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
|
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -S -passes=instcombine < %s | FileCheck %s
define i16 @foo(i16 %x) {
; CHECK-LABEL: @foo(
; CHECK-NEXT: [[T1:%.*]] = and i16 [[X:%.*]], 255
; CHECK-NEXT: ret i16 [[T1]]
;
%t1 = and i16 %x, 255
%t2 = zext i16 %t1 to i32
%t3 = icmp ult i32 %t2, 255
%t4 = select i1 %t3, i32 %t2, i32 255
%t5 = trunc i32 %t4 to i16
%t6 = and i16 %t5, 255
ret i16 %t6
}
; This contains a min/max pair to clamp a value to 12 bits.
; By analyzing the clamp pattern, we can tell the add doesn't have signed overflow.
define i16 @min_max_clamp(i16 %x) {
; CHECK-LABEL: @min_max_clamp(
; CHECK-NEXT: [[B:%.*]] = call i16 @llvm.smax.i16(i16 [[X:%.*]], i16 -2048)
; CHECK-NEXT: [[D:%.*]] = call i16 @llvm.smin.i16(i16 [[B]], i16 2047)
; CHECK-NEXT: [[E:%.*]] = add nsw i16 [[D]], 1
; CHECK-NEXT: ret i16 [[E]]
;
%a = icmp sgt i16 %x, -2048
%b = select i1 %a, i16 %x, i16 -2048
%c = icmp slt i16 %b, 2047
%d = select i1 %c, i16 %b, i16 2047
%e = add i16 %d, 1
ret i16 %e
}
; Same as above with min/max reversed.
define i16 @min_max_clamp_2(i16 %x) {
; CHECK-LABEL: @min_max_clamp_2(
; CHECK-NEXT: [[B:%.*]] = call i16 @llvm.smin.i16(i16 [[X:%.*]], i16 2047)
; CHECK-NEXT: [[D:%.*]] = call i16 @llvm.smax.i16(i16 [[B]], i16 -2048)
; CHECK-NEXT: [[E:%.*]] = add nsw i16 [[D]], 1
; CHECK-NEXT: ret i16 [[E]]
;
%a = icmp slt i16 %x, 2047
%b = select i1 %a, i16 %x, i16 2047
%c = icmp sgt i16 %b, -2048
%d = select i1 %c, i16 %b, i16 -2048
%e = add i16 %d, 1
ret i16 %e
}
; This contains a min/max pair to clamp a value to 12 bits.
; By analyzing the clamp pattern, we can tell that the second add doesn't
; overflow the original type and can be moved before the extend.
define i32 @min_max_clamp_3(i16 %x) {
; CHECK-LABEL: @min_max_clamp_3(
; CHECK-NEXT: [[B:%.*]] = call i16 @llvm.smax.i16(i16 [[X:%.*]], i16 -2048)
; CHECK-NEXT: [[D:%.*]] = call i16 @llvm.smin.i16(i16 [[B]], i16 2047)
; CHECK-NEXT: [[TMP1:%.*]] = sext i16 [[D]] to i32
; CHECK-NEXT: ret i32 [[TMP1]]
;
%a = icmp sgt i16 %x, -2048
%b = select i1 %a, i16 %x, i16 -2048
%c = icmp slt i16 %b, 2047
%d = select i1 %c, i16 %b, i16 2047
%e = add i16 %d, 1
%f = sext i16 %e to i32
%g = add i32 %f, -1
ret i32 %g
}
; Same as above with min/max order reversed
define i32 @min_max_clamp_4(i16 %x) {
; CHECK-LABEL: @min_max_clamp_4(
; CHECK-NEXT: [[B:%.*]] = call i16 @llvm.smin.i16(i16 [[X:%.*]], i16 2047)
; CHECK-NEXT: [[D:%.*]] = call i16 @llvm.smax.i16(i16 [[B]], i16 -2048)
; CHECK-NEXT: [[TMP1:%.*]] = sext i16 [[D]] to i32
; CHECK-NEXT: ret i32 [[TMP1]]
;
%a = icmp slt i16 %x, 2047
%b = select i1 %a, i16 %x, i16 2047
%c = icmp sgt i16 %b, -2048
%d = select i1 %c, i16 %b, i16 -2048
%e = add i16 %d, 1
%f = sext i16 %e to i32
%g = add i32 %f, -1
ret i32 %g
}
; Intrinsic versions of the above tests.
