1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py 2; PR34438 3; Loop has a short trip count of 8 iterations. It should be vectorized because no runtime checks or tail loop are necessary. 4; Two cases tested AVX (MaxVF=8 = TripCount) and AVX512 (MaxVF=16 > TripCount) 5 6; RUN: opt < %s -loop-vectorize -mtriple=x86_64-apple-macosx10.8.0 -mcpu=corei7-avx -S | FileCheck %s 7; RUN: opt < %s -loop-vectorize -mtriple=x86_64-apple-macosx10.8.0 -mcpu=skylake-avx512 -S | FileCheck %s 8 9target 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" 10target triple = "x86_64-apple-macosx10.8.0" 11 12define void @small_tc(float* noalias nocapture %A, float* noalias nocapture readonly %B) { 13; CHECK-LABEL: @small_tc( 14; CHECK-NEXT: entry: 15; CHECK-NEXT: br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]] 16; CHECK: vector.ph: 17; CHECK-NEXT: br label [[VECTOR_BODY:%.*]] 18; CHECK: vector.body: 19; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ] 20; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[INDEX]], 0 21; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds float, float* [[B:%.*]], i64 [[TMP0]] 22; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds float, float* [[TMP1]], i32 0 23; CHECK-NEXT: [[TMP3:%.*]] = bitcast float* [[TMP2]] to <8 x float>* 24; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <8 x float>, <8 x float>* [[TMP3]], align 4, !llvm.access.group !0 25; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds float, float* [[A:%.*]], i64 [[TMP0]] 26; CHECK-NEXT: [[TMP5:%.*]] = getelementptr inbounds float, float* [[TMP4]], i32 0 27; CHECK-NEXT: [[TMP6:%.*]] = bitcast float* [[TMP5]] to <8 x float>* 28; CHECK-NEXT: [[WIDE_LOAD1:%.*]] = load <8 x float>, <8 x float>* [[TMP6]], align 4, !llvm.access.group !0 29; CHECK-NEXT: [[TMP7:%.*]] = fadd fast <8 x float> [[WIDE_LOAD]], [[WIDE_LOAD1]] 30; CHECK-NEXT: [[TMP8:%.*]] = bitcast float* [[TMP5]] to <8 x float>* 31; CHECK-NEXT: store <8 x float> [[TMP7]], <8 x float>* [[TMP8]], align 4, !llvm.access.group !0 32; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8 33; CHECK-NEXT: br i1 true, label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], [[LOOP1:!llvm.loop !.*]] 34; CHECK: middle.block: 35; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 8, 8 36; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]] 37; CHECK: scalar.ph: 38; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ 8, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ] 39; CHECK-NEXT: br label [[FOR_BODY:%.*]] 40; CHECK: for.body: 41; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ] 42; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[B]], i64 [[INDVARS_IV]] 43; CHECK-NEXT: [[TMP10:%.*]] = load float, float* [[ARRAYIDX]], align 4, !llvm.access.group !0 44; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds float, float* [[A]], i64 [[INDVARS_IV]] 45; CHECK-NEXT: [[TMP11:%.*]] = load float, float* [[ARRAYIDX2]], align 4, !llvm.access.group !0 46; CHECK-NEXT: [[ADD:%.*]] = fadd fast float [[TMP10]], [[TMP11]] 47; CHECK-NEXT: store float [[ADD]], float* [[ARRAYIDX2]], align 4, !llvm.access.group !0 48; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1 49; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], 8 50; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_END]], label [[FOR_BODY]], [[LOOP3:!llvm.loop !.*]] 51; CHECK: for.end: 52; CHECK-NEXT: ret void 53; 54entry: 55 br label %for.body 56 57for.body: 58 %indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ] 59 %arrayidx = getelementptr inbounds float, float* %B, i64 %indvars.iv 60 %0 = load float, float* %arrayidx, align 4, !llvm.access.group !5 61 %arrayidx2 = getelementptr inbounds float, float* %A, i64 %indvars.iv 62 %1 = load float, float* %arrayidx2, align 4, !llvm.access.group !5 63 %add = fadd fast float %0, %1 64 store float %add, float* %arrayidx2, align 4, !llvm.access.group !5 65 %indvars.iv.next = add nuw nsw i64 %indvars.iv, 1 66 %exitcond = icmp eq i64 %indvars.iv.next, 8 67 br i1 %exitcond, label %for.end, label %for.body, !llvm.loop !4 68 69for.end: 70 ret void 71} 72 73!3 = !{!3, !{!"llvm.loop.parallel_accesses", !5}} 74!4 = !{!4} 75!5 = distinct !{} 76