[LoopVectorize] Enable hoisting of runtime checks by default (#71538)

With commit https://reviews.llvm.org/D152366 I introduced functionality
that permitted the hoisting of runtime memory checks from a vectorised
inner loop to the preheader of the next outer-most loop. This is useful
for benchmarks like SPEC2017's x264 where the inner loop is vectorised
and only has a small trip count. In such cases the runtime memory checks
become expensive and since the checks never fail in the case of x264 it
makes sense to do this. However, this behaviour was controlled by the
flag -hoist-runtime-checks which was off by default.

This patch enables this flag by default for all targets, since I believe
this is a generally beneficial thing to do. I have tested this with
SPEC2017 and I see 2.3% and 2.6% improvements with x264 on neoverse-v1
and neoverse-n1, respectively. Similarly, I saw slight improvements in
the overall geomean on both machines. The only other notable changes
were a 1% drop in the roms benchmark, which was compensated for by a 1%
improvement in fotonik3d.

llvm/lib/Analysis/LoopAccessAnalysis.cpp

@@ -143,7 +143,7 @@ static cl::opt<bool, true> HoistRuntimeChecks(
     "hoist-runtime-checks", cl::Hidden,
     cl::desc(
         "Hoist inner loop runtime memory checks to outer loop if possible"),
-    cl::location(VectorizerParams::HoistRuntimeChecks), cl::init(false));
+    cl::location(VectorizerParams::HoistRuntimeChecks), cl::init(true));
 bool VectorizerParams::HoistRuntimeChecks;
 
