1 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
2 ; RUN: opt < %s -passes=slp-vectorizer -S -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s
4 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"
5 target triple = "x86_64-apple-macosx10.8.0"
7 ; Simple 3-pair chain with loads and stores
8 define void @test1(ptr %a, ptr %b, ptr %c) {
11 ; CHECK-NEXT: [[AGG_TMP_I_I_SROA_0:%.*]] = alloca [3 x double], align 16
12 ; CHECK-NEXT: [[STORE1:%.*]] = getelementptr inbounds [3 x double], ptr [[AGG_TMP_I_I_SROA_0]], i64 0, i64 1
13 ; CHECK-NEXT: [[TMP1:%.*]] = load <2 x double>, ptr [[A:%.*]], align 8
14 ; CHECK-NEXT: [[TMP3:%.*]] = load <2 x double>, ptr [[B:%.*]], align 8
15 ; CHECK-NEXT: [[TMP4:%.*]] = fmul <2 x double> [[TMP1]], [[TMP3]]
16 ; CHECK-NEXT: store <2 x double> [[TMP4]], ptr [[STORE1]], align 8
17 ; CHECK-NEXT: ret void
20 %agg.tmp.i.i.sroa.0 = alloca [3 x double], align 16
21 %i0 = load double, ptr %a
22 %i1 = load double, ptr %b
23 %mul = fmul double %i0, %i1
24 %store1 = getelementptr inbounds [3 x double], ptr %agg.tmp.i.i.sroa.0, i64 0, i64 1
25 %store2 = getelementptr inbounds [3 x double], ptr %agg.tmp.i.i.sroa.0, i64 0, i64 2
26 %arrayidx3 = getelementptr inbounds double, ptr %a, i64 1
27 %i3 = load double, ptr %arrayidx3, align 8
28 %arrayidx4 = getelementptr inbounds double, ptr %b, i64 1
29 %i4 = load double, ptr %arrayidx4, align 8
30 %mul5 = fmul double %i3, %i4
31 store double %mul, ptr %store1
32 store double %mul5, ptr %store2, align 16
36 ; Float has 4 byte abi alignment on x86_64. We must use the alignment of the
37 ; value being loaded/stored not the alignment of the pointer type.
39 define void @test2(ptr %a, ptr %b) {
40 ; CHECK-LABEL: @test2(
42 ; CHECK-NEXT: [[TMP1:%.*]] = load <4 x float>, ptr [[A:%.*]], align 4
43 ; CHECK-NEXT: store <4 x float> [[TMP1]], ptr [[B:%.*]], align 4
44 ; CHECK-NEXT: ret void
47 %l0 = load float, ptr %a
48 %a1 = getelementptr inbounds float, ptr %a, i64 1
49 %l1 = load float, ptr %a1
50 %a2 = getelementptr inbounds float, ptr %a, i64 2
51 %l2 = load float, ptr %a2
52 %a3 = getelementptr inbounds float, ptr %a, i64 3
53 %l3 = load float, ptr %a3
54 store float %l0, ptr %b
55 %b1 = getelementptr inbounds float, ptr %b, i64 1
56 store float %l1, ptr %b1
57 %b2 = getelementptr inbounds float, ptr %b, i64 2
58 store float %l2, ptr %b2
59 %b3 = getelementptr inbounds float, ptr %b, i64 3
60 store float %l3, ptr %b3