llvm/test/Transforms/LoopVectorize/uniform-blend.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
; RUN: opt -passes=loop-vectorize -force-vector-width=4 -S %s | FileCheck %s

@dst = external global [32 x i16], align 1

define void @blend_uniform_iv_trunc(i1 %c) {
; CHECK-LABEL: define void @blend_uniform_iv_trunc(
; CHECK-SAME: i1 [[C:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP0:%.*]] = trunc i64 [[INDEX]] to i16
; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [32 x i16], ptr @dst, i16 0, i16 [[TMP0]]
; CHECK-NEXT:    store <4 x i16> zeroinitializer, ptr [[TMP7]], align 2
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP4:%.*]] = icmp eq i64 [[INDEX_NEXT]], 32
; CHECK-NEXT:    br i1 [[TMP4]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop.header

loop.header:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  %iv.trunc.2 = trunc i64 %iv to i16
  br i1 %c, label %loop.next, label %loop.latch

loop.next:
  br label %loop.latch

loop.latch:
  %blend = phi i16 [ %iv.trunc.2, %loop.header ], [ %iv.trunc.2, %loop.next ]
  %dst.ptr = getelementptr inbounds [32 x i16], ptr @dst, i16 0, i16 %blend
  store i16 0, ptr %dst.ptr
  %iv.next = add nuw nsw i64 %iv, 1
  %cmp439 = icmp ult i64 %iv, 31
  br i1 %cmp439, label %loop.header, label %exit

exit:
  ret void
}

define void @blend_uniform_iv(i1 %c) {
; CHECK-LABEL: define void @blend_uniform_iv(
; CHECK-SAME: i1 [[C:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[TMP6:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds [32 x i16], ptr @dst, i16 0, i64 [[TMP6]]
; CHECK-NEXT:    store <4 x i16> zeroinitializer, ptr [[TMP7]], align 2
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[TMP6]], 4
; CHECK-NEXT:    [[TMP3:%.*]] = icmp eq i64 [[INDEX_NEXT]], 32
; CHECK-NEXT:    br i1 [[TMP3]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;

entry:
  br label %loop.header

loop.header:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  br i1 %c, label %loop.next, label %loop.latch

loop.next:
  br label %loop.latch

loop.latch:
  %blend = phi i64 [ %iv, %loop.header ], [ %iv, %loop.next ]
  %dst.ptr = getelementptr inbounds [32 x i16], ptr @dst, i16 0, i64 %blend
  store i16 0, ptr %dst.ptr
  %iv.next = add nuw nsw i64 %iv, 1
  %cmp439 = icmp ult i64 %iv, 31
  br i1 %cmp439, label %loop.header, label %exit

exit:
  ret void
}

define void @blend_chain_iv(i1 %c) {
; CHECK-LABEL: define void @blend_chain_iv(
; CHECK-SAME: i1 [[C:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 1
; CHECK-NEXT:    [[TMP5:%.*]] = add i64 [[INDEX]], 2
; CHECK-NEXT:    [[TMP7:%.*]] = add i64 [[INDEX]], 3
; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds [32 x i16], ptr @dst, i16 0, i64 [[INDEX]]
; CHECK-NEXT:    [[TMP4:%.*]] = getelementptr inbounds [32 x i16], ptr @dst, i16 0, i64 [[TMP3]]
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds [32 x i16], ptr @dst, i16 0, i64 [[TMP5]]
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds [32 x i16], ptr @dst, i16 0, i64 [[TMP7]]
; CHECK-NEXT:    store i16 0, ptr [[TMP2]], align 2
; CHECK-NEXT:    store i16 0, ptr [[TMP4]], align 2
; CHECK-NEXT:    store i16 0, ptr [[TMP6]], align 2
; CHECK-NEXT:    store i16 0, ptr [[TMP8]], align 2
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT]], 32
; CHECK-NEXT:    br i1 [[TMP9]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;

entry:
  br label %loop.header

loop.header:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  br i1 %c, label %loop.next, label %loop.latch

loop.next:
  br i1 %c, label %loop.next.2, label %loop.next.3

loop.next.2:
  br label %loop.next.3

loop.next.3:
  %blend.1 = phi i64 [ %iv, %loop.next ], [ %iv, %loop.next.2 ]
  br label %loop.latch

loop.latch:
  %blend = phi i64 [ %iv, %loop.header ], [ %blend.1, %loop.next.3 ]
  %dst.ptr = getelementptr inbounds [32 x i16], ptr @dst, i16 0, i64 %blend
  store i16 0, ptr %dst.ptr
  %iv.next = add nuw nsw i64 %iv, 1
  %cmp439 = icmp ult i64 %iv, 31
  br i1 %cmp439, label %loop.header, label %exit

