[LV] Add epilogue test with variable induction start value.

Add additional test mentioned by @venkataramanan.kumar.llvm in
D92132.

diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/epilog-vectorization-widen-inductions.ll b/llvm/test/Transforms/LoopVectorize/AArch64/epilog-vectorization-widen-inductions.ll
index 6fd6304..44596ef 100644 (file)
--- a/llvm/test/Transforms/LoopVectorize/AArch64/epilog-vectorization-widen-inductions.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/epilog-vectorization-widen-inductions.ll
@@ -1,4 +1,4 @@
-; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --replace-value-regex "!llvm.loop ![0-9]+"
 ; RUN: opt -passes='loop-vectorize' -force-vector-width=2 -enable-epilogue-vectorization -epilogue-vectorization-force-VF=2 -S %s | FileCheck %s
 
 target triple = "arm64-apple-macosx"
@@ -42,7 +42,7 @@ define void @test_widen_ptr_induction(ptr %ptr.start.1) {
 ; CHECK-NEXT:    store <2 x i8> zeroinitializer, ptr [[TMP15]], align 1
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
 ; CHECK-NEXT:    [[TMP16:%.*]] = icmp eq i64 [[INDEX_NEXT]], 10000
-; CHECK-NEXT:    br i1 [[TMP16]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK-NEXT:    br i1 [[TMP16]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 10001, 10000
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
@@ -70,7 +70,7 @@ define void @test_widen_ptr_induction(ptr %ptr.start.1) {
 ; CHECK-NEXT:    store <2 x i8> zeroinitializer, ptr [[TMP24]], align 1
 ; CHECK-NEXT:    [[INDEX_NEXT10]] = add nuw i64 [[INDEX7]], 2
 ; CHECK-NEXT:    [[TMP25:%.*]] = icmp eq i64 [[INDEX_NEXT10]], 10000
-; CHECK-NEXT:    br i1 [[TMP25]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP2:![0-9]+]]
+; CHECK-NEXT:    br i1 [[TMP25]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       vec.epilog.middle.block:
 ; CHECK-NEXT:    [[CMP_N6:%.*]] = icmp eq i64 10001, 10000
 ; CHECK-NEXT:    br i1 [[CMP_N6]], label [[EXIT]], label [[VEC_EPILOG_SCALAR_PH]]
@@ -87,7 +87,7 @@ define void @test_widen_ptr_induction(ptr %ptr.start.1) {
 ; CHECK-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
 ; CHECK-NEXT:    [[PTR_IV_NEXT]] = getelementptr inbounds i8, ptr [[PTR_IV]], i64 1
 ; CHECK-NEXT:    [[EC:%.*]] = icmp eq i64 [[IV]], 10000
-; CHECK-NEXT:    br i1 [[EC]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP4:![0-9]+]]
+; CHECK-NEXT:    br i1 [[EC]], label [[EXIT]], label [[LOOP]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       exit:
 ; CHECK-NEXT:    ret void
 ;
@@ -111,7 +111,7 @@ exit:
 
 declare void @llvm.assume(i1 noundef)
 
-define void @test_widen_induction(i32 %init, ptr %A, i64 %N, i32 %step) {
+define void @test_widen_induction(ptr %A, i64 %N) {
 ; CHECK-LABEL: @test_widen_induction(
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N:%.*]], 4
@@ -135,7 +135,7 @@ define void @test_widen_induction(i32 %init, ptr %A, i64 %N, i32 %step) {
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
 ; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <2 x i64> [[STEP_ADD]], <i64 2, i64 2>
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
+; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
@@ -148,7 +148,7 @@ define void @test_widen_induction(i32 %init, ptr %A, i64 %N, i32 %step) {
 ; CHECK-NEXT:    store i64 [[IV_1]], ptr [[GEP_A]], align 4
 ; CHECK-NEXT:    [[IV_1_NEXT]] = add nuw nsw i64 [[IV_1]], 1
 ; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[IV_1_NEXT]], [[N]]
-; CHECK-NEXT:    br i1 [[EXITCOND]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP6:![0-9]+]]
+; CHECK-NEXT:    br i1 [[EXITCOND]], label [[EXIT]], label [[LOOP]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       exit:
 ; CHECK-NEXT:    ret void
 ;
@@ -167,7 +167,69 @@ exit:
   ret void
 }
 
