1 ; RUN: opt %loadPolly %defaultOpts -polly-codegen -enable-polly-vector -dce -S %s | FileCheck %s
2 ; Obtained from C source as:
3 ; clang -S -emit-llvm -O0 vect-42.c
4 ; opt -correlated-propagation -mem2reg -instcombine -loop-simplify -indvars -instnamer vect-42.s > vect-42.ll
6 ; ModuleID = 'vect-42.s'
7 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"
8 target triple = "x86_64-unknown-linux-gnu"
10 @main.b = internal unnamed_addr constant [256 x float] [float 0.000000e+00, float 3.000000e+00, float 6.000000e+00, float 9.000000e+00, float 1.200000e+01, float 1.500000e+01, float 1.800000e+01, float 2.100000e+01, float 2.400000e+01, float 2.700000e+01, float 3.000000e+01, float 3.300000e+01, float 3.600000e+01, float 3.900000e+01, float 4.200000e+01, float 4.500000e+01, float 4.800000e+01, float 5.100000e+01, float 5.400000e+01, float 5.700000e+01, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00], align 16
11 @main.c = internal unnamed_addr constant [256 x float] [float 0.000000e+00, float 1.000000e+00, float 2.000000e+00, float 3.000000e+00, float 4.000000e+00, float 5.000000e+00, float 6.000000e+00, float 7.000000e+00, float 8.000000e+00, float 9.000000e+00, float 1.000000e+01, float 1.100000e+01, float 1.200000e+01, float 1.300000e+01, float 1.400000e+01, float 1.500000e+01, float 1.600000e+01, float 1.700000e+01, float 1.800000e+01, float 1.900000e+01, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00], align 16
13 define void @bar(float* %pa, float* %pb, float* %pc) nounwind uwtable noinline {
17 for.cond: ; preds = %for.inc, %entry
18 %indvar = phi i64 [ %indvar.next, %for.inc ], [ 0, %entry ]
19 %arrayidx12 = getelementptr float* %pc, i64 %indvar
20 %arrayidx7 = getelementptr float* %pb, i64 %indvar
21 %arrayidx = getelementptr float* %pa, i64 %indvar
22 %i.0 = trunc i64 %indvar to i32
23 %cmp = icmp slt i32 %i.0, 256
24 br i1 %cmp, label %for.body, label %for.end
26 for.body: ; preds = %for.cond
27 %tmp3 = load float* %arrayidx, align 4
28 %tmp8 = load float* %arrayidx7, align 4
29 %tmp13 = load float* %arrayidx12, align 4
30 %mul = fmul float %tmp8, %tmp13
31 %cmp14 = fcmp une float %tmp3, %mul
32 br i1 %cmp14, label %if.then, label %if.end
34 if.then: ; preds = %for.body
35 call void @abort() noreturn nounwind
38 if.end: ; preds = %for.body
41 for.inc: ; preds = %if.end
42 %indvar.next = add i64 %indvar, 1
45 for.end: ; preds = %for.cond
49 declare void @abort() noreturn
51 define i32 @main1(float* noalias %pa, float* %pb, float* %pc) nounwind uwtable noinline {
53 %b = alloca [256 x float], align 16
54 %c = alloca [256 x float], align 16
57 for.cond: ; preds = %for.inc, %entry
58 %indvar1 = phi i64 [ %indvar.next2, %for.inc ], [ 0, %entry ]
59 %arrayidx14 = getelementptr [256 x float]* %c, i64 0, i64 %indvar1
60 %arrayidx10 = getelementptr float* %pc, i64 %indvar1
61 %arrayidx6 = getelementptr [256 x float]* %b, i64 0, i64 %indvar1
62 %arrayidx = getelementptr float* %pb, i64 %indvar1
63 %exitcond3 = icmp ne i64 %indvar1, 256
64 br i1 %exitcond3, label %for.body, label %for.end
66 for.body: ; preds = %for.cond
67 %tmp3 = load float* %arrayidx, align 4
68 store float %tmp3, float* %arrayidx6, align 4
69 %tmp11 = load float* %arrayidx10, align 4
70 store float %tmp11, float* %arrayidx14, align 4
73 for.inc: ; preds = %for.body
74 %indvar.next2 = add i64 %indvar1, 1
77 for.end: ; preds = %for.cond
80 for.cond16: ; preds = %for.inc32, %for.end
81 %indvar = phi i64 [ %indvar.next, %for.inc32 ], [ 0, %for.end ]
82 %arrayidx31 = getelementptr float* %pa, i64 %indvar
83 %arrayidx26 = getelementptr [256 x float]* %c, i64 0, i64 %indvar
84 %arrayidx22 = getelementptr [256 x float]* %b, i64 0, i64 %indvar
85 %exitcond = icmp ne i64 %indvar, 256
86 br i1 %exitcond, label %for.body19, label %for.end35
88 for.body19: ; preds = %for.cond16
89 %tmp23 = load float* %arrayidx22, align 4
90 %tmp27 = load float* %arrayidx26, align 4
91 %mul = fmul float %tmp23, %tmp27
92 store float %mul, float* %arrayidx31, align 4
95 for.inc32: ; preds = %for.body19
96 %indvar.next = add i64 %indvar, 1
99 for.end35: ; preds = %for.cond16