 bool VectorizerParams::isInterleaveForced() {

llvm/test/Transforms/LoopVectorize/invariant-store-vectorization.ll

@@ -346,74 +346,75 @@ define i32 @multiple_uniform_stores(ptr nocapture %var1, ptr nocapture readonly
 ; CHECK-NEXT:    [[CMP20:%.*]] = icmp eq i32 [[ITR:%.*]], 0
 ; CHECK-NEXT:    br i1 [[CMP20]], label [[FOR_END10:%.*]], label [[FOR_COND1_PREHEADER_PREHEADER:%.*]]
 ; CHECK:       for.cond1.preheader.preheader:
-; CHECK-NEXT:    [[SCEVGEP3:%.*]] = getelementptr i8, ptr [[VAR2:%.*]], i64 4
-; CHECK-NEXT:    [[INVARIANT_GEP5:%.*]] = getelementptr i8, ptr [[VAR1:%.*]], i64 4
+; CHECK-NEXT:    [[TMP0:%.*]] = add i32 [[ITR]], -1
+; CHECK-NEXT:    [[TMP1:%.*]] = zext i32 [[TMP0]] to i64
+; CHECK-NEXT:    [[TMP2:%.*]] = shl nuw nsw i64 [[TMP1]], 2
+; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr i8, ptr [[VAR1:%.*]], i64 [[TMP2]]
+; CHECK-NEXT:    [[SCEVGEP:%.*]] = getelementptr i8, ptr [[TMP3]], i64 4
+; CHECK-NEXT:    [[SCEVGEP2:%.*]] = getelementptr i8, ptr [[VAR2:%.*]], i64 4
 ; CHECK-NEXT:    br label [[FOR_COND1_PREHEADER:%.*]]
 ; CHECK:       for.cond1.preheader:
 ; CHECK-NEXT:    [[INDVARS_IV23:%.*]] = phi i64 [ [[INDVARS_IV_NEXT24:%.*]], [[FOR_INC8:%.*]] ], [ 0, [[FOR_COND1_PREHEADER_PREHEADER]] ]
 ; CHECK-NEXT:    [[J_022:%.*]] = phi i32 [ [[J_1_LCSSA:%.*]], [[FOR_INC8]] ], [ 0, [[FOR_COND1_PREHEADER_PREHEADER]] ]
-; CHECK-NEXT:    [[TMP0:%.*]] = shl nuw nsw i64 [[INDVARS_IV23]], 2
-; CHECK-NEXT:    [[SCEVGEP:%.*]] = getelementptr i8, ptr [[VAR1]], i64 [[TMP0]]
-; CHECK-NEXT:    [[GEP6:%.*]] = getelementptr i8, ptr [[INVARIANT_GEP5]], i64 [[TMP0]]
 ; CHECK-NEXT:    [[CMP218:%.*]] = icmp ult i32 [[J_022]], [[ITR]]
 ; CHECK-NEXT:    br i1 [[CMP218]], label [[FOR_BODY3_LR_PH:%.*]], label [[FOR_INC8]]
 ; CHECK:       for.body3.lr.ph:
 ; CHECK-NEXT:    [[ARRAYIDX5:%.*]] = getelementptr inbounds i32, ptr [[VAR1]], i64 [[INDVARS_IV23]]
-; CHECK-NEXT:    [[TMP1:%.*]] = zext i32 [[J_022]] to i64
+; CHECK-NEXT:    [[TMP4:%.*]] = zext i32 [[J_022]] to i64
 ; CHECK-NEXT:    [[ARRAYIDX5_PROMOTED:%.*]] = load i32, ptr [[ARRAYIDX5]], align 4
-; CHECK-NEXT:    [[TMP2:%.*]] = xor i32 [[J_022]], -1
-; CHECK-NEXT:    [[TMP3:%.*]] = add i32 [[TMP2]], [[ITR]]
-; CHECK-NEXT:    [[TMP4:%.*]] = zext i32 [[TMP3]] to i64
-; CHECK-NEXT:    [[TMP5:%.*]] = add nuw nsw i64 [[TMP4]], 1
-; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[TMP3]], 3
+; CHECK-NEXT:    [[TMP5:%.*]] = xor i32 [[J_022]], -1
+; CHECK-NEXT:    [[TMP6:%.*]] = add i32 [[TMP5]], [[ITR]]
+; CHECK-NEXT:    [[TMP7:%.*]] = zext i32 [[TMP6]] to i64
+; CHECK-NEXT:    [[TMP8:%.*]] = add nuw nsw i64 [[TMP7]], 1
+; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[TMP6]], 3
 ; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_MEMCHECK:%.*]]
 ; CHECK:       vector.memcheck:
-; CHECK-NEXT:    [[TMP6:%.*]] = shl nuw nsw i64 [[TMP1]], 2
-; CHECK-NEXT:    [[SCEVGEP2:%.*]] = getelementptr i8, ptr [[VAR2]], i64 [[TMP6]]
