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.c 4.9 kB

1 year ago
123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150
  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. #ifdef FUNCTION_PROFILE
  41. #include "functable.h"
  42. #endif
  43. #ifdef XDOUBLE
  44. #define ERROR_NAME "QPOTRF"
  45. #elif defined(DOUBLE)
  46. #define ERROR_NAME "DPOTRF"
  47. #else
  48. #define ERROR_NAME "SPOTRF"
  49. #endif
  50. static blasint (*potrf_single[])(blas_arg_t *, BLASLONG *, BLASLONG *, FLOAT *, FLOAT *, BLASLONG) = {
  51. POTRF_U_SINGLE, POTRF_L_SINGLE,
  52. };
  53. #ifdef SMP
  54. static blasint (*potrf_parallel[])(blas_arg_t *, BLASLONG *, BLASLONG *, FLOAT *, FLOAT *, BLASLONG) = {
  55. POTRF_U_PARALLEL, POTRF_L_PARALLEL,
  56. };
  57. #endif
  58. int NAME(char *UPLO, blasint *N, FLOAT *a, blasint *ldA, blasint *Info){
  59. blas_arg_t args;
  60. blasint uplo_arg = *UPLO;
  61. blasint uplo;
  62. blasint info;
  63. FLOAT *buffer;
  64. #ifdef PPC440
  65. extern
  66. #endif
  67. FLOAT *sa, *sb;
  68. PRINT_DEBUG_NAME;
  69. args.n = *N;
  70. args.a = (void *)a;
  71. args.lda = *ldA;
  72. TOUPPER(uplo_arg);
  73. uplo = -1;
  74. if (uplo_arg == 'U') uplo = 0;
  75. if (uplo_arg == 'L') uplo = 1;
  76. info = 0;
  77. if (args.lda < MAX(1,args.n)) info = 4;
  78. if (args.n < 0) info = 2;
  79. if (uplo < 0) info = 1;
  80. if (info) {
  81. BLASFUNC(xerbla)(ERROR_NAME, &info, sizeof(ERROR_NAME) - 1);
  82. *Info = - info;
  83. return 0;
  84. }
  85. *Info = 0;
  86. if (args.n == 0) return 0;
  87. IDEBUG_START;
  88. FUNCTION_PROFILE_START();
  89. #ifndef PPC440
  90. buffer = (FLOAT *)blas_memory_alloc(1);
  91. sa = (FLOAT *)((BLASLONG)buffer + GEMM_OFFSET_A);
  92. sb = (FLOAT *)(((BLASLONG)sa + ((GEMM_P * GEMM_Q * COMPSIZE * SIZE + GEMM_ALIGN) & ~GEMM_ALIGN)) + GEMM_OFFSET_B);
  93. #endif
  94. #ifdef SMP
  95. args.common = NULL;
  96. #ifndef DOUBLE
  97. int nmax = 128;
  98. #else
  99. int nmax = 64;
  100. #endif
  101. if (args.n <nmax) {
  102. args.nthreads = 1;
  103. } else {
  104. args.nthreads = num_cpu_avail(4);
  105. if (args.n/args.nthreads <nmax)
  106. args.nthreads = args.n/nmax;
  107. }
  108. if (args.nthreads == 1) {
  109. #endif
  110. *Info = (potrf_single[uplo])(&args, NULL, NULL, sa, sb, 0);
  111. #ifdef SMP
  112. } else {
  113. *Info = (potrf_parallel[uplo])(&args, NULL, NULL, sa, sb, 0);
  114. }
  115. #endif
  116. #ifndef PPC440
  117. blas_memory_free(buffer);
  118. #endif
  119. FUNCTION_PROFILE_END(1, .5 * args.n * args.n,
  120. args.n * (1./3. + args.n * ( 1./2. + args.n * 1./6.))
  121. + 1./6. * args.n * (args.n * args.n - 1));
  122. IDEBUG_END;
  123. return 0;
  124. }