1; NOTE: Assertions have been autogenerated by utils/update_test_checks.py 2; RUN: opt < %s -loop-vectorize -force-vector-interleave=4 -force-vector-width=4 -force-reduction-intrinsics -dce -instcombine -S | FileCheck %s 3 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" 5 6define i32 @reduction_sum_single(i32* noalias nocapture %A) { 7; CHECK-LABEL: @reduction_sum_single( 8; CHECK-NEXT: entry: 9; CHECK-NEXT: br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]] 10; CHECK: vector.ph: 11; CHECK-NEXT: br label [[VECTOR_BODY:%.*]] 12; CHECK: vector.body: 13; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ] 14; CHECK-NEXT: [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP8:%.*]], [[VECTOR_BODY]] ] 15; CHECK-NEXT: [[VEC_PHI1:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP9:%.*]], [[VECTOR_BODY]] ] 16; CHECK-NEXT: [[VEC_PHI2:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP10:%.*]], [[VECTOR_BODY]] ] 17; CHECK-NEXT: [[VEC_PHI3:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP11:%.*]], [[VECTOR_BODY]] ] 18; CHECK-NEXT: [[TMP0:%.*]] = getelementptr inbounds i32, i32* [[A:%.*]], i64 [[INDEX]] 19; CHECK-NEXT: [[TMP1:%.*]] = bitcast i32* [[TMP0]] to <4 x i32>* 20; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i32>, <4 x i32>* [[TMP1]], align 4 21; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 4 22; CHECK-NEXT: [[TMP3:%.*]] = bitcast i32* [[TMP2]] to <4 x i32>* 23; CHECK-NEXT: [[WIDE_LOAD4:%.*]] = load <4 x i32>, <4 x i32>* [[TMP3]], align 4 24; CHECK-NEXT: [[TMP4:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 8 25; CHECK-NEXT: [[TMP5:%.*]] = bitcast i32* [[TMP4]] to <4 x i32>* 26; CHECK-NEXT: [[WIDE_LOAD5:%.*]] = load <4 x i32>, <4 x i32>* [[TMP5]], align 4 27; CHECK-NEXT: [[TMP6:%.*]] = getelementptr inbounds i32, i32* [[TMP0]], i64 12 28; CHECK-NEXT: [[TMP7:%.*]] = bitcast i32* [[TMP6]] to <4 x i32>* 29; CHECK-NEXT: [[WIDE_LOAD6:%.*]] = load <4 x i32>, <4 x i32>* [[TMP7]], align 4 30; CHECK-NEXT: [[TMP8]] = add <4 x i32> [[VEC_PHI]], [[WIDE_LOAD]] 31; CHECK-NEXT: [[TMP9]] = add <4 x i32> [[VEC_PHI1]], [[WIDE_LOAD4]] 32; CHECK-NEXT: [[TMP10]] = add <4 x i32> [[VEC_PHI2]], [[WIDE_LOAD5]] 33; CHECK-NEXT: [[TMP11]] = add <4 x i32> [[VEC_PHI3]], [[WIDE_LOAD6]] 34; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 16 35; CHECK-NEXT: [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], 256 36; CHECK-NEXT: br i1 [[TMP12]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop !0 37; CHECK: middle.block: 38; CHECK-NEXT: [[BIN_RDX:%.*]] = add <4 x i32> [[TMP9]], [[TMP8]] 39; CHECK-NEXT: [[BIN_RDX7:%.*]] = add <4 x i32> [[TMP10]], [[BIN_RDX]] 40; CHECK-NEXT: [[BIN_RDX8:%.*]] = add <4 x i32> [[TMP11]], [[BIN_RDX7]] 41; CHECK-NEXT: [[TMP13:%.*]] = call i32 @llvm.experimental.vector.reduce.add.v4i32(<4 x i32> [[BIN_RDX8]]) 42; CHECK-NEXT: br i1 true, label [[DOT_CRIT_EDGE:%.*]], label [[SCALAR_PH]] 43; CHECK: scalar.ph: 44; CHECK-NEXT: br label [[DOTLR_PH:%.*]] 45; CHECK: .lr.ph: 46; CHECK-NEXT: br i1 undef, label [[DOT_CRIT_EDGE]], label [[DOTLR_PH]], !llvm.loop !2 47; CHECK: ._crit_edge: 48; CHECK-NEXT: [[SUM_0_LCSSA:%.*]] = phi i32 [ undef, [[DOTLR_PH]] ], [ [[TMP13]], [[MIDDLE_BLOCK]] ] 49; CHECK-NEXT: ret i32 [[SUM_0_LCSSA]] 50; 51entry: 52 br label %.lr.ph 53 54.lr.ph: ; preds = %entry, %.lr.ph 55 %indvars.iv = phi i64 [ %indvars.iv.next, %.lr.ph ], [ 0, %entry ] 56 %sum.02 = phi i32 [ %l7, %.lr.ph ], [ 0, %entry ] 57 %l2 = getelementptr inbounds i32, i32* %A, i64 %indvars.iv 58 %l3 = load i32, i32* %l2, align 4 59 %l7 = add i32 %sum.02, %l3 60 %indvars.iv.next = add i64 %indvars.iv, 1 61 %lftr.wideiv = trunc i64 %indvars.iv.next to i32 62 %exitcond = icmp eq i32 %lftr.wideiv, 256 63 br i1 %exitcond, label %._crit_edge, label %.lr.ph 64 65._crit_edge: ; preds = %.lr.ph 66 %sum.0.lcssa = phi i32 [ %l7, %.lr.ph ] 67 ret i32 %sum.0.lcssa 68} 69