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
|
; RUN: llc -march=x86-64 -mattr=+sse2 < %s | FileCheck %s
; Verify that we correctly fold target specific packed vector shifts by
; immediate count into a simple build_vector when the elements of the vector
; in input to the packed shift are all constants or undef.
define <8 x i16> @test1() {
%1 = tail call <8 x i16> @llvm.x86.sse2.pslli.w(<8 x i16> <i16 1, i16 2, i16 4, i16 8, i16 1, i16 2, i16 4, i16 8>, i32 3)
ret <8 x i16> %1
}
; CHECK-LABEL: test1
; CHECK-NOT: psll
; CHECK: movaps
; CHECK-NEXT: ret
define <8 x i16> @test2() {
%1 = tail call <8 x i16> @llvm.x86.sse2.psrli.w(<8 x i16> <i16 4, i16 8, i16 16, i16 32, i16 4, i16 8, i16 16, i16 32>, i32 3)
ret <8 x i16> %1
}
; CHECK-LABEL: test2
; CHECK-NOT: psrl
; CHECK: movaps
; CHECK-NEXT: ret
define <8 x i16> @test3() {
%1 = tail call <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16> <i16 4, i16 8, i16 16, i16 32, i16 4, i16 8, i16 16, i16 32>, i32 3)
ret <8 x i16> %1
}
; CHECK-LABEL: test3
; CHECK-NOT: psra
; CHECK: movaps
; CHECK-NEXT: ret
define <4 x i32> @test4() {
%1 = tail call <4 x i32> @llvm.x86.sse2.pslli.d(<4 x i32> <i32 1, i32 2, i32 4, i32 8>, i32 3)
ret <4 x i32> %1
}
; CHECK-LABEL: test4
; CHECK-NOT: psll
; CHECK: movaps
; CHECK-NEXT: ret
define <4 x i32> @test5() {
%1 = tail call <4 x i32> @llvm.x86.sse2.psrli.d(<4 x i32> <i32 4, i32 8, i32 16, i32 32>, i32 3)
ret <4 x i32> %1
}
; CHECK-LABEL: test5
; CHECK-NOT: psrl
; CHECK: movaps
; CHECK-NEXT: ret
define <4 x i32> @test6() {
%1 = tail call <4 x i32> @llvm.x86.sse2.psrai.d(<4 x i32> <i32 4, i32 8, i32 16, i32 32>, i32 3)
ret <4 x i32> %1
}
; CHECK-LABEL: test6
; CHECK-NOT: psra
; CHECK: movaps
; CHECK-NEXT: ret
define <2 x i64> @test7() {
%1 = tail call <2 x i64> @llvm.x86.sse2.pslli.q(<2 x i64> <i64 1, i64 2>, i32 3)
ret <2 x i64> %1
}
; CHECK-LABEL: test7
; CHECK-NOT: psll
; CHECK: movaps
; CHECK-NEXT: ret
define <2 x i64> @test8() {
%1 = tail call <2 x i64> @llvm.x86.sse2.psrli.q(<2 x i64> <i64 8, i64 16>, i32 3)
ret <2 x i64> %1
}
; CHECK-LABEL: test8
; CHECK-NOT: psrl
; CHECK: movaps
; CHECK-NEXT: ret
define <8 x i16> @test9() {
%1 = tail call <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16> <i16 15, i16 8, i16 undef, i16 undef, i16 31, i16 undef, i16 64, i16 128>, i32 3)
ret <8 x i16> %1
}
; CHECK-LABEL: test9
; CHECK-NOT: psra
; CHECK: movaps
; CHECK-NEXT: ret
define <4 x i32> @test10() {
%1 = tail call <4 x i32> @llvm.x86.sse2.psrai.d(<4 x i32> <i32 undef, i32 8, i32 undef, i32 32>, i32 3)
ret <4 x i32> %1
}
; CHECK-LABEL: test10
; CHECK-NOT: psra
; CHECK: movaps
; CHECK-NEXT: ret
define <2 x i64> @test11() {
%1 = tail call <2 x i64> @llvm.x86.sse2.psrli.q(<2 x i64> <i64 undef, i64 31>, i32 3)
ret <2 x i64> %1
}
; CHECK-LABEL: test11
; CHECK-NOT: psrl
; CHECK: movaps
; CHECK-NEXT: ret
define <8 x i16> @test12() {
%1 = tail call <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16> <i16 15, i16 8, i16 undef, i16 undef, i16 31, i16 undef, i16 64, i16 128>, i32 3)
ret <8 x i16> %1
}
; CHECK-LABEL: test12
; CHECK-NOT: psra
; CHECK: movaps
; CHECK-NEXT: ret
define <4 x i32> @test13() {
%1 = tail call <4 x i32> @llvm.x86.sse2.psrli.d(<4 x i32> <i32 undef, i32 8, i32 undef, i32 32>, i32 3)
ret <4 x i32> %1
}
; CHECK-LABEL: test13
; CHECK-NOT: psrl
; CHECK: movaps
; CHECK-NEXT: ret
define <8 x i16> @test14() {
%1 = tail call <8 x i16> @llvm.x86.sse2.psrli.w(<8 x i16> <i16 15, i16 8, i16 undef, i16 undef, i16 31, i16 undef, i16 64, i16 128>, i32 3)
ret <8 x i16> %1
}
; CHECK-LABEL: test14
; CHECK-NOT: psrl
; CHECK: movaps
; CHECK-NEXT: ret
define <4 x i32> @test15() {
%1 = tail call <4 x i32> @llvm.x86.sse2.pslli.d(<4 x i32> <i32 undef, i32 8, i32 undef, i32 32>, i32 3)
ret <4 x i32> %1
}
; CHECK-LABEL: test15
; CHECK-NOT: psll
; CHECK: movaps
; CHECK-NEXT: ret
define <2 x i64> @test16() {
%1 = tail call <2 x i64> @llvm.x86.sse2.pslli.q(<2 x i64> <i64 undef, i64 31>, i32 3)
ret <2 x i64> %1
}
; CHECK-LABEL: test16
; CHECK-NOT: psll
; CHECK: movaps
; CHECK-NEXT: ret
declare <8 x i16> @llvm.x86.sse2.pslli.w(<8 x i16>, i32)
declare <8 x i16> @llvm.x86.sse2.psrli.w(<8 x i16>, i32)
declare <8 x i16> @llvm.x86.sse2.psrai.w(<8 x i16>, i32)
declare <4 x i32> @llvm.x86.sse2.pslli.d(<4 x i32>, i32)
declare <4 x i32> @llvm.x86.sse2.psrli.d(<4 x i32>, i32)
declare <4 x i32> @llvm.x86.sse2.psrai.d(<4 x i32>, i32)
declare <2 x i64> @llvm.x86.sse2.pslli.q(<2 x i64>, i32)
declare <2 x i64> @llvm.x86.sse2.psrli.q(<2 x i64>, i32)
|
