[llvm] [LV] Transform to handle exits in the scalar loop (PR #148626)

[Graham Hunter via llvm-commits]

================
@@ -0,0 +1,79 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
+; RUN: opt -S < %s -p loop-vectorize -handle-early-exits-in-scalar-tail | FileCheck %s --check-prefixes=CHECK
+
+target triple = "aarch64-unknown-linux-gnu"
+
+define i32 @simple_contains(ptr align 4 dereferenceable(100) readonly %array, i32 %elt) {
+; CHECK-LABEL: define i32 @simple_contains(
+; CHECK-SAME: ptr readonly align 4 dereferenceable(100) [[ARRAY:%.*]], i32 [[ELT:%.*]]) {
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
+; CHECK:       vector.ph:
+; CHECK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> poison, i32 [[ELT]], i64 0
+; CHECK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT]], <4 x i32> poison, <4 x i32> zeroinitializer
+; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr inbounds i32, ptr [[ARRAY]], i32 0
+; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP0]], align 4
+; CHECK-NEXT:    [[TMP1:%.*]] = icmp eq <4 x i32> [[WIDE_LOAD]], [[BROADCAST_SPLAT]]
+; CHECK-NEXT:    [[TMP2:%.*]] = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> [[TMP1]])
+; CHECK-NEXT:    br i1 [[TMP2]], label [[SCALAR_PH]], label [[VECTOR_PH_SPLIT:%.*]]
+; CHECK:       vector.ph.split:
+; CHECK-NEXT:    br label [[FOR_BODY:%.*]]
+; CHECK:       vector.body:
+; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH_SPLIT]] ], [ [[INDEX_NEXT:%.*]], [[FOR_BODY]] ]
+; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
+; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i32, ptr [[ARRAY]], i64 [[INDEX_NEXT]]
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i32, ptr [[TMP3]], i32 0
+; CHECK-NEXT:    [[WIDE_LOAD1:%.*]] = load <4 x i32>, ptr [[TMP4]], align 4
+; CHECK-NEXT:    [[TMP5:%.*]] = icmp eq <4 x i32> [[WIDE_LOAD1]], [[BROADCAST_SPLAT]]
+; CHECK-NEXT:    [[TMP6:%.*]] = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> [[TMP5]])
+; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], 24
+; CHECK-NEXT:    [[TMP8:%.*]] = or i1 [[TMP6]], [[TMP7]]
+; CHECK-NEXT:    br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[FOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK:       middle.block:
+; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT]], 25
----------------
huntergr-arm wrote:

The comparison in the vector body is against the vector tripcount, whereas this is against the original tripcount. It's a bit more obvious in the vplan representation.

https://github.com/llvm/llvm-project/pull/148626