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
140
|
2011-05-02 Ira Rosen <ira.rosen@linaro.org>
Backport from FSF:
2011-03-27 Ira Rosen <ira.rosen@linaro.org>
gcc/
* config/arm/arm.c (arm_autovectorize_vector_sizes): New function.
(TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_SIZES): Define.
gcc/testsuite/
* gcc.dg/vect/vect-outer-5.c: Reduce the distance between data
accesses to preserve the meaning of the test for doubleword vectors.
* gcc.dg/vect/no-vfa-pr29145.c: Likewise.
* gcc.dg/vect/slp-3.c: Reduce the loop bound for the same reason.
=== modified file 'gcc/config/arm/arm.c'
--- old/gcc/config/arm/arm.c 2011-03-02 11:38:43 +0000
+++ new/gcc/config/arm/arm.c 2011-04-28 11:46:58 +0000
@@ -250,6 +250,7 @@
bool is_packed);
static void arm_conditional_register_usage (void);
static reg_class_t arm_preferred_rename_class (reg_class_t rclass);
+static unsigned int arm_autovectorize_vector_sizes (void);
/* Table of machine attributes. */
@@ -395,6 +396,9 @@
#define TARGET_VECTOR_MODE_SUPPORTED_P arm_vector_mode_supported_p
#undef TARGET_VECTORIZE_PREFERRED_SIMD_MODE
#define TARGET_VECTORIZE_PREFERRED_SIMD_MODE arm_preferred_simd_mode
+#undef TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_SIZES
+#define TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_SIZES \
+ arm_autovectorize_vector_sizes
#undef TARGET_MACHINE_DEPENDENT_REORG
#define TARGET_MACHINE_DEPENDENT_REORG arm_reorg
@@ -23511,6 +23515,12 @@
}
}
+static unsigned int
+arm_autovectorize_vector_sizes (void)
+{
+ return TARGET_NEON_VECTORIZE_QUAD ? 16 | 8 : 0;
+}
+
static bool
arm_vector_alignment_reachable (const_tree type, bool is_packed)
{
=== modified file 'gcc/testsuite/gcc.dg/vect/no-vfa-pr29145.c'
--- old/gcc/testsuite/gcc.dg/vect/no-vfa-pr29145.c 2010-10-04 14:59:30 +0000
+++ new/gcc/testsuite/gcc.dg/vect/no-vfa-pr29145.c 2011-04-28 11:46:58 +0000
@@ -8,7 +8,7 @@
void with_restrict(int * __restrict p)
{
int i;
- int *q = p - 2;
+ int *q = p - 1;
for (i = 0; i < 1000; ++i) {
p[i] = q[i];
@@ -19,7 +19,7 @@
void without_restrict(int * p)
{
int i;
- int *q = p - 2;
+ int *q = p - 1;
for (i = 0; i < 1000; ++i) {
p[i] = q[i];
@@ -38,8 +38,8 @@
a[i] = b[i] = i;
}
- with_restrict(a + 2);
- without_restrict(b + 2);
+ with_restrict(a + 1);
+ without_restrict(b + 1);
for (i = 0; i < 1002; ++i) {
if (a[i] != b[i])
=== modified file 'gcc/testsuite/gcc.dg/vect/slp-3.c'
--- old/gcc/testsuite/gcc.dg/vect/slp-3.c 2010-11-22 12:16:52 +0000
+++ new/gcc/testsuite/gcc.dg/vect/slp-3.c 2011-04-28 11:46:58 +0000
@@ -4,9 +4,9 @@
#include <stdarg.h>
#include "tree-vect.h"
-#define N 8
+#define N 12
-unsigned short in[N*8] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63};
+unsigned short in[N*8] = {0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31};
int
main1 ()
@@ -101,7 +101,7 @@
}
/* SLP with unrolling by 8. */
- for (i = 0; i < N/2; i++)
+ for (i = 0; i < N/4; i++)
{
out[i*9] = in[i*9];
out[i*9 + 1] = in[i*9 + 1];
@@ -115,7 +115,7 @@
}
/* check results: */
- for (i = 0; i < N/2; i++)
+ for (i = 0; i < N/4; i++)
{
if (out[i*9] != in[i*9]
|| out[i*9 + 1] != in[i*9 + 1]
=== modified file 'gcc/testsuite/gcc.dg/vect/vect-outer-5.c'
--- old/gcc/testsuite/gcc.dg/vect/vect-outer-5.c 2010-11-22 12:16:52 +0000
+++ new/gcc/testsuite/gcc.dg/vect/vect-outer-5.c 2011-04-28 11:46:58 +0000
@@ -17,7 +17,7 @@
float B[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
float C[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
float D[N] __attribute__ ((__aligned__(__BIGGEST_ALIGNMENT__)));
- float E[4] = {0,1,2,480};
+ float E[4] = {0,480,960,1440};
float s;
int i, j;
@@ -55,7 +55,7 @@
s = 0;
for (j=0; j<N; j+=4)
s += C[j];
- B[i+3] = B[i] + s;
+ B[i+1] = B[i] + s;
}
/* check results: */
|
