adx-intrinsics.ll
5.77 KB
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
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=corei7 --show-mc-encoding | FileCheck %s --check-prefix=CHECK --check-prefix=NOADX
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=broadwell --show-mc-encoding | FileCheck %s --check-prefix=CHECK --check-prefix=ADX
define i8 @test_addcarry_32_load(i8 %c, i32* %aptr, i32 %b, i8* %ptr) {
; CHECK-LABEL: test_addcarry_32_load:
; CHECK: ## %bb.0:
; CHECK-NEXT: addb $-1, %dil ## encoding: [0x40,0x80,0xc7,0xff]
; CHECK-NEXT: adcl (%rsi), %edx ## encoding: [0x13,0x16]
; CHECK-NEXT: setb %al ## encoding: [0x0f,0x92,0xc0]
; CHECK-NEXT: movl %edx, (%rcx) ## encoding: [0x89,0x11]
; CHECK-NEXT: retq ## encoding: [0xc3]
%a = load i32, i32* %aptr
%ret = call { i8, i32 } @llvm.x86.addcarry.32(i8 %c, i32 %a, i32 %b)
%1 = extractvalue { i8, i32 } %ret, 1
%2 = bitcast i8* %ptr to i32*
store i32 %1, i32* %2, align 1
%3 = extractvalue { i8, i32 } %ret, 0
ret i8 %3
}
define i8 @test_addcarry_32_load2(i8 %c, i32 %a, i32* %bptr, i8* %ptr) {
; CHECK-LABEL: test_addcarry_32_load2:
; CHECK: ## %bb.0:
; CHECK-NEXT: addb $-1, %dil ## encoding: [0x40,0x80,0xc7,0xff]
; CHECK-NEXT: adcl (%rdx), %esi ## encoding: [0x13,0x32]
; CHECK-NEXT: setb %al ## encoding: [0x0f,0x92,0xc0]
; CHECK-NEXT: movl %esi, (%rcx) ## encoding: [0x89,0x31]
; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load i32, i32* %bptr
%ret = call { i8, i32 } @llvm.x86.addcarry.32(i8 %c, i32 %a, i32 %b)
%1 = extractvalue { i8, i32 } %ret, 1
%2 = bitcast i8* %ptr to i32*
store i32 %1, i32* %2, align 1
%3 = extractvalue { i8, i32 } %ret, 0
ret i8 %3
}
declare { i8, i32 } @llvm.x86.addcarry.32(i8, i32, i32)
define i8 @test_addcarry_32(i8 %c, i32 %a, i32 %b, i8* %ptr) {
; CHECK-LABEL: test_addcarry_32:
; CHECK: ## %bb.0:
; CHECK-NEXT: addb $-1, %dil ## encoding: [0x40,0x80,0xc7,0xff]
; CHECK-NEXT: adcl %edx, %esi ## encoding: [0x11,0xd6]
; CHECK-NEXT: setb %al ## encoding: [0x0f,0x92,0xc0]
; CHECK-NEXT: movl %esi, (%rcx) ## encoding: [0x89,0x31]
; CHECK-NEXT: retq ## encoding: [0xc3]
%ret = call { i8, i32 } @llvm.x86.addcarry.32(i8 %c, i32 %a, i32 %b)
%1 = extractvalue { i8, i32 } %ret, 1
%2 = bitcast i8* %ptr to i32*
store i32 %1, i32* %2, align 1
%3 = extractvalue { i8, i32 } %ret, 0
ret i8 %3
}
declare { i8, i64 } @llvm.x86.addcarry.64(i8, i64, i64)
define i8 @test_addcarry_64(i8 %c, i64 %a, i64 %b, i8* %ptr) {
; CHECK-LABEL: test_addcarry_64:
; CHECK: ## %bb.0:
; CHECK-NEXT: addb $-1, %dil ## encoding: [0x40,0x80,0xc7,0xff]
; CHECK-NEXT: adcq %rdx, %rsi ## encoding: [0x48,0x11,0xd6]
; CHECK-NEXT: setb %al ## encoding: [0x0f,0x92,0xc0]
; CHECK-NEXT: movq %rsi, (%rcx) ## encoding: [0x48,0x89,0x31]
; CHECK-NEXT: retq ## encoding: [0xc3]
%ret = call { i8, i64 } @llvm.x86.addcarry.64(i8 %c, i64 %a, i64 %b)