-; CHECK-NEXT:    [[TMP7:%.*]] = xor i32 [[J_022]], -1
-; CHECK-NEXT:    [[TMP8:%.*]] = add i32 [[TMP7]], [[ITR]]
-; CHECK-NEXT:    [[TMP9:%.*]] = zext i32 [[TMP8]] to i64
-; CHECK-NEXT:    [[TMP10:%.*]] = add nuw nsw i64 [[TMP1]], [[TMP9]]
-; CHECK-NEXT:    [[TMP11:%.*]] = shl nuw nsw i64 [[TMP10]], 2
-; CHECK-NEXT:    [[SCEVGEP4:%.*]] = getelementptr i8, ptr [[SCEVGEP3]], i64 [[TMP11]]
-; CHECK-NEXT:    [[BOUND0:%.*]] = icmp ult ptr [[SCEVGEP]], [[SCEVGEP4]]
-; CHECK-NEXT:    [[BOUND1:%.*]] = icmp ult ptr [[SCEVGEP2]], [[GEP6]]
+; CHECK-NEXT:    [[TMP9:%.*]] = shl nuw nsw i64 [[TMP4]], 2
+; CHECK-NEXT:    [[SCEVGEP1:%.*]] = getelementptr i8, ptr [[VAR2]], i64 [[TMP9]]
+; CHECK-NEXT:    [[TMP10:%.*]] = xor i32 [[J_022]], -1
+; CHECK-NEXT:    [[TMP11:%.*]] = add i32 [[TMP10]], [[ITR]]
+; CHECK-NEXT:    [[TMP12:%.*]] = zext i32 [[TMP11]] to i64
+; CHECK-NEXT:    [[TMP13:%.*]] = add nuw nsw i64 [[TMP4]], [[TMP12]]
+; CHECK-NEXT:    [[TMP14:%.*]] = shl nuw nsw i64 [[TMP13]], 2
+; CHECK-NEXT:    [[SCEVGEP3:%.*]] = getelementptr i8, ptr [[SCEVGEP2]], i64 [[TMP14]]
+; CHECK-NEXT:    [[BOUND0:%.*]] = icmp ugt ptr [[SCEVGEP3]], [[VAR1]]
+; CHECK-NEXT:    [[BOUND1:%.*]] = icmp ult ptr [[SCEVGEP1]], [[SCEVGEP]]
 ; CHECK-NEXT:    [[FOUND_CONFLICT:%.*]] = and i1 [[BOUND0]], [[BOUND1]]
 ; CHECK-NEXT:    br i1 [[FOUND_CONFLICT]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
 ; CHECK:       vector.ph:
-; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[TMP5]], 8589934588
-; CHECK-NEXT:    [[IND_END:%.*]] = add nuw nsw i64 [[N_VEC]], [[TMP1]]
-; CHECK-NEXT:    [[TMP12:%.*]] = insertelement <4 x i32> <i32 poison, i32 0, i32 0, i32 0>, i32 [[ARRAYIDX5_PROMOTED]], i64 0
-; CHECK-NEXT:    [[INVARIANT_GEP:%.*]] = getelementptr i32, ptr [[VAR2]], i64 [[TMP1]]
+; CHECK-NEXT:    [[N_VEC:%.*]] = and i64 [[TMP8]], 8589934588
+; CHECK-NEXT:    [[IND_END:%.*]] = add nuw nsw i64 [[N_VEC]], [[TMP4]]
+; CHECK-NEXT:    [[TMP15:%.*]] = insertelement <4 x i32> <i32 poison, i32 0, i32 0, i32 0>, i32 [[ARRAYIDX5_PROMOTED]], i64 0
+; CHECK-NEXT:    [[INVARIANT_GEP:%.*]] = getelementptr i32, ptr [[VAR2]], i64 [[TMP4]]
 ; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
 ; CHECK:       vector.body:
 ; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ [[TMP12]], [[VECTOR_PH]] ], [ [[TMP14:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_PHI:%.*]] = phi <4 x i32> [ [[TMP15]], [[VECTOR_PH]] ], [ [[TMP17:%.*]], [[VECTOR_BODY]] ]
 ; CHECK-NEXT:    [[GEP:%.*]] = getelementptr i32, ptr [[INVARIANT_GEP]], i64 [[INDEX]]
 ; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[GEP]], align 4, !alias.scope !23
-; CHECK-NEXT:    [[TMP13:%.*]] = add <4 x i32> [[VEC_PHI]], [[WIDE_LOAD]]
-; CHECK-NEXT:    [[TMP14]] = add <4 x i32> [[TMP13]], <i32 1, i32 1, i32 1, i32 1>
+; CHECK-NEXT:    [[TMP16:%.*]] = add <4 x i32> [[VEC_PHI]], [[WIDE_LOAD]]
+; CHECK-NEXT:    [[TMP17]] = add <4 x i32> [[TMP16]], <i32 1, i32 1, i32 1, i32 1>
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