-define void @test_widen_induction_step_2(i32 %init, ptr %A, i64 %N, i32 %step) {
+define void @test_widen_induction_variable_start(ptr %A, i64 %N, i64 %start) {
+; CHECK-LABEL: @test_widen_induction_variable_start(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[TMP0:%.*]] = sub i64 [[N:%.*]], [[START:%.*]]
+; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 4
+; CHECK-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
+; CHECK:       vector.ph:
+; CHECK-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 4
+; CHECK-NEXT:    [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
+; CHECK-NEXT:    [[IND_END:%.*]] = add i64 [[START]], [[N_VEC]]
+; CHECK-NEXT:    [[DOTSPLATINSERT:%.*]] = insertelement <2 x i64> poison, i64 [[START]], i32 0
+; CHECK-NEXT:    [[DOTSPLAT:%.*]] = shufflevector <2 x i64> [[DOTSPLATINSERT]], <2 x i64> poison, <2 x i32> zeroinitializer
+; CHECK-NEXT:    [[INDUCTION:%.*]] = add <2 x i64> [[DOTSPLAT]], <i64 0, i64 1>
+; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
+; CHECK:       vector.body:
+; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[VEC_IND:%.*]] = phi <2 x i64> [ [[INDUCTION]], [[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], [[VECTOR_BODY]] ]
+; CHECK-NEXT:    [[STEP_ADD:%.*]] = add <2 x i64> [[VEC_IND]], <i64 2, i64 2>
+; CHECK-NEXT:    [[OFFSET_IDX:%.*]] = add i64 [[START]], [[INDEX]]
+; CHECK-NEXT:    [[TMP1:%.*]] = add i64 [[OFFSET_IDX]], 0
+; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[OFFSET_IDX]], 2
+; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i64, ptr [[A:%.*]], i64 [[TMP1]]
+; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[TMP2]]
+; CHECK-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i64, ptr [[TMP3]], i32 0
+; CHECK-NEXT:    store <2 x i64> [[VEC_IND]], ptr [[TMP5]], align 4
+; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i64, ptr [[TMP3]], i32 2
+; CHECK-NEXT:    store <2 x i64> [[STEP_ADD]], ptr [[TMP6]], align 4
+; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
+; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <2 x i64> [[STEP_ADD]], <i64 2, i64 2>
+; CHECK-NEXT:    [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT:    br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], {{!llvm.loop ![0-9]+}}
+; CHECK:       middle.block:
+; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
+; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
+; CHECK:       scalar.ph:
+; CHECK-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[START]], [[ENTRY:%.*]] ]
+; CHECK-NEXT:    br label [[LOOP:%.*]]
+; CHECK:       loop:
+; CHECK-NEXT:    [[IV_1:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_1_NEXT:%.*]], [[LOOP]] ]
+; CHECK-NEXT:    [[GEP_A:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[IV_1]]
+; CHECK-NEXT:    store i64 [[IV_1]], ptr [[GEP_A]], align 4
+; CHECK-NEXT:    [[IV_1_NEXT]] = add nuw nsw i64 [[IV_1]], 1
+; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[IV_1_NEXT]], [[N]]
+; CHECK-NEXT:    br i1 [[EXITCOND]], label [[EXIT]], label [[LOOP]], {{!llvm.loop ![0-9]+}}
+; CHECK:       exit:
+; CHECK-NEXT:    ret void
+;
+entry:
+  br label %loop
+
+loop:
+  %iv.1 = phi i64 [ %start, %entry ], [ %iv.1.next, %loop ]
+  %gep.A = getelementptr inbounds i64, ptr %A, i64 %iv.1
+  store i64 %iv.1, ptr %gep.A, align 4
+  %iv.1.next = add nuw nsw i64 %iv.1, 1
+  %exitcond = icmp eq i64 %iv.1.next, %N
+  br i1 %exitcond, label %exit, label %loop
+
+exit:
+  ret void
+}
+
