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
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+sse4.1 -O3 | FileCheck %s --check-prefixes=SSE41
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 -O3 | FileCheck %s --check-prefixes=SSE41
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx -O3 | FileCheck %s --check-prefixes=AVX
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx -O3 | FileCheck %s --check-prefixes=AVX
; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512f -O3 | FileCheck %s --check-prefixes=AVX
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f -O3 | FileCheck %s --check-prefixes=AVX
declare <4 x float> @llvm.experimental.constrained.ceil.v4f32(<4 x float>, metadata)
declare <2 x double> @llvm.experimental.constrained.ceil.v2f64(<2 x double>, metadata)
declare <4 x float> @llvm.experimental.constrained.floor.v4f32(<4 x float>, metadata)
declare <2 x double> @llvm.experimental.constrained.floor.v2f64(<2 x double>, metadata)
declare <4 x float> @llvm.experimental.constrained.trunc.v4f32(<4 x float>, metadata)
declare <2 x double> @llvm.experimental.constrained.trunc.v2f64(<2 x double>, metadata)
declare <4 x float> @llvm.experimental.constrained.rint.v4f32(<4 x float>, metadata, metadata)
declare <2 x double> @llvm.experimental.constrained.rint.v2f64(<2 x double>, metadata, metadata)
declare <4 x float> @llvm.experimental.constrained.nearbyint.v4f32(<4 x float>, metadata, metadata)
declare <2 x double> @llvm.experimental.constrained.nearbyint.v2f64(<2 x double>, metadata, metadata)
define <4 x float> @fceilv4f32(<4 x float> %f) #0 {
; SSE41-LABEL: fceilv4f32:
; SSE41: # %bb.0:
; SSE41-NEXT: roundps $10, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: fceilv4f32:
; AVX: # %bb.0:
; AVX-NEXT: vroundps $10, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <4 x float> @llvm.experimental.constrained.ceil.v4f32(
<4 x float> %f, metadata !"fpexcept.strict") #0
ret <4 x float> %res
}
define <2 x double> @fceilv2f64(<2 x double> %f) #0 {
; SSE41-LABEL: fceilv2f64:
; SSE41: # %bb.0:
; SSE41-NEXT: roundpd $10, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: fceilv2f64:
; AVX: # %bb.0:
; AVX-NEXT: vroundpd $10, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <2 x double> @llvm.experimental.constrained.ceil.v2f64(
<2 x double> %f, metadata !"fpexcept.strict") #0
ret <2 x double> %res
}
define <4 x float> @ffloorv4f32(<4 x float> %f) #0 {
; SSE41-LABEL: ffloorv4f32:
; SSE41: # %bb.0:
; SSE41-NEXT: roundps $9, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: ffloorv4f32:
; AVX: # %bb.0:
; AVX-NEXT: vroundps $9, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <4 x float> @llvm.experimental.constrained.floor.v4f32(
<4 x float> %f, metadata !"fpexcept.strict") #0
ret <4 x float> %res
}
define <2 x double> @ffloorv2f64(<2 x double> %f) #0 {
; SSE41-LABEL: ffloorv2f64:
; SSE41: # %bb.0:
; SSE41-NEXT: roundpd $9, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: ffloorv2f64:
; AVX: # %bb.0:
; AVX-NEXT: vroundpd $9, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <2 x double> @llvm.experimental.constrained.floor.v2f64(
<2 x double> %f, metadata !"fpexcept.strict") #0
ret <2 x double> %res
}
define <4 x float> @ftruncv4f32(<4 x float> %f) #0 {
; SSE41-LABEL: ftruncv4f32:
; SSE41: # %bb.0:
; SSE41-NEXT: roundps $11, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: ftruncv4f32:
; AVX: # %bb.0:
; AVX-NEXT: vroundps $11, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <4 x float> @llvm.experimental.constrained.trunc.v4f32(
<4 x float> %f, metadata !"fpexcept.strict") #0
ret <4 x float> %res
}
define <2 x double> @ftruncv2f64(<2 x double> %f) #0 {
; SSE41-LABEL: ftruncv2f64:
; SSE41: # %bb.0:
; SSE41-NEXT: roundpd $11, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: ftruncv2f64:
; AVX: # %bb.0:
; AVX-NEXT: vroundpd $11, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <2 x double> @llvm.experimental.constrained.trunc.v2f64(
<2 x double> %f, metadata !"fpexcept.strict") #0
ret <2 x double> %res
}
define <4 x float> @frintv4f32(<4 x float> %f) #0 {
; SSE41-LABEL: frintv4f32:
; SSE41: # %bb.0:
; SSE41-NEXT: roundps $4, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: frintv4f32:
; AVX: # %bb.0:
; AVX-NEXT: vroundps $4, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <4 x float> @llvm.experimental.constrained.rint.v4f32(
<4 x float> %f,
metadata !"round.dynamic", metadata !"fpexcept.strict") #0
ret <4 x float> %res
}
define <2 x double> @frintv2f64(<2 x double> %f) #0 {
; SSE41-LABEL: frintv2f64:
; SSE41: # %bb.0:
; SSE41-NEXT: roundpd $4, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: frintv2f64:
; AVX: # %bb.0:
; AVX-NEXT: vroundpd $4, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <2 x double> @llvm.experimental.constrained.rint.v2f64(
<2 x double> %f,
metadata !"round.dynamic", metadata !"fpexcept.strict") #0
ret <2 x double> %res
}
define <4 x float> @fnearbyintv4f32(<4 x float> %f) #0 {
; SSE41-LABEL: fnearbyintv4f32:
; SSE41: # %bb.0:
; SSE41-NEXT: roundps $12, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: fnearbyintv4f32:
; AVX: # %bb.0:
; AVX-NEXT: vroundps $12, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <4 x float> @llvm.experimental.constrained.nearbyint.v4f32(
<4 x float> %f,
metadata !"round.dynamic", metadata !"fpexcept.strict") #0
ret <4 x float> %res
}
define <2 x double> @fnearbyintv2f64(<2 x double> %f) #0 {
; SSE41-LABEL: fnearbyintv2f64:
; SSE41: # %bb.0:
; SSE41-NEXT: roundpd $12, %xmm0, %xmm0
; SSE41-NEXT: ret{{[l|q]}}
;
; AVX-LABEL: fnearbyintv2f64:
; AVX: # %bb.0:
; AVX-NEXT: vroundpd $12, %xmm0, %xmm0
; AVX-NEXT: ret{{[l|q]}}
%res = call <2 x double> @llvm.experimental.constrained.nearbyint.v2f64(
<2 x double> %f,
metadata !"round.dynamic", metadata !"fpexcept.strict") #0
ret <2 x double> %res
}
attributes #0 = { strictfp }
|
