; NOTE: Assertions have been autogenerated by utils/update_test_checks.py ; RUN: opt < %s -passes=slp-vectorizer -S | FileCheck %s ; RUN: opt < %s -passes=slp-vectorizer -S -mattr=+avx | FileCheck %s target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128" target triple = "x86_64-apple-macosx10.10.0" define void @testfunc(ptr nocapture %dest, ptr nocapture readonly %src) { ; CHECK-LABEL: @testfunc( ; CHECK-NEXT: entry: ; CHECK-NEXT: br label [[FOR_BODY:%.*]] ; CHECK: for.body: ; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ] ; CHECK-NEXT: [[ACC1_056:%.*]] = phi float [ 0.000000e+00, [[ENTRY]] ], [ [[ADD13:%.*]], [[FOR_BODY]] ] ; CHECK-NEXT: [[TMP0:%.*]] = phi <2 x float> [ zeroinitializer, [[ENTRY]] ], [ [[TMP18:%.*]], [[FOR_BODY]] ] ; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[SRC:%.*]], i64 [[INDVARS_IV]] ; CHECK-NEXT: [[TMP1:%.*]] = load float, ptr [[ARRAYIDX]], align 4 ; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1 ; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds float, ptr [[DEST:%.*]], i64 [[INDVARS_IV]] ; CHECK-NEXT: store float [[ACC1_056]], ptr [[ARRAYIDX2]], align 4 ; CHECK-NEXT: [[TMP2:%.*]] = insertelement <2 x float> poison, float [[TMP1]], i32 0 ; CHECK-NEXT: [[SHUFFLE:%.*]] = shufflevector <2 x float> [[TMP2]], <2 x float> poison, <2 x i32> zeroinitializer ; CHECK-NEXT: [[TMP3:%.*]] = fadd <2 x float> [[TMP0]], [[SHUFFLE]] ; CHECK-NEXT: [[SHUFFLE1:%.*]] = shufflevector <2 x float> [[TMP3]], <2 x float> poison, <2 x i32> ; CHECK-NEXT: [[TMP4:%.*]] = fmul <2 x float> [[TMP0]], zeroinitializer ; CHECK-NEXT: [[TMP5:%.*]] = fadd <2 x float> [[TMP4]], [[SHUFFLE1]] ; CHECK-NEXT: [[TMP6:%.*]] = fcmp olt <2 x float> [[TMP5]], ; CHECK-NEXT: [[TMP7:%.*]] = select <2 x i1> [[TMP6]], <2 x float> [[TMP5]], <2 x float> ; CHECK-NEXT: [[TMP8:%.*]] = fcmp olt <2 x float> [[TMP7]], ; CHECK-NEXT: [[TMP9:%.*]] = fmul <2 x float> [[TMP7]], zeroinitializer ; CHECK-NEXT: [[TMP10:%.*]] = select <2 x i1> [[TMP8]], <2 x float> , <2 x float> [[TMP9]] ; CHECK-NEXT: [[TMP11:%.*]] = extractelement <2 x float> [[TMP10]], i32 0 ; CHECK-NEXT: [[TMP12:%.*]] = extractelement <2 x float> [[TMP10]], i32 1 ; CHECK-NEXT: [[ADD13]] = fadd float [[TMP11]], [[TMP12]] ; CHECK-NEXT: [[TMP13:%.*]] = shufflevector <2 x float> [[TMP10]], <2 x float> poison, <2 x i32> ; CHECK-NEXT: [[TMP14:%.*]] = insertelement <2 x float> [[TMP13]], float [[ADD13]], i32 1 ; CHECK-NEXT: [[TMP15:%.*]] = fcmp olt <2 x float> [[TMP14]], ; CHECK-NEXT: [[TMP16:%.*]] = select <2 x i1> [[TMP15]], <2 x float> [[TMP14]], <2 x float> ; CHECK-NEXT: [[TMP17:%.*]] = fcmp olt <2 x float> [[TMP16]], ; CHECK-NEXT: [[TMP18]] = select <2 x i1> [[TMP17]], <2 x float> , <2 x float> [[TMP16]] ; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], 32 ; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_END:%.*]], label [[FOR_BODY]] ; CHECK: for.end: ; CHECK-NEXT: ret void ; entry: br label %for.body for.body: %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ] %acc1.056 = phi float [ 0.000000e+00, %entry ], [ %add13, %for.body ] %s1.055 = phi float [ 0.000000e+00, %entry ], [ %cond.i40, %for.body ] %s0.054 = phi float [ 0.000000e+00, %entry ], [ %cond.i44, %for.body ] %arrayidx = getelementptr inbounds float, ptr %src, i64 %indvars.iv %0 = load float, ptr %arrayidx, align 4 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1 %arrayidx2 = getelementptr inbounds float, ptr %dest, i64 %indvars.iv store float %acc1.056, ptr %arrayidx2, align 4 %add = fadd float %s0.054, %0 %add3 = fadd float %s1.055, %0 %mul = fmul float %s0.054, 0.000000e+00 %add4 = fadd float %mul, %add3 %mul5 = fmul float %s1.055, 0.000000e+00 %add6 = fadd float %mul5, %add %cmp.i = fcmp olt float %add6, 1.000000e+00 %cond.i = select i1 %cmp.i, float %add6, float 1.000000e+00 %cmp.i51 = fcmp olt float %cond.i, -1.000000e+00 %cmp.i49 = fcmp olt float %add4, 1.000000e+00 %cond.i50 = select i1 %cmp.i49, float %add4, float 1.000000e+00 %cmp.i47 = fcmp olt float %cond.i50, -1.000000e+00 %cond.i.op = fmul float %cond.i, 0.000000e+00 %mul10 = select i1 %cmp.i51, float -0.000000e+00, float %cond.i.op %cond.i50.op = fmul float %cond.i50, 0.000000e+00 %mul11 = select i1 %cmp.i47, float -0.000000e+00, float %cond.i50.op %add13 = fadd float %mul10, %mul11 ; The SLPVectorizer crashed in vectorizeChainsInBlock() because it tried ; to access the second operand of the following cmp after the cmp itself ; was already vectorized and deleted. %cmp.i45 = fcmp olt float %add13, 1.000000e+00 %cond.i46 = select i1 %cmp.i45, float %add13, float 1.000000e+00 %cmp.i43 = fcmp olt float %cond.i46, -1.000000e+00 %cond.i44 = select i1 %cmp.i43, float -1.000000e+00, float %cond.i46 %cmp.i41 = fcmp olt float %mul11, 1.000000e+00 %cond.i42 = select i1 %cmp.i41, float %mul11, float 1.000000e+00 %cmp.i39 = fcmp olt float %cond.i42, -1.000000e+00 %cond.i40 = select i1 %cmp.i39, float -1.000000e+00, float %cond.i42 %exitcond = icmp eq i64 %indvars.iv.next, 32 br i1 %exitcond, label %for.end, label %for.body for.end: ret void }