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
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefixes=CHECK,SSE
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.2 | FileCheck %s --check-prefixes=CHECK,SSE
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx | FileCheck %s --check-prefixes=CHECK,AVX
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,AVX
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,AVX
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512bw | FileCheck %s --check-prefixes=CHECK,AVX
declare i32 @llvm.ssub.sat.i32 (i32, i32)
declare i64 @llvm.ssub.sat.i64 (i64, i64)
declare <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16>, <8 x i16>)
; fold (ssub_sat x, undef) -> 0
define i32 @combine_undef_i32(i32 %a0) {
; CHECK-LABEL: combine_undef_i32:
; CHECK: # %bb.0:
; CHECK-NEXT: xorl %eax, %eax
; CHECK-NEXT: retq
%res = call i32 @llvm.ssub.sat.i32(i32 %a0, i32 undef)
ret i32 %res
}
define <8 x i16> @combine_undef_v8i16(<8 x i16> %a0) {
; SSE-LABEL: combine_undef_v8i16:
; SSE: # %bb.0:
; SSE-NEXT: xorps %xmm0, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_undef_v8i16:
; AVX: # %bb.0:
; AVX-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX-NEXT: retq
%res = call <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16> undef, <8 x i16> %a0)
ret <8 x i16> %res
}
; fold (ssub_sat c1, c2) -> c3
define i32 @combine_constfold_i32() {
; CHECK-LABEL: combine_constfold_i32:
; CHECK: # %bb.0:
; CHECK-NEXT: movl $-2147483547, %eax # imm = 0x80000065
; CHECK-NEXT: retq
%res = call i32 @llvm.ssub.sat.i32(i32 100, i32 2147483647)
ret i32 %res
}
define <8 x i16> @combine_constfold_v8i16() {
; SSE-LABEL: combine_constfold_v8i16:
; SSE: # %bb.0:
; SSE-NEXT: movaps {{.*#+}} xmm0 = [65535,2,254,0,65534,65282,32786,2]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_constfold_v8i16:
; AVX: # %bb.0:
; AVX-NEXT: vmovaps {{.*#+}} xmm0 = [65535,2,254,0,65534,65282,32786,2]
; AVX-NEXT: retq
%res = call <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16> <i16 0, i16 1, i16 255, i16 65535, i16 -1, i16 -255, i16 -32760, i16 1>, <8 x i16> <i16 1, i16 65535, i16 1, i16 65535, i16 1, i16 65535, i16 -10, i16 65535>)
ret <8 x i16> %res
}
define <8 x i16> @combine_constfold_undef_v8i16() {
; SSE-LABEL: combine_constfold_undef_v8i16:
; SSE: # %bb.0:
; SSE-NEXT: movaps {{.*#+}} xmm0 = [0,0,0,0,65534,65282,32786,2]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_constfold_undef_v8i16:
; AVX: # %bb.0:
; AVX-NEXT: vmovaps {{.*#+}} xmm0 = [0,0,0,0,65534,65282,32786,2]
; AVX-NEXT: retq
%res = call <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16> <i16 undef, i16 1, i16 undef, i16 65535, i16 -1, i16 -255, i16 -32760, i16 1>, <8 x i16> <i16 1, i16 undef, i16 undef, i16 65535, i16 1, i16 65535, i16 -10, i16 65535>)
ret <8 x i16> %res
}
; fold (ssub_sat x, 0) -> x
define i32 @combine_zero_i32(i32 %a0) {
; CHECK-LABEL: combine_zero_i32:
; CHECK: # %bb.0:
; CHECK-NEXT: movl %edi, %eax
; CHECK-NEXT: retq
%1 = call i32 @llvm.ssub.sat.i32(i32 %a0, i32 0)
ret i32 %1
}
define <8 x i16> @combine_zero_v8i16(<8 x i16> %a0) {
; CHECK-LABEL: combine_zero_v8i16:
; CHECK: # %bb.0:
; CHECK-NEXT: retq
%1 = call <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16> %a0, <8 x i16> zeroinitializer)
ret <8 x i16> %1
}
; fold (ssub_sat x, x) -> 0
define i32 @combine_self_i32(i32 %a0) {
; CHECK-LABEL: combine_self_i32:
; CHECK: # %bb.0:
; CHECK-NEXT: xorl %eax, %eax
; CHECK-NEXT: retq
%1 = call i32 @llvm.ssub.sat.i32(i32 %a0, i32 %a0)
ret i32 %1
}
define <8 x i16> @combine_self_v8i16(<8 x i16> %a0) {
; SSE-LABEL: combine_self_v8i16:
; SSE: # %bb.0:
; SSE-NEXT: xorps %xmm0, %xmm0
; SSE-NEXT: retq
;
; AVX-LABEL: combine_self_v8i16:
; AVX: # %bb.0:
; AVX-NEXT: vxorps %xmm0, %xmm0, %xmm0
; AVX-NEXT: retq
%1 = call <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16> %a0, <8 x i16> %a0)
ret <8 x i16> %1
}
; fold (ssub_sat (shuffle x, u, m), (shuffle y, u, m)) -> (shuffle (ssub_sat x, y), u, m)
define <8 x i16> @combine_shuffle_shuffle_v8i16(<8 x i16> %x0, <8 x i16> %y0) {
; SSE-LABEL: combine_shuffle_shuffle_v8i16:
; SSE: # %bb.0:
; SSE-NEXT: psubsw %xmm1, %xmm0
; SSE-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7]
; SSE-NEXT: retq
;
; AVX-LABEL: combine_shuffle_shuffle_v8i16:
; AVX: # %bb.0:
; AVX-NEXT: vpsubsw %xmm1, %xmm0, %xmm0
; AVX-NEXT: vpshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7]
; AVX-NEXT: retq
%x1= shufflevector <8 x i16> %x0, <8 x i16> poison, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7>
%y1 = shufflevector <8 x i16> %y0, <8 x i16> poison, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 4, i32 5, i32 6, i32 7>
%res = tail call <8 x i16> @llvm.ssub.sat.v8i16(<8 x i16> %x1, <8 x i16> %y1)
ret <8 x i16> %res
}
|
