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
113
114
115
116
117
118
119
120
121
122
123
124
125
|
//==============================================================================
//=== sfmult_vector_1 ==========================================================
//==============================================================================
// SFMULT, Copyright (c) 2009, Timothy A Davis. All Rights Reserved.
// SPDX-License-Identifier: BSD-3-clause
// y = A*x or A'*x where x is a vector
// sfmult_AN_x_1 y = A*x x is n-by-1, y is m-by-1
// sfmult_AT_x_1 y = A'*x x is m-by-1, y is n-by-1
#include "sfmult.h"
//==============================================================================
//=== sfmult_AN_x_1 ============================================================
//==============================================================================
void sfmult_AN_x_1 // y = A*x x is n-by-1 unit stride, y is m-by-1
(
// --- outputs, not initialized on input
double *Yx, // m-by-1
double *Yz, // m-by-1 if Y is complex (TO DO)
// --- inputs, not modified
const Int *Ap, // size n+1 column pointers
const Int *Ai, // size nz = Ap[n] row indices
const double *Ax, // size nz values
const double *Az, // size nz imaginary values if A is complex (TO DO)
Int m, // A is m-by-n
Int n,
const double *Xx, // n-by-1
const double *Xz, // n-by-1 if X complex
int ac, // true: use conj(A), otherwise use A (TO DO)
int xc, // true: use conj(X), otherwise use X (TO DO)
int yc // true: compute conj(Y), otherwise compute Y (TO DO)
)
{
double y [4], x ;
Int p, pend, i, j, i0, i1, i2, i3 ;
for (i = 0 ; i < m ; i++)
{
Yx [i] = 0 ;
}
p = 0 ;
for (j = 0 ; j < n ; j++)
{
pend = Ap [j+1] ;
x = Xx [j] ;
switch ((pend - p) % 4)
{
case 3: Yx [Ai [p]] += Ax [p] * x ; p++ ;
case 2: Yx [Ai [p]] += Ax [p] * x ; p++ ;
case 1: Yx [Ai [p]] += Ax [p] * x ; p++ ;
case 0: ;
}
for ( ; p < pend ; p += 4)
{
i0 = Ai [p ] ;
i1 = Ai [p+1] ;
i2 = Ai [p+2] ;
i3 = Ai [p+3] ;
y [0] = Yx [i0] + Ax [p ] * x ;
y [1] = Yx [i1] + Ax [p+1] * x ;
y [2] = Yx [i2] + Ax [p+2] * x ;
y [3] = Yx [i3] + Ax [p+3] * x ;
Yx [i0] = y [0] ;
Yx [i1] = y [1] ;
Yx [i2] = y [2] ;
Yx [i3] = y [3] ;
}
}
}
//==============================================================================
//=== sfmult_AT_x_1 ============================================================
//==============================================================================
void sfmult_AT_x_1 // y = A'*x x is m-by-1 unit stride, y is n-by-1
(
// --- outputs, not initialized on input
double *Yx, // n-by-1
double *Yz, // n-by-1 if Y is complex (TO DO)
// --- inputs, not modified
const Int *Ap, // size n+1 column pointers
const Int *Ai, // size nz = Ap[n] row indices
const double *Ax, // size nz values
const double *Az, // size nz imaginary values if A is complex (TO DO)
Int m, // A is m-by-n
Int n,
const double *Xx, // m-by-1
const double *Xz, // m-by-1 if X complex
int ac, // true: use conj(A), otherwise use A (TO DO)
int xc, // true: use conj(X), otherwise use X (TO DO)
int yc // true: compute conj(Y), otherwise compute Y (TO DO)
)
{
double y ;
Int p, pend, j ;
p = 0 ;
for (j = 0 ; j < n ; j++)
{
pend = Ap [j+1] ;
y = 0 ;
switch ((pend - p) % 4)
{
case 3: y += Ax [p] * Xx [Ai [p]] ; p++ ;
case 2: y += Ax [p] * Xx [Ai [p]] ; p++ ;
case 1: y += Ax [p] * Xx [Ai [p]] ; p++ ;
case 0: ;
}
for ( ; p < pend ; p += 4)
{
y += Ax [p ] * Xx [Ai [p ]] ;
y += Ax [p+1] * Xx [Ai [p+1]] ;
y += Ax [p+2] * Xx [Ai [p+2]] ;
y += Ax [p+3] * Xx [Ai [p+3]] ;
}
Yx [j] = y ;
}
}
|
