File: if-conversion.ll

package info (click to toggle)
llvm-toolchain-19 1%3A19.1.7-3~deb12u1
  • links: PTS, VCS
  • area: main
  • in suites: bookworm-proposed-updates
  • size: 1,998,492 kB
  • sloc: cpp: 6,951,680; ansic: 1,486,157; asm: 913,598; python: 232,024; f90: 80,126; objc: 75,281; lisp: 37,276; pascal: 16,990; sh: 10,009; ml: 5,058; perl: 4,724; awk: 3,523; makefile: 3,167; javascript: 2,504; xml: 892; fortran: 664; cs: 573
file content (140 lines) | stat: -rw-r--r-- 4,173 bytes parent folder | download | duplicates (9)
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
; RUN: opt < %s -passes=loop-vectorize,dce,instcombine -force-vector-interleave=1 -force-vector-width=4 -enable-if-conversion -S | FileCheck %s

target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128"

; This is the loop in this example:
;
;int function0(int *a, int *b, int start, int end) {
;
;  for (int i=start; i<end; ++i) {
;    unsigned k = a[i];
;
;    if (a[i] > b[i])   <------ notice the IF inside the loop.
;      k = k * 5 + 3;
;
;    a[i] = k;  <---- K is a phi node that becomes vector-select.
;  }
;}

;CHECK-LABEL: @function0(
;CHECK: load <4 x i32>
;CHECK: icmp sgt <4 x i32>
;CHECK: mul <4 x i32>
;CHECK: add <4 x i32>
;CHECK: select <4 x i1>
;CHECK: ret i32
define i32 @function0(ptr nocapture %a, ptr nocapture %b, i32 %start, i32 %end) nounwind uwtable ssp {
entry:
  %cmp16 = icmp slt i32 %start, %end
  br i1 %cmp16, label %for.body.lr.ph, label %for.end

for.body.lr.ph:
  %0 = sext i32 %start to i64
  br label %for.body

for.body:
  %indvars.iv = phi i64 [ %0, %for.body.lr.ph ], [ %indvars.iv.next, %if.end ]
  %arrayidx = getelementptr inbounds i32, ptr %a, i64 %indvars.iv
  %1 = load i32, ptr %arrayidx, align 4
  %arrayidx4 = getelementptr inbounds i32, ptr %b, i64 %indvars.iv
  %2 = load i32, ptr %arrayidx4, align 4
  %cmp5 = icmp sgt i32 %1, %2
  br i1 %cmp5, label %if.then, label %if.end

if.then:
  %mul = mul i32 %1, 5
  %add = add i32 %mul, 3
  br label %if.end

if.end:
  %k.0 = phi i32 [ %add, %if.then ], [ %1, %for.body ]
  store i32 %k.0, ptr %arrayidx, align 4
  %indvars.iv.next = add i64 %indvars.iv, 1
  %3 = trunc i64 %indvars.iv.next to i32
  %cmp = icmp slt i32 %3, %end
  br i1 %cmp, label %for.body, label %for.end

for.end:
  ret i32 undef
}



; int func(int *A, int n) {
;   unsigned sum = 0;
;   for (int i = 0; i < n; ++i)
;     if (A[i] > 30)
;       sum += A[i] + 2;
;
;   return sum;
; }

;CHECK-LABEL: @reduction_func(
;CHECK: load <4 x i32>
;CHECK: icmp slt <4 x i32>
;CHECK: add <4 x i32>
;CHECK: select <4 x i1>
;CHECK: ret i32
define i32 @reduction_func(ptr nocapture %A, i32 %n) nounwind uwtable readonly ssp {
entry:
  %cmp10 = icmp sgt i32 %n, 0
  br i1 %cmp10, label %for.body, label %for.end

for.body:                                         ; preds = %entry, %for.inc
  %indvars.iv = phi i64 [ %indvars.iv.next, %for.inc ], [ 0, %entry ]
  %sum.011 = phi i32 [ %sum.1, %for.inc ], [ 0, %entry ]
  %arrayidx = getelementptr inbounds i32, ptr %A, i64 %indvars.iv
  %0 = load i32, ptr %arrayidx, align 4
  %cmp1 = icmp sgt i32 %0, 30
  br i1 %cmp1, label %if.then, label %for.inc

if.then:                                          ; preds = %for.body
  %add = add i32 %sum.011, 2
  %add4 = add i32 %add, %0
  br label %for.inc

for.inc:                                          ; preds = %for.body, %if.then
  %sum.1 = phi i32 [ %add4, %if.then ], [ %sum.011, %for.body ]
  %indvars.iv.next = add i64 %indvars.iv, 1
  %lftr.wideiv = trunc i64 %indvars.iv.next to i32
  %exitcond = icmp eq i32 %lftr.wideiv, %n
  br i1 %exitcond, label %for.end, label %for.body

for.end:                                          ; preds = %for.inc, %entry
  %sum.0.lcssa = phi i32 [ 0, %entry ], [ %sum.1, %for.inc ]
  ret i32 %sum.0.lcssa
}

@a = common global [1 x ptr] zeroinitializer, align 8
@c = common global ptr null, align 8

; We use to if convert this loop. This is not safe because there is a trapping
; constant expression.
; PR16729

; Handle PHI with single incoming value having a full mask.
; PR34523

; CHECK-LABEL: PR34523
; CHECK: vector.body

define void @PR34523() {
bb1:
  br label %bb2

bb2:                                             ; preds = %bb4, %bb1
  %i = phi i16 [ undef, %bb1 ], [ %_tmp2, %bb4 ]
  br label %bb3

bb3:                                             ; preds = %bb2
  %_tmp1 = phi ptr [ undef, %bb2 ]
  br label %bb4

bb4:                                             ; preds = %bb3
  %_tmp2 = add i16 %i, 1
  %_tmp3 = icmp slt i16 %_tmp2, 2
  br i1 %_tmp3, label %bb2, label %bb5

bb5:                                             ; preds = %bb4
  unreachable
}