-; CHECK-NEXT:    [[TMP15:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT:    br i1 [[TMP15]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP26:![0-9]+]]
+; CHECK-NEXT:    [[TMP18:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT:    br i1 [[TMP18]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP26:![0-9]+]]
 ; CHECK:       middle.block:
-; CHECK-NEXT:    [[DOTLCSSA:%.*]] = phi <4 x i32> [ [[TMP14]], [[VECTOR_BODY]] ]
-; CHECK-NEXT:    [[TMP16:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[DOTLCSSA]])
-; CHECK-NEXT:    store i32 [[TMP16]], ptr [[ARRAYIDX5]], align 4
-; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[TMP5]], [[N_VEC]]
+; CHECK-NEXT:    [[DOTLCSSA:%.*]] = phi <4 x i32> [ [[TMP17]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[TMP19:%.*]] = call i32 @llvm.vector.reduce.add.v4i32(<4 x i32> [[DOTLCSSA]])
+; CHECK-NEXT:    store i32 [[TMP19]], ptr [[ARRAYIDX5]], align 4
+; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[TMP8]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[FOR_INC8_LOOPEXIT:%.*]], label [[SCALAR_PH]]
 ; CHECK:       scalar.ph:
-; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[TMP1]], [[FOR_BODY3_LR_PH]] ], [ [[TMP1]], [[VECTOR_MEMCHECK]] ]
-; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ [[TMP16]], [[MIDDLE_BLOCK]] ], [ [[ARRAYIDX5_PROMOTED]], [[FOR_BODY3_LR_PH]] ], [ [[ARRAYIDX5_PROMOTED]], [[VECTOR_MEMCHECK]] ]
+; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[TMP4]], [[FOR_BODY3_LR_PH]] ], [ [[TMP4]], [[VECTOR_MEMCHECK]] ]
+; CHECK-NEXT:    [[BC_MERGE_RDX:%.*]] = phi i32 [ [[TMP19]], [[MIDDLE_BLOCK]] ], [ [[ARRAYIDX5_PROMOTED]], [[FOR_BODY3_LR_PH]] ], [ [[ARRAYIDX5_PROMOTED]], [[VECTOR_MEMCHECK]] ]
 ; CHECK-NEXT:    br label [[FOR_BODY3:%.*]]
 ; CHECK:       for.body3:
-; CHECK-NEXT:    [[TMP17:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TMP19:%.*]], [[FOR_BODY3]] ]
+; CHECK-NEXT:    [[TMP20:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TMP22:%.*]], [[FOR_BODY3]] ]
 ; CHECK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY3]] ]
 ; CHECK-NEXT:    [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[VAR2]], i64 [[INDVARS_IV]]
-; CHECK-NEXT:    [[TMP18:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
-; CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP17]], [[TMP18]]
-; CHECK-NEXT:    [[TMP19]] = add nsw i32 [[ADD]], 1
-; CHECK-NEXT:    store i32 [[TMP19]], ptr [[ARRAYIDX5]], align 4
+; CHECK-NEXT:    [[TMP21:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
+; CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 [[TMP20]], [[TMP21]]
+; CHECK-NEXT:    [[TMP22]] = add nsw i32 [[ADD]], 1
+; CHECK-NEXT:    store i32 [[TMP22]], ptr [[ARRAYIDX5]], align 4
 ; CHECK-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
 ; CHECK-NEXT:    [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32
 ; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[ITR]]

