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
126
127
128
129
130
131
132
133
134
135
136
137
138
139
|
/* APPLE LOCAL file AltiVec */
/* { dg-do compile { target powerpc*-*-* } } */
/* For 64-bit we need 64-bit headers. */
/* { dg-xfail-if "" { powerpc*-*-darwin* } { "-m64" } { "" } } */
/* { dg-options "-O3 -finline-limit=9999 -maltivec -Wa,-force_cpusubtype_ALL -fdump-ipa-cgraph -S -faltivec" } */
/* Inliner should inline AltiVec(tm) functions normally when -maltivec is on. */
/* <rdar://problem/3837835> Selective inlining of functions that use Altivec */
#include <Carbon/Carbon.h>
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
#define N 400
#define N4 ((N+3)/4)
#define N8 ((N+7)/8)
typedef union
{
signed short sInt[8];
vector signed short vInt;
} IntegerToVector;
static signed long vIntDotProduct (vector signed short [], vector signed short [], long int);
static int mainInt();
static Ptr getMemory ( size_t amount );
static int
mainInt()
{
long int n = N, n8 = N8, m, i, j;
signed long vDotProduct, sDotProduct;
signed short *sx, *sy;
vector signed short *x, *y;
IntegerToVector *sX, *sY;
sx = (short *) getMemory( 4*(n+3) );
if (sx == nil)
return 0;
sy = (short *) getMemory( 4*(n+3) );
if (sy == nil)
return 0;
x = (vector signed short *) getMemory( n8*16 );
if ( x == nil)
return 0;
y = (vector signed short *) getMemory( n8*16 );
if ( y == nil)
return 0;
sX = (IntegerToVector *) getMemory( n8*16 );
if (sX == nil)
return 0;
sY = (IntegerToVector *) getMemory( n8*16 );
if (sY == nil)
return 0;
for ( i = 0; i < n; i++ )
{
sx[i] = ( signed short ) scalb(( M_PI * ( double ) ( i ) / ( double ) n ), 8) + 0.5;
sy[i] = ( signed short ) scalb(( M_PI * ( double ) ( n - i ) / ( double ) n ), 8) + 0.5;
}
m = n % 8;
if (m != 0)
for (i = n; i < n + 8 - m; i++)
{
sx[i] = 0.0;
sy[i] = 0.0;
}
for ( i = 0; i < n8; i++ )
for ( j = 0; j < 8; j++ )
{
sX[i].sInt[j] = sx[i*8+j];
sY[i].sInt[j] = sy[i*8+j];
}
for ( i = 0; i < n8; i++ )
{
x[i] = sX[i].vInt;
y[i] = sY[i].vInt;
}
vDotProduct = vIntDotProduct ( x, y, n8 );
printf ( "\nVector dot product = %10d\n", (int) vDotProduct );
return 0;
}
static Ptr
getMemory ( size_t amount )
{
Ptr ptr;
ptr = malloc(amount);
if (ptr == nil)
printf ("\nUnable to allocate sufficient memory.");
return (ptr);
}
static signed long
vIntDotProduct ( vector signed short x[], vector signed short y[], long int n )
{
typedef union
{
signed long xElem[4];
vector signed int vWord;
} WordToVector;
long int i;
vector signed int partialProduct, zero = ( vector signed int ) { 0,0,0,0 };
WordToVector sum;
partialProduct = zero;
for ( i = 0; i < n ; i++ )
partialProduct = vec_msums ( x[i], y[i], partialProduct );
sum.vWord = vec_sums( partialProduct, zero);
return sum.xElem[3];
}
int
main()
{
mainInt();
exit(0);
}
/* { dg-final { global compiler_flags; if ![string match "*-m64 *" $compiler_flags]
{ scan-tree-dump-times "callee has AltiVec" 0 "cgraph" } } } */
/* { dg-final { global compiler_flags; if ![string match "*-m64 *" $compiler_flags]
{ scan-assembler-not "vIntDotProduct" } } } */
/* { dg-final { global compiler_flags; if ![string match "*-m64 *" $compiler_flags]
{ scan-assembler-not "mainInt" } } } */
|