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 fast-math-pr43074.c
4 ; opt -correlated-propagation -mem2reg -instcombine -loop-simplify -indvars -instnamer fast-math-pr43074.s > fast-math-pr43074.ll
6 ; ModuleID = 'fast-math-pr43074.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 define float @pvslockprocess(float* %fout, float* %fin, i32 %framesize) nounwind uwtable {
14 for.cond: ; preds = %for.inc, %entry
15 %indvar = phi i64 [ %indvar.next, %for.inc ], [ 0, %entry ]
16 %mag.0 = phi float [ 0.000000e+00, %entry ], [ %add, %for.inc ]
17 %tmp = mul i64 %indvar, 2
18 %tmp3 = add i64 %tmp, 1
19 %arrayidx25 = getelementptr float* %fout, i64 %tmp3
20 %arrayidx19 = getelementptr float* %fin, i64 %tmp3
21 %arrayidx14 = getelementptr float* %fout, i64 %tmp
22 %arrayidx = getelementptr float* %fin, i64 %tmp
23 %i.0 = trunc i64 %tmp to i32
24 %cmp = icmp slt i32 %i.0, %framesize
25 br i1 %cmp, label %for.body, label %for.end
27 for.body: ; preds = %for.cond
28 %tmp4 = load float* %arrayidx, align 4
29 %tmp10 = load float* %arrayidx, align 4
30 store float %tmp10, float* %arrayidx14, align 4
31 %tmp20 = load float* %arrayidx19, align 4
32 store float %tmp20, float* %arrayidx25, align 4
35 for.inc: ; preds = %for.body
36 %add = fadd float %mag.0, %tmp4
37 %indvar.next = add i64 %indvar, 1
40 for.end: ; preds = %for.cond
41 %mag.0.lcssa = phi float [ %mag.0, %for.cond ]
42 ret float %mag.0.lcssa