llvm/test/Transforms/LoopVectorize/multiple-strides-vectorization.ll

@@ -1,5 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt -passes=loop-vectorize -force-vector-width=4 -S < %s | FileCheck %s
+; RUN: opt -passes=loop-vectorize -force-vector-width=4 -hoist-runtime-checks=false -S < %s | FileCheck %s --check-prefix=CHECK
+; RUN: opt -passes=loop-vectorize -force-vector-width=4 -S < %s | FileCheck %s --check-prefix=CHECK-HOIST
 
 ; This is the test case from PR26314.
 ; When we were retrying dependence checking with memchecks only,
@@ -31,29 +32,29 @@ define void @Test(ptr nocapture %obj, i64 %z) #0 {
 ; CHECK-LABEL: @Test(
 ; CHECK-NEXT:    [[TMP1:%.*]] = shl i64 [[Z:%.*]], 2
 ; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[TMP1]], 256
-; CHECK-NEXT:    [[UGLYGEP2:%.*]] = getelementptr i8, ptr [[OBJ:%.*]], i64 [[TMP1]]
+; CHECK-NEXT:    [[SCEVGEP2:%.*]] = getelementptr i8, ptr [[OBJ:%.*]], i64 [[TMP1]]
 ; CHECK-NEXT:    br label [[DOTOUTER_PREHEADER:%.*]]
 ; CHECK:       .outer.preheader:
 ; CHECK-NEXT:    [[I:%.*]] = phi i64 [ 0, [[TMP0:%.*]] ], [ [[I_NEXT:%.*]], [[DOTOUTER:%.*]] ]
 ; CHECK-NEXT:    [[TMP3:%.*]] = shl nuw nsw i64 [[I]], 7
 ; CHECK-NEXT:    [[TMP4:%.*]] = add i64 [[TMP3]], 256
-; CHECK-NEXT:    [[UGLYGEP:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP4]]
+; CHECK-NEXT:    [[SCEVGEP:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP4]]
 ; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[TMP2]], [[TMP3]]
-; CHECK-NEXT:    [[UGLYGEP1:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP5]]
+; CHECK-NEXT:    [[SCEVGEP1:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP5]]
 ; CHECK-NEXT:    [[TMP6:%.*]] = shl nuw nsw i64 [[I]], 2
 ; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[TMP6]], 128
-; CHECK-NEXT:    [[UGLYGEP3:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP7]]
+; CHECK-NEXT:    [[SCEVGEP3:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP7]]
 ; CHECK-NEXT:    [[TMP8:%.*]] = add i64 [[TMP6]], 132
-; CHECK-NEXT:    [[UGLYGEP4:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP8]]
+; CHECK-NEXT:    [[SCEVGEP4:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP8]]
 ; CHECK-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], ptr [[OBJ]], i64 0, i32 1, i64 [[I]]
 ; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[Z]], 4
 ; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_MEMCHECK:%.*]]
 ; CHECK:       vector.memcheck:
-; CHECK-NEXT:    [[BOUND0:%.*]] = icmp ult ptr [[UGLYGEP]], [[UGLYGEP2]]
-; CHECK-NEXT:    [[BOUND1:%.*]] = icmp ult ptr [[OBJ]], [[UGLYGEP1]]
+; CHECK-NEXT:    [[BOUND0:%.*]] = icmp ult ptr [[SCEVGEP]], [[SCEVGEP2]]
+; CHECK-NEXT:    [[BOUND1:%.*]] = icmp ult ptr [[OBJ]], [[SCEVGEP1]]
 ; CHECK-NEXT:    [[FOUND_CONFLICT:%.*]] = and i1 [[BOUND0]], [[BOUND1]]
-; CHECK-NEXT:    [[BOUND05:%.*]] = icmp ult ptr [[UGLYGEP]], [[UGLYGEP4]]
-; CHECK-NEXT:    [[BOUND16:%.*]] = icmp ult ptr [[UGLYGEP3]], [[UGLYGEP1]]
+; CHECK-NEXT:    [[BOUND05:%.*]] = icmp ult ptr [[SCEVGEP]], [[SCEVGEP4]]
+; CHECK-NEXT:    [[BOUND16:%.*]] = icmp ult ptr [[SCEVGEP3]], [[SCEVGEP1]]
 ; CHECK-NEXT:    [[FOUND_CONFLICT7:%.*]] = and i1 [[BOUND05]], [[BOUND16]]
 ; CHECK-NEXT:    [[CONFLICT_RDX:%.*]] = or i1 [[FOUND_CONFLICT]], [[FOUND_CONFLICT7]]