103 define i32 @main() nounwind uwtable {
105 %a = alloca [256 x float], align 16
106 %b = alloca [256 x float], align 16
107 %c = alloca [256 x float], align 16
108 %tmp = bitcast [256 x float]* %b to i8*
109 call void @llvm.memcpy.p0i8.p0i8.i64(i8* %tmp, i8* bitcast ([256 x float]* @main.b to i8*), i64 1024, i32 16, i1 false)
110 %tmp2 = bitcast [256 x float]* %c to i8*
111 call void @llvm.memcpy.p0i8.p0i8.i64(i8* %tmp2, i8* bitcast ([256 x float]* @main.c to i8*), i64 1024, i32 16, i1 false)
112 call void @check_vect()
113 %arraydecay = getelementptr inbounds [256 x float]* %a, i64 0, i64 0
114 %arraydecay3 = getelementptr inbounds [256 x float]* %b, i64 0, i64 0
115 %arraydecay4 = getelementptr inbounds [256 x float]* %c, i64 0, i64 0
116 %call = call i32 @main1(float* %arraydecay, float* %arraydecay3, float* %arraydecay4)
117 %arraydecay5 = getelementptr inbounds [256 x float]* %a, i64 0, i64 0
118 %arraydecay6 = getelementptr inbounds [256 x float]* %b, i64 0, i64 0
119 %arraydecay7 = getelementptr inbounds [256 x float]* %c, i64 0, i64 0
120 call void @bar(float* %arraydecay5, float* %arraydecay6, float* %arraydecay7)
124 declare void @llvm.memcpy.p0i8.p0i8.i64(i8* nocapture, i8* nocapture, i64, i32, i1) nounwind
126 define internal void @check_vect() nounwind uwtable noinline {
128 %a = alloca i32, align 4
129 %b = alloca i32, align 4
130 %c = alloca i32, align 4
131 %d = alloca i32, align 4
132 %call = call void (i32)* (i32, void (i32)*)* @signal(i32 4, void (i32)* @sig_ill_handler) nounwind
133 %call1 = call i32 @__get_cpuid(i32 1, i32* %a, i32* %b, i32* %c, i32* %d)
134 %tobool = icmp eq i32 %call1, 0
135 br i1 %tobool, label %if.then, label %lor.lhs.false
137 lor.lhs.false: ; preds = %entry
138 %tmp4 = load i32* %d, align 4
139 %and6 = and i32 %tmp4, 67108864
140 %cmp = icmp eq i32 %and6, 0
141 br i1 %cmp, label %if.then, label %if.end
143 if.then: ; preds = %entry, %lor.lhs.false
144 call void @exit(i32 0) noreturn nounwind
147 if.end: ; preds = %lor.lhs.false
148 %call7 = call void (i32)* (i32, void (i32)*)* @signal(i32 4, void (i32)* null) nounwind
152 declare void (i32)* @signal(i32, void (i32)*) nounwind
154 define internal void @sig_ill_handler(i32 %sig) nounwind uwtable {
156 call void @exit(i32 0) noreturn nounwind
159 return: ; No predecessors!
163 define internal i32 @__get_cpuid(i32 %__level, i32* %__eax, i32* %__ebx, i32* %__ecx, i32* %__edx) nounwind uwtable inlinehint {
165 %and = and i32 %__level, -2147483648
166 %call = call i32 @__get_cpuid_max(i32 %and, i32* null)
167 %cmp = icmp ult i32 %call, %__level
168 br i1 %cmp, label %if.then, label %if.end
170 if.then: ; preds = %entry
173 if.end: ; preds = %entry
174 %tmp = call { i32, i32, i32, i32 } asm "cpuid\0A\09", "={ax},={bx},={cx},={dx},0,~{dirflag},~{fpsr},~{flags}"(i32 %__level) nounwind, !srcloc !0
175 %asmresult = extractvalue { i32, i32, i32, i32 } %tmp, 0
176 %asmresult8 = extractvalue { i32, i32, i32, i32 } %tmp, 1
177 %asmresult9 = extractvalue { i32, i32, i32, i32 } %tmp, 2
178 %asmresult10 = extractvalue { i32, i32, i32, i32 } %tmp, 3
179 store i32 %asmresult, i32* %__eax, align 4
180 store i32 %asmresult8, i32* %__ebx, align 4
181 store i32 %asmresult9, i32* %__ecx, align 4
182 store i32 %asmresult10, i32* %__edx, align 4
185 return: ; preds = %if.end, %if.then
186 %retval.0 = phi i32 [ 0, %if.then ], [ 1, %if.end ]
190 declare void @exit(i32) noreturn
192 define internal i32 @__get_cpuid_max(i32 %__ext, i32* %__sig) nounwind uwtable inlinehint {
194 %tmp = call { i32, i32, i32, i32 } asm "cpuid\0A\09", "={ax},={bx},={cx},={dx},0,~{dirflag},~{fpsr},~{flags}"(i32 %__ext) nounwind, !srcloc !1
195 %asmresult = extractvalue { i32, i32, i32, i32 } %tmp, 0
196 %tobool = icmp eq i32* %__sig, null
197 br i1 %tobool, label %if.end, label %if.then
199 if.then: ; preds = %entry
200 %asmresult1 = extractvalue { i32, i32, i32, i32 } %tmp, 1
201 store i32 %asmresult1, i32* %__sig, align 4
204 if.end: ; preds = %entry, %if.then
208 !0 = metadata !{i32 -2147342121, i32 -2147342113}
209 !1 = metadata !{i32 -2147342244, i32 -2147342236}