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
39 #include <stdio.h>
40 #include <stdlib.h>
41 #include "common.h"
42
43 static const int divide_rule[][2] =
44 {{ 0, 0},
45 { 1, 1}, { 1, 2}, { 1, 3}, { 2, 2},
46 { 1, 5}, { 2, 3}, { 1, 7}, { 2, 4},
47 { 3, 3}, { 2, 5}, { 1, 11}, { 2, 6},
48 { 1, 13}, { 2, 7}, { 3, 5}, { 4, 4},
49 { 1, 17}, { 3, 6}, { 1, 19}, { 4, 5},
50 { 3, 7}, { 2, 11}, { 1, 23}, { 4, 6},
51 { 5, 5}, { 2, 13}, { 3, 9}, { 4, 7},
52 { 1, 29}, { 5, 6}, { 1, 31}, { 4, 8},
53 { 3, 11}, { 2, 17}, { 5, 7}, { 6, 6},
54 { 1, 37}, { 2, 19}, { 3, 13}, { 5, 8},
55 { 1, 41}, { 6, 7}, { 1, 43}, { 4, 11},
56 { 5, 9}, { 2, 23}, { 1, 47}, { 6, 8},
57 { 7, 7}, { 5, 10}, { 3, 17}, { 4, 13},
58 { 1, 53}, { 6, 9}, { 5, 11}, { 7, 8},
59 { 3, 19}, { 2, 29}, { 1, 59}, { 6, 10},
60 { 1, 61}, { 2, 31}, { 7, 9}, { 8, 8},
61 };
62
CNAME(int mode,blas_arg_t * arg,BLASLONG * range_m,BLASLONG * range_n,int (* function)(),void * sa,void * sb,BLASLONG nthreads)63 int CNAME(int mode, blas_arg_t *arg, BLASLONG *range_m, BLASLONG *range_n, int (*function)(), void *sa, void *sb, BLASLONG nthreads) {
64
65 blas_queue_t queue[MAX_CPU_NUMBER];
66
67 BLASLONG range_M[MAX_CPU_NUMBER + 1], range_N[MAX_CPU_NUMBER + 1];
68 BLASLONG procs, total_procs, num_cpu_m, num_cpu_n;
69
70 BLASLONG width, i, j;
71 BLASLONG divM, divN;
72
73 divM = divide_rule[nthreads][0];
74 divN = divide_rule[nthreads][1];
75
76 if (!range_m) {
77 range_M[0] = 0;
78 i = arg -> m;
79 } else {
80 range_M[0] = range_M[0];
81 i = range_M[1] - range_M[0];
82 }
83
84 num_cpu_m = 0;
85
86 while (i > 0){
87
88 width = blas_quickdivide(i + divM - num_cpu_m - 1, divM - num_cpu_m);
89
90 i -= width;
91 if (i < 0) width = width + i;
92
93 range_M[num_cpu_m + 1] = range_M[num_cpu_m] + width;
94
95 num_cpu_m ++;
96 }
97
98 if (!range_n) {
99 range_N[0] = 0;
100 i = arg -> n;
101 } else {
102 range_N[0] = range_n[0];
103 i = range_n[1] - range_n[0];
104 }
105
106 num_cpu_n = 0;
107
108 while (i > 0){
109
110 width = blas_quickdivide(i + divN - num_cpu_n - 1, divN - num_cpu_n);
111
112 i -= width;
113 if (i < 0) width = width + i;
114
115 range_N[num_cpu_n + 1] = range_N[num_cpu_n] + width;
116
117 num_cpu_n ++;
118 }
119
120 procs = 0;
121
122 for (j = 0; j < num_cpu_n; j++) {
123 for (i = 0; i < num_cpu_m; i++) {
124
125 queue[procs].mode = mode;
126 queue[procs].routine = function;
127 queue[procs].args = arg;
128 queue[procs].range_m = &range_M[i];
129 queue[procs].range_n = &range_N[j];
130 queue[procs].sa = NULL;
131 queue[procs].sb = NULL;
132 queue[procs].next = &queue[procs + 1];
133
134 procs ++;
135 }
136 }
137
138 if (procs) {
139 queue[0].sa = sa;
140 queue[0].sb = sb;
141
142 queue[procs - 1].next = NULL;
143
144 exec_blas(procs, queue);
145 }
146
147 return 0;
148 }
149