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
|
//------------------------------------------------------------------------------
// GB_concat_sparse_template: concatenate a tile into a sparse matrix
//------------------------------------------------------------------------------
// SuiteSparse:GraphBLAS, Timothy A. Davis, (c) 2017-2025, All Rights Reserved.
// SPDX-License-Identifier: Apache-2.0
//------------------------------------------------------------------------------
// The tile A is hypersparse, sparse, or full, not bitmap. If C is iso, then
// so is A, and the values are not copied here.
{
//--------------------------------------------------------------------------
// get C and the tile A
//--------------------------------------------------------------------------
ASSERT (GB_IS_SPARSE (A) || GB_IS_HYPERSPARSE (A) || GB_IS_FULL (A)) ;
#ifndef GB_ISO_CONCAT
const GB_A_TYPE *restrict Ax = (GB_A_TYPE *) A->x ;
GB_C_TYPE *restrict Cx = (GB_C_TYPE *) C->x ;
#endif
#ifdef GB_JIT_KERNEL
int64_t avlen = A->vlen ;
GB_Ap_DECLARE (Ap, const) ; GB_Ap_PTR (Ap, A) ;
GB_Ah_DECLARE (Ah, const) ; GB_Ah_PTR (Ah, A) ;
GB_Ai_DECLARE (Ai, const) ; GB_Ai_PTR (Ai, A) ;
GB_Ci_DECLARE (Ci, ) ; GB_Ci_PTR (Ci, C) ;
const int64_t *restrict kfirst_Aslice = A_ek_slicing ;
const int64_t *restrict klast_Aslice = A_ek_slicing + A_ntasks ;
const int64_t *restrict pstart_Aslice = A_ek_slicing + A_ntasks * 2 ;
#if GB_Cp_IS_32
const uint32_t *restrict W = W_parameter ;
#else
const uint64_t *restrict W = W_parameter ;
#endif
#endif
//--------------------------------------------------------------------------
// copy the tile A into C
//--------------------------------------------------------------------------
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_A (Ah, k) ;
const int64_t pC_start = GB_IGET (W, j) ;
//------------------------------------------------------------------
// find the part of the kth vector A(:,j) for this task
//------------------------------------------------------------------
const int64_t p0 = GBp_A (Ap, k, avlen) ;
GB_GET_PA (pA_start, pA_end, tid, k, kfirst, klast, pstart_Aslice,
p0, GBp_A (Ap, k+1, avlen)) ;
//------------------------------------------------------------------
// append A(:,j) onto C(:,j)
//------------------------------------------------------------------
GB_PRAGMA_SIMD
for (int64_t pA = pA_start ; pA < pA_end ; pA++)
{
int64_t i = GBi_A (Ai, pA, avlen) ; // i = Ai [pA]
int64_t pC = pC_start + pA - p0 ;
int64_t ci = cistart + i ;
GB_ISET (Ci, pC, ci) ; // Ci [pC] = ci ;
// Cx [pC] = Ax [pA] ;
GB_COPY (pC, pA, A_iso) ;
}
}
}
}
#undef GB_C_TYPE
#undef GB_A_TYPE
#undef GB_ISO_CONCAT
|