 ; CHECK-NEXT:    br i1 [[CONFLICT_RDX]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
@@ -103,7 +104,74 @@ define void @Test(ptr nocapture %obj, i64 %z) #0 {
 ; CHECK-NEXT:    store i32 [[TMP25]], ptr [[TMP23]], align 4
 ; CHECK-NEXT:    [[J_NEXT]] = add nuw nsw i64 [[J]], 1
 ; CHECK-NEXT:    [[EXITCOND_INNER:%.*]] = icmp eq i64 [[J_NEXT]], [[Z]]
-; CHECK-NEXT:    br i1 [[EXITCOND_INNER]], label [[DOTOUTER]], label [[DOTINNER]], !llvm.loop [[LOOP10:![0-9]+]]
+; CHECK-NEXT:    br i1 [[EXITCOND_INNER]], label [[DOTOUTER]], label [[DOTINNER]], !llvm.loop [[LOOP11:![0-9]+]]
+;
+; CHECK-HOIST-LABEL: @Test(
+; CHECK-HOIST-NEXT:    [[SCEVGEP:%.*]] = getelementptr i8, ptr [[OBJ:%.*]], i64 256
+; CHECK-HOIST-NEXT:    [[TMP1:%.*]] = shl i64 [[Z:%.*]], 2
+; CHECK-HOIST-NEXT:    [[TMP2:%.*]] = add i64 [[TMP1]], 4224
+; CHECK-HOIST-NEXT:    [[SCEVGEP1:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP2]]
+; CHECK-HOIST-NEXT:    [[SCEVGEP2:%.*]] = getelementptr i8, ptr [[OBJ]], i64 [[TMP1]]
+; CHECK-HOIST-NEXT:    [[SCEVGEP3:%.*]] = getelementptr i8, ptr [[OBJ]], i64 128
+; CHECK-HOIST-NEXT:    br label [[DOTOUTER_PREHEADER:%.*]]
+; CHECK-HOIST:       .outer.preheader:
+; CHECK-HOIST-NEXT:    [[I:%.*]] = phi i64 [ 0, [[TMP0:%.*]] ], [ [[I_NEXT:%.*]], [[DOTOUTER:%.*]] ]
+; CHECK-HOIST-NEXT:    [[TMP3:%.*]] = getelementptr inbounds [[STRUCT_S:%.*]], ptr [[OBJ]], i64 0, i32 1, i64 [[I]]
+; CHECK-HOIST-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[Z]], 4
+; CHECK-HOIST-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_MEMCHECK:%.*]]
+; CHECK-HOIST:       vector.memcheck:
+; CHECK-HOIST-NEXT:    [[BOUND0:%.*]] = icmp ult ptr [[SCEVGEP]], [[SCEVGEP2]]
+; CHECK-HOIST-NEXT:    [[BOUND1:%.*]] = icmp ult ptr [[OBJ]], [[SCEVGEP1]]
+; CHECK-HOIST-NEXT:    [[FOUND_CONFLICT:%.*]] = and i1 [[BOUND0]], [[BOUND1]]
+; CHECK-HOIST-NEXT:    [[BOUND14:%.*]] = icmp ult ptr [[SCEVGEP3]], [[SCEVGEP1]]
+; CHECK-HOIST-NEXT:    br i1 [[FOUND_CONFLICT]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
+; CHECK-HOIST:       vector.ph:
+; CHECK-HOIST-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[Z]], 4
+; CHECK-HOIST-NEXT:    [[N_VEC:%.*]] = sub i64 [[Z]], [[N_MOD_VF]]
+; CHECK-HOIST-NEXT:    br label [[VECTOR_BODY:%.*]]
+; CHECK-HOIST:       vector.body:
+; CHECK-HOIST-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-HOIST-NEXT:    [[TMP4:%.*]] = add i64 [[INDEX]], 0
+; CHECK-HOIST-NEXT:    [[TMP5:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[OBJ]], i64 0, i32 0, i64 [[TMP4]]
+; CHECK-HOIST-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, ptr [[TMP5]], i32 0
+; CHECK-HOIST-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP6]], align 4, !alias.scope !0
+; CHECK-HOIST-NEXT:    [[TMP7:%.*]] = load i32, ptr [[TMP3]], align 4, !alias.scope !3
+; CHECK-HOIST-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> poison, i32 [[TMP7]], i64 0
+; CHECK-HOIST-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT]], <4 x i32> poison, <4 x i32> zeroinitializer
+; CHECK-HOIST-NEXT:    [[TMP8:%.*]] = add nsw <4 x i32> [[BROADCAST_SPLAT]], [[WIDE_LOAD]]
+; CHECK-HOIST-NEXT:    [[TMP9:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[OBJ]], i64 0, i32 2, i64 [[I]], i64 [[TMP4]]
+; CHECK-HOIST-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i32, ptr [[TMP9]], i32 0
