;
  88. newarg.n = bk;
  89. newarg.a = a + (i + i * lda) * COMPSIZE;
  90. info = CNAME(&newarg, NULL, NULL, sa, sb, 0);
  91. if (info) return info + i;
  92. if (n - i - bk > 0) {
  93. newarg.m = bk;
  94. newarg.n = n - i - bk;
  95. newarg.a = a + (i + i * lda) * COMPSIZE;
  96. newarg.b = a + (i + (i + bk) * lda) * COMPSIZE;
  97. gemm_thread_n(mode | BLAS_TRANSA_T,
  98. &newarg, NULL, NULL, (int (*)(blas_arg_t *, BLASLONG *, BLASLONG *, FLOAT *, FLOAT *, BLASLONG))TRSM_LCUN, sa, sb, args -> nthreads);
  99. newarg.n = n - i - bk;
  100. newarg.k = bk;
  101. newarg.a = a + ( i + (i + bk) * lda) * COMPSIZE;
  102. newarg.c = a + ((i + bk) + (i + bk) * lda) * COMPSIZE;
  103. #ifndef USE_SIMPLE_THREADED_LEVEL3
  104. HERK_THREAD_UC(&newarg, NULL, NULL, sa, sb, 0);
  105. #else
  106. syrk_thread(mode | BLAS_TRANSA_N | BLAS_TRANSB_T,
  107. &newarg, NULL, NULL, (int (*)(blas_arg_t *, BLASLONG *, BLASLONG *, FLOAT *, FLOAT *, BLASLONG))HERK_UC, sa, sb, args -> nthreads);
  108. #endif
  109. }
  110. }
  111. return 0;
  112. }