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
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
|
//------------------------------------------------------------------------------
// GB_concat_bitmap_template: concatenate a tile into a bitmap matrix
//------------------------------------------------------------------------------
// SuiteSparse:GraphBLAS, Timothy A. Davis, (c) 2017-2022, All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0
//------------------------------------------------------------------------------
{
//--------------------------------------------------------------------------
// get C and the tile A
//--------------------------------------------------------------------------
#ifndef GB_ISO_CONCAT
const bool A_iso = A->iso ;
const GB_CTYPE *restrict Ax = (GB_CTYPE *) A->x ;
GB_CTYPE *restrict Cx = (GB_CTYPE *) C->x ;
#endif
int8_t *restrict Cb = C->b ;
//--------------------------------------------------------------------------
// copy the tile A into C
//--------------------------------------------------------------------------
switch (GB_sparsity (A))
{
case GxB_FULL : // A is full
{
int A_nthreads = GB_nthreads (anz, chunk, nthreads_max) ;
int64_t pA ;
#pragma omp parallel for num_threads(A_nthreads) schedule(static)
for (pA = 0 ; pA < anz ; pA++)
{
int64_t i = pA % avlen ;
int64_t j = pA / avlen ;
int64_t iC = cistart + i ;
int64_t jC = cvstart + j ;
int64_t pC = iC + jC * cvlen ;
// Cx [pC] = Ax [pA] ;
GB_COPY (pC, pA, A_iso) ;
Cb [pC] = 1 ;
}
}
break ;
case GxB_BITMAP : // A is bitmap
{
int A_nthreads = GB_nthreads (anz, chunk, nthreads_max) ;
const int8_t *restrict Ab = A->b ;
int64_t pA ;
#pragma omp parallel for num_threads(A_nthreads) schedule(static)
for (pA = 0 ; pA < anz ; pA++)
{
if (Ab [pA])
{
int64_t i = pA % avlen ;
int64_t j = pA / avlen ;
int64_t iC = cistart + i ;
int64_t jC = cvstart + j ;
int64_t pC = iC + jC * cvlen ;
// Cx [pC] = Ax [pA] ;
GB_COPY (pC, pA, A_iso) ;
Cb [pC] = 1 ;
}
}
}
break ;
default : // A is sparse or hypersparse
{
int A_nthreads, A_ntasks ;
GB_SLICE_MATRIX (A, 1, chunk) ;
const int64_t *restrict Ap = A->p ;
const int64_t *restrict Ah = A->h ;
const int64_t *restrict Ai = A->i ;
int tid ;
#pragma omp parallel for num_threads(A_nthreads) schedule(static)
for (tid = 0 ; tid < A_ntasks ; tid++)
{
int64_t kfirst = kfirst_Aslice [tid] ;
int64_t klast = klast_Aslice [tid] ;
for (int64_t k = kfirst ; k <= klast ; k++)
{
int64_t j = GBH (Ah, k) ;
int64_t jC = cvstart + j ;
int64_t pC_start = cistart + jC * cvlen ;
int64_t pA_start, pA_end ;
GB_get_pA (&pA_start, &pA_end, tid, k,
kfirst, klast, pstart_Aslice, Ap, avlen) ;
GB_PRAGMA_SIMD
for (int64_t pA = pA_start ; pA < pA_end ; pA++)
{
int64_t i = Ai [pA] ;
int64_t pC = pC_start + i ;
// Cx [pC] = Ax [pA] ;
GB_COPY (pC, pA, A_iso) ;
Cb [pC] = 1 ;
}
}
}
}
break ;
}
done = true ;
}
#undef GB_CTYPE
#undef GB_ISO_CONCAT
|
