brintos

brintos / llvm-project-archived public Read only

0
0
Text · 2.6 KiB · 58bcc3f Raw
63 lines · plain
1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py2; RUN: opt < %s -passes=slp-vectorizer -S -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx | FileCheck %s3 4target 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"5target triple = "x86_64-apple-macosx10.8.0"6 7; Simple 3-pair chain with loads and stores8define void @test1(ptr %a, ptr %b, ptr %c) {9; CHECK-LABEL: @test1(10; CHECK-NEXT:  entry:11; CHECK-NEXT:    [[AGG_TMP_I_I_SROA_0:%.*]] = alloca [3 x double], align 1612; CHECK-NEXT:    [[STORE1:%.*]] = getelementptr inbounds [3 x double], ptr [[AGG_TMP_I_I_SROA_0]], i64 0, i64 113; CHECK-NEXT:    [[TMP1:%.*]] = load <2 x double>, ptr [[A:%.*]], align 814; CHECK-NEXT:    [[TMP3:%.*]] = load <2 x double>, ptr [[B:%.*]], align 815; CHECK-NEXT:    [[TMP4:%.*]] = fmul <2 x double> [[TMP1]], [[TMP3]]16; CHECK-NEXT:    store <2 x double> [[TMP4]], ptr [[STORE1]], align 817; CHECK-NEXT:    ret void18;19entry:20  %agg.tmp.i.i.sroa.0 = alloca [3 x double], align 1621  %i0 = load double, ptr %a22  %i1 = load double, ptr %b23  %mul = fmul double %i0, %i124  %store1 = getelementptr inbounds [3 x double], ptr %agg.tmp.i.i.sroa.0, i64 0, i64 125  %store2 = getelementptr inbounds [3 x double], ptr %agg.tmp.i.i.sroa.0, i64 0, i64 226  %arrayidx3 = getelementptr inbounds double, ptr %a, i64 127  %i3 = load double, ptr %arrayidx3, align 828  %arrayidx4 = getelementptr inbounds double, ptr %b, i64 129  %i4 = load double, ptr %arrayidx4, align 830  %mul5 = fmul double %i3, %i431  store double %mul, ptr %store132  store double %mul5, ptr %store2, align 1633  ret void34}35 36; Float has 4 byte abi alignment on x86_64. We must use the alignment of the37; value being loaded/stored not the alignment of the pointer type.38 39define void @test2(ptr %a, ptr %b) {40; CHECK-LABEL: @test2(41; CHECK-NEXT:  entry:42; CHECK-NEXT:    [[TMP1:%.*]] = load <4 x float>, ptr [[A:%.*]], align 443; CHECK-NEXT:    store <4 x float> [[TMP1]], ptr [[B:%.*]], align 444; CHECK-NEXT:    ret void45;46entry:47  %l0 = load float, ptr %a48  %a1 = getelementptr inbounds float, ptr %a, i64 149  %l1 = load float, ptr %a150  %a2 = getelementptr inbounds float, ptr %a, i64 251  %l2 = load float, ptr %a252  %a3 = getelementptr inbounds float, ptr %a, i64 353  %l3 = load float, ptr %a354  store float %l0, ptr %b55  %b1 = getelementptr inbounds float, ptr %b, i64 156  store float %l1, ptr %b157  %b2 = getelementptr inbounds float, ptr %b, i64 258  store float %l2, ptr %b259  %b3 = getelementptr inbounds float, ptr %b, i64 360  store float %l3, ptr %b361  ret void62}63