exit:
  ret void
}

define void @redundant_branch_and_blends_without_mask(ptr %A) {
; CHECK-LABEL: define void @redundant_branch_and_blends_without_mask(
; CHECK-SAME: ptr [[A:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[TMP6:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 1
; CHECK-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 2
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds i32, ptr [[A]], i64 3
; CHECK-NEXT:    br i1 true, label %[[PRED_LOAD_IF:.*]], label %[[PRED_LOAD_CONTINUE:.*]]
; CHECK:       [[PRED_LOAD_IF]]:
; CHECK-NEXT:    [[TMP10:%.*]] = load i32, ptr [[A]], align 4
; CHECK-NEXT:    [[TMP13:%.*]] = insertelement <4 x i32> poison, i32 [[TMP10]], i64 0
; CHECK-NEXT:    br label %[[PRED_LOAD_CONTINUE]]
; CHECK:       [[PRED_LOAD_CONTINUE]]:
; CHECK-NEXT:    [[TMP9:%.*]] = phi <4 x i32> [ poison, %[[VECTOR_BODY]] ], [ [[TMP13]], %[[PRED_LOAD_IF]] ]
; CHECK-NEXT:    br i1 true, label %[[PRED_LOAD_IF1:.*]], label %[[PRED_LOAD_CONTINUE2:.*]]
; CHECK:       [[PRED_LOAD_IF1]]:
; CHECK-NEXT:    [[TMP14:%.*]] = load i32, ptr [[TMP6]], align 4
; CHECK-NEXT:    [[TMP11:%.*]] = insertelement <4 x i32> [[TMP9]], i32 [[TMP14]], i64 1
; CHECK-NEXT:    br label %[[PRED_LOAD_CONTINUE2]]
; CHECK:       [[PRED_LOAD_CONTINUE2]]:
; CHECK-NEXT:    [[TMP12:%.*]] = phi <4 x i32> [ [[TMP9]], %[[PRED_LOAD_CONTINUE]] ], [ [[TMP11]], %[[PRED_LOAD_IF1]] ]
; CHECK-NEXT:    br i1 false, label %[[PRED_LOAD_IF3:.*]], label %[[PRED_LOAD_CONTINUE4:.*]]
; CHECK:       [[PRED_LOAD_IF3]]:
; CHECK-NEXT:    [[TMP18:%.*]] = load i32, ptr [[TMP7]], align 4
; CHECK-NEXT:    [[TMP16:%.*]] = insertelement <4 x i32> [[TMP12]], i32 [[TMP18]], i64 2
; CHECK-NEXT:    br label %[[PRED_LOAD_CONTINUE4]]
; CHECK:       [[PRED_LOAD_CONTINUE4]]:
; CHECK-NEXT:    [[TMP15:%.*]] = phi <4 x i32> [ [[TMP12]], %[[PRED_LOAD_CONTINUE2]] ], [ [[TMP16]], %[[PRED_LOAD_IF3]] ]
; CHECK-NEXT:    br i1 false, label %[[PRED_LOAD_IF5:.*]], label %[[PRED_LOAD_CONTINUE6:.*]]
; CHECK:       [[PRED_LOAD_IF5]]:
; CHECK-NEXT:    [[TMP22:%.*]] = load i32, ptr [[TMP8]], align 4
; CHECK-NEXT:    [[TMP17:%.*]] = insertelement <4 x i32> [[TMP15]], i32 [[TMP22]], i64 3
; CHECK-NEXT:    br label %[[PRED_LOAD_CONTINUE6]]
; CHECK:       [[PRED_LOAD_CONTINUE6]]:
; CHECK-NEXT:    [[TMP24:%.*]] = phi <4 x i32> [ [[TMP15]], %[[PRED_LOAD_CONTINUE4]] ], [ [[TMP17]], %[[PRED_LOAD_IF5]] ]
; CHECK-NEXT:    [[TMP25:%.*]] = add <4 x i32> [[TMP24]], splat (i32 10)
; CHECK-NEXT:    [[TMP26:%.*]] = add <4 x i32> [[TMP24]], [[TMP25]]
; CHECK-NEXT:    br i1 true, label %[[PRED_STORE_IF:.*]], label %[[PRED_STORE_CONTINUE:.*]]
; CHECK:       [[PRED_STORE_IF]]:
; CHECK-NEXT:    [[TMP28:%.*]] = extractelement <4 x i32> [[TMP26]], i64 0
; CHECK-NEXT:    store i32 [[TMP28]], ptr [[A]], align 4
; CHECK-NEXT:    br label %[[PRED_STORE_CONTINUE]]
; CHECK:       [[PRED_STORE_CONTINUE]]:
; CHECK-NEXT:    br i1 true, label %[[PRED_STORE_IF7:.*]], label %[[PRED_STORE_CONTINUE8:.*]]
; CHECK:       [[PRED_STORE_IF7]]:
; CHECK-NEXT:    [[TMP30:%.*]] = extractelement <4 x i32> [[TMP26]], i64 1
; CHECK-NEXT:    store i32 [[TMP30]], ptr [[TMP6]], align 4
; CHECK-NEXT:    br label %[[PRED_STORE_CONTINUE8]]
; CHECK:       [[PRED_STORE_CONTINUE8]]:
; CHECK-NEXT:    br i1 false, label %[[PRED_STORE_IF9:.*]], label %[[PRED_STORE_CONTINUE10:.*]]
; CHECK:       [[PRED_STORE_IF9]]:
; CHECK-NEXT:    [[TMP32:%.*]] = extractelement <4 x i32> [[TMP26]], i64 2
; CHECK-NEXT:    store i32 [[TMP32]], ptr [[TMP7]], align 4
; CHECK-NEXT:    br label %[[PRED_STORE_CONTINUE10]]
; CHECK:       [[PRED_STORE_CONTINUE10]]:
; CHECK-NEXT:    br i1 false, label %[[PRED_STORE_IF11:.*]], label %[[PRED_STORE_CONTINUE12:.*]]
; CHECK:       [[PRED_STORE_IF11]]:
; CHECK-NEXT:    [[TMP34:%.*]] = extractelement <4 x i32> [[TMP26]], i64 3
; CHECK-NEXT:    store i32 [[TMP34]], ptr [[TMP8]], align 4
; CHECK-NEXT:    br label %[[PRED_STORE_CONTINUE12]]
; CHECK:       [[PRED_STORE_CONTINUE12]]:
; CHECK-NEXT:    br label %[[MIDDLE_BLOCK:.*]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop.header

