1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2; RUN: opt -slp-vectorizer -slp-vectorize-hor -slp-vectorize-hor-store -S < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=haswell | FileCheck %s
3target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128-ni:1"
4
5define void @test() #0 {
6; CHECK-LABEL: @test(
7; CHECK-NEXT:  entry:
8; CHECK-NEXT:    br label [[LOOP:%.*]]
9; CHECK:       loop:
10; CHECK-NEXT:    [[DUMMY_PHI:%.*]] = phi i64 [ 1, [[ENTRY:%.*]] ], [ [[OP_RDX1:%.*]], [[LOOP]] ]
11; CHECK-NEXT:    [[TMP0:%.*]] = phi i64 [ 2, [[ENTRY]] ], [ [[TMP3:%.*]], [[LOOP]] ]
12; CHECK-NEXT:    [[TMP1:%.*]] = insertelement <4 x i64> poison, i64 [[TMP0]], i32 0
13; CHECK-NEXT:    [[SHUFFLE:%.*]] = shufflevector <4 x i64> [[TMP1]], <4 x i64> poison, <4 x i32> zeroinitializer
14; CHECK-NEXT:    [[TMP2:%.*]] = add <4 x i64> [[SHUFFLE]], <i64 2, i64 3, i64 1, i64 0>
15; CHECK-NEXT:    [[TMP3]] = extractelement <4 x i64> [[TMP2]], i32 3
16; CHECK-NEXT:    [[DUMMY_ADD:%.*]] = add i16 0, 0
17; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <4 x i64> [[TMP2]], i32 1
18; CHECK-NEXT:    [[DUMMY_SHL:%.*]] = shl i64 [[TMP4]], 32
19; CHECK-NEXT:    [[TMP5:%.*]] = add <4 x i64> <i64 1, i64 1, i64 1, i64 1>, [[TMP2]]
20; CHECK-NEXT:    [[TMP6:%.*]] = ashr exact <4 x i64> [[TMP5]], <i64 32, i64 32, i64 32, i64 32>
21; CHECK-NEXT:    [[TMP7:%.*]] = call i64 @llvm.vector.reduce.add.v4i64(<4 x i64> [[TMP6]])
22; CHECK-NEXT:    [[OP_RDX:%.*]] = add i64 [[TMP7]], [[TMP3]]
23; CHECK-NEXT:    [[OP_RDX1]] = add i64 [[OP_RDX]], 0
24; CHECK-NEXT:    br label [[LOOP]]
25;
26entry:
27  br label %loop
28
29loop:
30  %dummy_phi = phi i64 [ 1, %entry ], [ %last, %loop ]
31  %0 = phi i64 [ 2, %entry ], [ %fork, %loop ]
32  %inc1 = add i64 %0, 1
33  %inc2 = add i64 %0, 2
34  %inc11 = add i64 1, %inc1
35  %exact1 = ashr exact i64 %inc11, 32
36  %inc3 = add i64 %0, 3
37  %dummy_add = add i16 0, 0
38  %inc12 = add i64 1, %inc2
39  %exact2 = ashr exact i64 %inc12, 32
40  %dummy_shl = shl i64 %inc3, 32
41  %inc13 = add i64 1, %inc3
42  %exact3 = ashr exact i64 %inc13, 32
43  %fork = add i64 %0, 0
44  %sum1 = add i64 %exact3, %exact2
45  %sum2 = add i64 %sum1, %exact1
46  %zsum = add i64 %sum2, 0
47  %sext22 = add i64 1, %fork
48  %exact4 = ashr exact i64 %sext22, 32
49  %join = add i64 %fork, %zsum
50  %last = add i64 %join, %exact4
51  br label %loop
52}
53
54