declare i16 @llvm.smin.i16(i16, i16)
declare i16 @llvm.smax.i16(i16, i16)
define i16 @min_max_clamp_intrinsic(i16 %x) {
; CHECK-LABEL: @min_max_clamp_intrinsic(
; CHECK-NEXT: [[A:%.*]] = call i16 @llvm.smax.i16(i16 [[X:%.*]], i16 -2048)
; CHECK-NEXT: [[B:%.*]] = call i16 @llvm.smin.i16(i16 [[A]], i16 2047)
; CHECK-NEXT: [[C:%.*]] = add nsw i16 [[B]], 1
; CHECK-NEXT: ret i16 [[C]]
;
%a = call i16 @llvm.smax.i16(i16 %x, i16 -2048)
%b = call i16 @llvm.smin.i16(i16 %a, i16 2047)
%c = add i16 %b, 1
ret i16 %c
}
define i16 @min_max_clamp_intrinsic_2(i16 %x) {
; CHECK-LABEL: @min_max_clamp_intrinsic_2(
; CHECK-NEXT: [[A:%.*]] = call i16 @llvm.smin.i16(i16 [[X:%.*]], i16 2047)
; CHECK-NEXT: [[B:%.*]] = call i16 @llvm.smax.i16(i16 [[A]], i16 -2048)
; CHECK-NEXT: [[C:%.*]] = add nsw i16 [[B]], 1
; CHECK-NEXT: ret i16 [[C]]
;
%a = call i16 @llvm.smin.i16(i16 %x, i16 2047)
%b = call i16 @llvm.smax.i16(i16 %a, i16 -2048)
%c = add i16 %b, 1
ret i16 %c
}
define i32 @min_max_clamp_intrinsic_3(i16 %x) {
; CHECK-LABEL: @min_max_clamp_intrinsic_3(
; CHECK-NEXT: [[A:%.*]] = call i16 @llvm.smax.i16(i16 [[X:%.*]], i16 -2048)
; CHECK-NEXT: [[B:%.*]] = call i16 @llvm.smin.i16(i16 [[A]], i16 2047)
; CHECK-NEXT: [[TMP1:%.*]] = sext i16 [[B]] to i32
; CHECK-NEXT: ret i32 [[TMP1]]
;
%a = call i16 @llvm.smax.i16(i16 %x, i16 -2048)
%b = call i16 @llvm.smin.i16(i16 %a, i16 2047)
%c = add i16 %b, 1
%d = sext i16 %c to i32
%e = add i32 %d, -1
ret i32 %e
}
define i32 @min_max_clamp_intrinsic_4(i16 %x) {
; CHECK-LABEL: @min_max_clamp_intrinsic_4(
; CHECK-NEXT: [[A:%.*]] = call i16 @llvm.smin.i16(i16 [[X:%.*]], i16 2047)
; CHECK-NEXT: [[B:%.*]] = call i16 @llvm.smax.i16(i16 [[A]], i16 -2048)
; CHECK-NEXT: [[TMP1:%.*]] = sext i16 [[B]] to i32
; CHECK-NEXT: ret i32 [[TMP1]]
;
%a = call i16 @llvm.smin.i16(i16 %x, i16 2047)
%b = call i16 @llvm.smax.i16(i16 %a, i16 -2048)
%c = add i16 %b, 1
%d = sext i16 %c to i32
%e = add i32 %d, -1
ret i32 %e
}
|