+; CHECK-HOIST-NEXT:    [[WIDE_LOAD5:%.*]] = load <4 x i32>, ptr [[TMP10]], align 4, !alias.scope !5, !noalias !7
+; CHECK-HOIST-NEXT:    [[TMP11:%.*]] = add nsw <4 x i32> [[TMP8]], [[WIDE_LOAD5]]
+; CHECK-HOIST-NEXT:    store <4 x i32> [[TMP11]], ptr [[TMP10]], align 4, !alias.scope !5, !noalias !7
+; CHECK-HOIST-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
+; CHECK-HOIST-NEXT:    [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-HOIST-NEXT:    br i1 [[TMP12]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
+; CHECK-HOIST:       middle.block:
+; CHECK-HOIST-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[Z]], [[N_VEC]]
+; CHECK-HOIST-NEXT:    br i1 [[CMP_N]], label [[DOTOUTER]], label [[SCALAR_PH]]
+; CHECK-HOIST:       scalar.ph:
+; CHECK-HOIST-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[DOTOUTER_PREHEADER]] ], [ 0, [[VECTOR_MEMCHECK]] ]
+; CHECK-HOIST-NEXT:    br label [[DOTINNER:%.*]]
+; CHECK-HOIST:       .exit:
+; CHECK-HOIST-NEXT:    ret void
+; CHECK-HOIST:       .outer:
+; CHECK-HOIST-NEXT:    [[I_NEXT]] = add nuw nsw i64 [[I]], 1
+; CHECK-HOIST-NEXT:    [[EXITCOND_OUTER:%.*]] = icmp eq i64 [[I_NEXT]], 32
+; CHECK-HOIST-NEXT:    br i1 [[EXITCOND_OUTER]], label [[DOTEXIT:%.*]], label [[DOTOUTER_PREHEADER]]
+; CHECK-HOIST:       .inner:
+; CHECK-HOIST-NEXT:    [[J:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[J_NEXT:%.*]], [[DOTINNER]] ]
+; CHECK-HOIST-NEXT:    [[TMP13:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[OBJ]], i64 0, i32 0, i64 [[J]]
+; CHECK-HOIST-NEXT:    [[TMP14:%.*]] = load i32, ptr [[TMP13]], align 4
+; CHECK-HOIST-NEXT:    [[TMP15:%.*]] = load i32, ptr [[TMP3]], align 4
+; CHECK-HOIST-NEXT:    [[TMP16:%.*]] = add nsw i32 [[TMP15]], [[TMP14]]
+; CHECK-HOIST-NEXT:    [[TMP17:%.*]] = getelementptr inbounds [[STRUCT_S]], ptr [[OBJ]], i64 0, i32 2, i64 [[I]], i64 [[J]]
+; CHECK-HOIST-NEXT:    [[TMP18:%.*]] = load i32, ptr [[TMP17]], align 4
+; CHECK-HOIST-NEXT:    [[TMP19:%.*]] = add nsw i32 [[TMP16]], [[TMP18]]
+; CHECK-HOIST-NEXT:    store i32 [[TMP19]], ptr [[TMP17]], align 4
+; CHECK-HOIST-NEXT:    [[J_NEXT]] = add nuw nsw i64 [[J]], 1
+; CHECK-HOIST-NEXT:    [[EXITCOND_INNER:%.*]] = icmp eq i64 [[J_NEXT]], [[Z]]
+; CHECK-HOIST-NEXT:    br i1 [[EXITCOND_INNER]], label [[DOTOUTER]], label [[DOTINNER]], !llvm.loop [[LOOP11:![0-9]+]]
 ;
   br label %.outer.preheader
 

llvm/test/Transforms/LoopVectorize/runtime-checks-difference.ll

@@ -1,5 +1,5 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
-; RUN: opt %s -passes=loop-vectorize -force-vector-width=4 -force-vector-interleave=1 -S | FileCheck %s
+; RUN: opt %s -passes=loop-vectorize -hoist-runtime-checks=false -force-vector-width=4 -force-vector-interleave=1 -S | FileCheck %s
 
 target datalayout = "e-m:e-i64:64-i128:128-n32:64-S128"
 

llvm/test/Transforms/LoopVectorize/runtime-checks-hoist.ll

@@ -1,6 +1,6 @@
 ; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 2
 ; REQUIRES: asserts
-; RUN: opt < %s -hoist-runtime-checks -p 'loop-vectorize' -force-vector-interleave=1 -S \
+; RUN: opt < %s -p 'loop-vectorize' -force-vector-interleave=1 -S \
 ; RUN:   -force-vector-width=4 -debug-only=loop-accesses,loop-vectorize,loop-utils 2> %t | FileCheck %s
 ; RUN: cat %t | FileCheck %s --check-prefix=DEBUG