+define void @test_widen_induction_step_2(ptr %A, i64 %N, i32 %step) {
 ; CHECK-LABEL: @test_widen_induction_step_2(
 ; CHECK-NEXT:  entry:
 ; CHECK-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N:%.*]], 4
@@ -193,7 +255,7 @@ define void @test_widen_induction_step_2(i32 %init, ptr %A, i64 %N, i32 %step) {
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
 ; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <2 x i64> [[STEP_ADD]], <i64 2, i64 2>
 ; CHECK-NEXT:    [[TMP8:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[IND_END]]
-; CHECK-NEXT:    br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
+; CHECK-NEXT:    br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[N]], [[IND_END]]
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
@@ -210,7 +272,7 @@ define void @test_widen_induction_step_2(i32 %init, ptr %A, i64 %N, i32 %step) {
 ; CHECK-NEXT:    [[IV_1_NEXT]] = add nuw nsw i64 [[IV_1]], 1
 ; CHECK-NEXT:    [[IV_2_NEXT:%.*]] = add nuw nsw i64 [[IV_1]], 2
 ; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[IV_1_NEXT]], [[N]]
-; CHECK-NEXT:    br i1 [[EXITCOND]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP8:![0-9]+]]
+; CHECK-NEXT:    br i1 [[EXITCOND]], label [[EXIT]], label [[LOOP]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       exit:
 ; CHECK-NEXT:    ret void
 ;
@@ -258,7 +320,7 @@ define void @test_widen_extended_induction(ptr %dst) {
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i32 [[INDEX]], 4
 ; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <2 x i8> [[STEP_ADD]], <i8 2, i8 2>
 ; CHECK-NEXT:    [[TMP8:%.*]] = icmp eq i32 [[INDEX_NEXT]], 10000
-; CHECK-NEXT:    br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP9:![0-9]+]]
+; CHECK-NEXT:    br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i32 10000, 10000
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
@@ -273,7 +335,7 @@ define void @test_widen_extended_induction(ptr %dst) {
 ; CHECK-NEXT:    [[IV_NEXT]] = add i8 [[IV]], 1
 ; CHECK-NEXT:    [[IV_NEXT_EXT:%.*]] = zext i8 [[IV_NEXT]] to i32
 ; CHECK-NEXT:    [[EC:%.*]] = icmp eq i32 [[IV_NEXT_EXT]], 10000
-; CHECK-NEXT:    br i1 [[EC]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP10:![0-9]+]]
+; CHECK-NEXT:    br i1 [[EC]], label [[EXIT]], label [[LOOP]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       exit:
 ; CHECK-NEXT:    ret void
 ;
@@ -315,7 +377,7 @@ define void @test_widen_truncated_induction(ptr %A) {
 ; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
 ; CHECK-NEXT:    [[VEC_IND_NEXT]] = add <2 x i8> [[STEP_ADD]], <i8 2, i8 2>
 ; CHECK-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], 10000
-; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP11:![0-9]+]]
+; CHECK-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       middle.block:
 ; CHECK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 10000, 10000
 ; CHECK-NEXT:    br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
@@ -329,7 +391,7 @@ define void @test_widen_truncated_induction(ptr %A) {
 ; CHECK-NEXT:    store i8 [[IV_TRUNC]], ptr [[ARRAYIDX]], align 1
 ; CHECK-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
 ; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp ne i64 [[IV_NEXT]], 10000
-; CHECK-NEXT:    br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT]], !llvm.loop [[LOOP12:![0-9]+]]
+; CHECK-NEXT:    br i1 [[EXITCOND]], label [[LOOP]], label [[EXIT]], {{!llvm.loop ![0-9]+}}
 ; CHECK:       exit:
 ; CHECK-NEXT:    ret void
 ;