%1 = extractvalue { i8, i64 } %ret, 1
%2 = bitcast i8* %ptr to i64*
store i64 %1, i64* %2, align 1
%3 = extractvalue { i8, i64 } %ret, 0
ret i8 %3
}
declare { i8, i32 } @llvm.x86.subborrow.32(i8, i32, i32)
define i8 @test_subborrow_32(i8 %c, i32 %a, i32 %b, i8* %ptr) {
; CHECK-LABEL: test_subborrow_32:
; CHECK: ## %bb.0:
; CHECK-NEXT: addb $-1, %dil ## encoding: [0x40,0x80,0xc7,0xff]
; CHECK-NEXT: sbbl %edx, %esi ## encoding: [0x19,0xd6]
; CHECK-NEXT: setb %al ## encoding: [0x0f,0x92,0xc0]
; CHECK-NEXT: movl %esi, (%rcx) ## encoding: [0x89,0x31]
; CHECK-NEXT: retq ## encoding: [0xc3]
%ret = call { i8, i32 } @llvm.x86.subborrow.32(i8 %c, i32 %a, i32 %b)
%1 = extractvalue { i8, i32 } %ret, 1
%2 = bitcast i8* %ptr to i32*
store i32 %1, i32* %2, align 1
%3 = extractvalue { i8, i32 } %ret, 0
ret i8 %3
}
declare { i8, i64 } @llvm.x86.subborrow.64(i8, i64, i64)
define i8 @test_subborrow_64(i8 %c, i64 %a, i64 %b, i8* %ptr) {
; CHECK-LABEL: test_subborrow_64:
; CHECK: ## %bb.0:
; CHECK-NEXT: addb $-1, %dil ## encoding: [0x40,0x80,0xc7,0xff]
; CHECK-NEXT: sbbq %rdx, %rsi ## encoding: [0x48,0x19,0xd6]
; CHECK-NEXT: setb %al ## encoding: [0x0f,0x92,0xc0]
; CHECK-NEXT: movq %rsi, (%rcx) ## encoding: [0x48,0x89,0x31]
; CHECK-NEXT: retq ## encoding: [0xc3]
%ret = call { i8, i64 } @llvm.x86.subborrow.64(i8 %c, i64 %a, i64 %b)
%1 = extractvalue { i8, i64 } %ret, 1
%2 = bitcast i8* %ptr to i64*
store i64 %1, i64* %2, align 1
%3 = extractvalue { i8, i64 } %ret, 0
ret i8 %3
}
; Try a version with loads. Previously we crashed on this.
define i32 @load_crash(i64* nocapture readonly %a, i64* nocapture readonly %b, i64* %res) {
; CHECK-LABEL: load_crash:
; CHECK: ## %bb.0:
; CHECK-NEXT: movq (%rdi), %rcx ## encoding: [0x48,0x8b,0x0f]
; CHECK-NEXT: xorl %eax, %eax ## encoding: [0x31,0xc0]
; CHECK-NEXT: addq (%rsi), %rcx ## encoding: [0x48,0x03,0x0e]
; CHECK-NEXT: setb %al ## encoding: [0x0f,0x92,0xc0]
; CHECK-NEXT: movq %rcx, (%rdx) ## encoding: [0x48,0x89,0x0a]
; CHECK-NEXT: retq ## encoding: [0xc3]
%1 = load i64, i64* %a, align 8
%2 = load i64, i64* %b, align 8
%3 = bitcast i64* %res to i8*
%4 = call { i8, i64 } @llvm.x86.addcarry.64(i8 0, i64 %1, i64 %2)
%5 = extractvalue { i8, i64 } %4, 1
%6 = bitcast i8* %3 to i64*
store i64 %5, i64* %6, align 1
%7 = extractvalue { i8, i64 } %4, 0
%conv = zext i8 %7 to i32
ret i32 %conv
}
; Try a really simple all zero input case, which also used to crash
define void @allzeros() {
; CHECK-LABEL: allzeros:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: movq $0, 0 ## encoding: [0x48,0xc7,0x04,0x25,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
; CHECK-NEXT: retq ## encoding: [0xc3]
entry:
%0 = call { i8, i64 } @llvm.x86.addcarry.64(i8 0, i64 0, i64 0)
%1 = extractvalue { i8, i64 } %0, 1
store i64 %1, i64* null, align 1
%2 = extractvalue { i8, i64 } %0, 0
ret void
}