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
|
// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 5
// RUN: %clang_cc1 -triple arm64-apple-ios7 -target-feature +neon -ffreestanding -flax-vector-conversions=none -emit-llvm -o - %s | FileCheck %s
// REQUIRES: aarch64-registered-target || arm-registered-target
#include <arm_neon.h>
// CHECK-LABEL: define <2 x double> @rnd5(
// CHECK-SAME: <2 x double> noundef [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[__P0_ADDR_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[__RET_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[REF_TMP_I:%.*]] = alloca <16 x i8>, align 16
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: store <2 x double> [[A]], ptr [[A_ADDR]], align 16
// CHECK-NEXT: [[TMP0:%.*]] = load <2 x double>, ptr [[A_ADDR]], align 16
// CHECK-NEXT: store <2 x double> [[TMP0]], ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[VRNDZ_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <2 x double>
// CHECK-NEXT: [[VRNDZ1_I:%.*]] = call <2 x double> @llvm.trunc.v2f64(<2 x double> [[VRNDZ_I]])
// CHECK-NEXT: store <2 x double> [[VRNDZ1_I]], ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: store <2 x double> [[TMP2]], ptr [[__RET_I]], align 16
// CHECK-NEXT: [[TMP3:%.*]] = load <2 x double>, ptr [[__RET_I]], align 16
// CHECK-NEXT: ret <2 x double> [[TMP3]]
//
float64x2_t rnd5(float64x2_t a) { return vrndq_f64(a); }
// CHECK-LABEL: define <2 x double> @rnd9(
// CHECK-SAME: <2 x double> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[__P0_ADDR_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[__RET_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[REF_TMP_I:%.*]] = alloca <16 x i8>, align 16
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: store <2 x double> [[A]], ptr [[A_ADDR]], align 16
// CHECK-NEXT: [[TMP0:%.*]] = load <2 x double>, ptr [[A_ADDR]], align 16
// CHECK-NEXT: store <2 x double> [[TMP0]], ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[VRNDN_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <2 x double>
// CHECK-NEXT: [[VRNDN1_I:%.*]] = call <2 x double> @llvm.roundeven.v2f64(<2 x double> [[VRNDN_I]])
// CHECK-NEXT: store <2 x double> [[VRNDN1_I]], ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: store <2 x double> [[TMP2]], ptr [[__RET_I]], align 16
// CHECK-NEXT: [[TMP3:%.*]] = load <2 x double>, ptr [[__RET_I]], align 16
// CHECK-NEXT: ret <2 x double> [[TMP3]]
//
float64x2_t rnd9(float64x2_t a) { return vrndnq_f64(a); }
// CHECK-LABEL: define <2 x double> @rnd13(
// CHECK-SAME: <2 x double> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[__P0_ADDR_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[__RET_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[REF_TMP_I:%.*]] = alloca <16 x i8>, align 16
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: store <2 x double> [[A]], ptr [[A_ADDR]], align 16
// CHECK-NEXT: [[TMP0:%.*]] = load <2 x double>, ptr [[A_ADDR]], align 16
// CHECK-NEXT: store <2 x double> [[TMP0]], ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[VRNDM_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <2 x double>
// CHECK-NEXT: [[VRNDM1_I:%.*]] = call <2 x double> @llvm.floor.v2f64(<2 x double> [[VRNDM_I]])
// CHECK-NEXT: store <2 x double> [[VRNDM1_I]], ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: store <2 x double> [[TMP2]], ptr [[__RET_I]], align 16
// CHECK-NEXT: [[TMP3:%.*]] = load <2 x double>, ptr [[__RET_I]], align 16
// CHECK-NEXT: ret <2 x double> [[TMP3]]
//
float64x2_t rnd13(float64x2_t a) { return vrndmq_f64(a); }
// CHECK-LABEL: define <2 x double> @rnd18(
// CHECK-SAME: <2 x double> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[__P0_ADDR_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[__RET_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[REF_TMP_I:%.*]] = alloca <16 x i8>, align 16
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: store <2 x double> [[A]], ptr [[A_ADDR]], align 16
// CHECK-NEXT: [[TMP0:%.*]] = load <2 x double>, ptr [[A_ADDR]], align 16
// CHECK-NEXT: store <2 x double> [[TMP0]], ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[VRNDP_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <2 x double>
// CHECK-NEXT: [[VRNDP1_I:%.*]] = call <2 x double> @llvm.ceil.v2f64(<2 x double> [[VRNDP_I]])
// CHECK-NEXT: store <2 x double> [[VRNDP1_I]], ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: store <2 x double> [[TMP2]], ptr [[__RET_I]], align 16
// CHECK-NEXT: [[TMP3:%.*]] = load <2 x double>, ptr [[__RET_I]], align 16
// CHECK-NEXT: ret <2 x double> [[TMP3]]
//
float64x2_t rnd18(float64x2_t a) { return vrndpq_f64(a); }
// CHECK-LABEL: define <2 x double> @rnd22(
// CHECK-SAME: <2 x double> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[__P0_ADDR_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[__RET_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[REF_TMP_I:%.*]] = alloca <16 x i8>, align 16
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: store <2 x double> [[A]], ptr [[A_ADDR]], align 16
// CHECK-NEXT: [[TMP0:%.*]] = load <2 x double>, ptr [[A_ADDR]], align 16
// CHECK-NEXT: store <2 x double> [[TMP0]], ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[VRNDA_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <2 x double>
// CHECK-NEXT: [[VRNDA1_I:%.*]] = call <2 x double> @llvm.round.v2f64(<2 x double> [[VRNDA_I]])
// CHECK-NEXT: store <2 x double> [[VRNDA1_I]], ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: store <2 x double> [[TMP2]], ptr [[__RET_I]], align 16
// CHECK-NEXT: [[TMP3:%.*]] = load <2 x double>, ptr [[__RET_I]], align 16
// CHECK-NEXT: ret <2 x double> [[TMP3]]
//
float64x2_t rnd22(float64x2_t a) { return vrndaq_f64(a); }
// CHECK-LABEL: define <2 x double> @rnd25(
// CHECK-SAME: <2 x double> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
// CHECK-NEXT: [[__P0_ADDR_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[__RET_I:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: [[REF_TMP_I:%.*]] = alloca <16 x i8>, align 16
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca <2 x double>, align 16
// CHECK-NEXT: store <2 x double> [[A]], ptr [[A_ADDR]], align 16
// CHECK-NEXT: [[TMP0:%.*]] = load <2 x double>, ptr [[A_ADDR]], align 16
// CHECK-NEXT: store <2 x double> [[TMP0]], ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[TMP1:%.*]] = load <16 x i8>, ptr [[__P0_ADDR_I]], align 16
// CHECK-NEXT: [[VRNDX_I:%.*]] = bitcast <16 x i8> [[TMP1]] to <2 x double>
// CHECK-NEXT: [[VRNDX1_I:%.*]] = call <2 x double> @llvm.rint.v2f64(<2 x double> [[VRNDX_I]])
// CHECK-NEXT: store <2 x double> [[VRNDX1_I]], ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: [[TMP2:%.*]] = load <2 x double>, ptr [[REF_TMP_I]], align 16
// CHECK-NEXT: store <2 x double> [[TMP2]], ptr [[__RET_I]], align 16
// CHECK-NEXT: [[TMP3:%.*]] = load <2 x double>, ptr [[__RET_I]], align 16
// CHECK-NEXT: ret <2 x double> [[TMP3]]
//
float64x2_t rnd25(float64x2_t a) { return vrndxq_f64(a); }
|