loop.header:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.latch ]
  %gep.iv = getelementptr inbounds i32, ptr %A, i64 %iv
  %l = load i32, ptr %gep.iv
  %add = add i32 %l, 10
  br label %loop.latch

loop.latch:
  %p.1 = phi i32 [ %l, %loop.header ]
  %p.2 = phi i32 [ %add, %loop.header ]
  %res = add i32 %p.1, %p.2
  store i32 %res, ptr %gep.iv
  %iv.next = add i64 %iv, 1
  %ec = icmp eq i64 %iv, 1
  br i1 %ec, label %exit, label %loop.header

exit:
  ret void
}

; Test that blend is optimized away when other incoming values are poison.

define void @blend_poison(ptr %p, i1 %c) {
; CHECK-LABEL: define void @blend_poison(
; CHECK-SAME: ptr [[P:%.*]], i1 [[C:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; 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 <4 x i64> [ <i64 0, i64 1, i64 2, i64 3>, %[[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr i64, ptr [[P]], i64 [[INDEX]]
; CHECK-NEXT:    store <4 x i64> [[VEC_IND]], ptr [[TMP0]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[VEC_IND_NEXT]] = add nuw nsw <4 x i64> [[VEC_IND]], splat (i64 4)
; CHECK-NEXT:    [[TMP1:%.*]] = icmp eq i64 [[INDEX_NEXT]], 32
; CHECK-NEXT:    br i1 [[TMP1]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %latch ]
  br i1 %c, label %if, label %latch

if:
  br label %latch

latch:
  %blend = phi i64 [ poison, %loop ], [ %iv, %if ]
  %gep = getelementptr i64, ptr %p, i64 %iv
  store i64 %blend, ptr %gep
  %iv.next = add nuw nsw i64 %iv, 1
  %ec = icmp ult i64 %iv, 31
  br i1 %ec, label %loop, label %exit

exit:
  ret void
}

define void @blend_poison_first_argument(ptr %p, i1 %c) {
; CHECK-LABEL: define void @blend_poison_first_argument(
; CHECK-SAME: ptr [[P:%.*]], i1 [[C:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; 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 <4 x i64> [ <i64 0, i64 1, i64 2, i64 3>, %[[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr i64, ptr [[P]], i64 [[INDEX]]
; CHECK-NEXT:    store <4 x i64> [[VEC_IND]], ptr [[TMP0]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[VEC_IND_NEXT]] = add nuw nsw <4 x i64> [[VEC_IND]], splat (i64 4)
; CHECK-NEXT:    [[TMP1:%.*]] = icmp eq i64 [[INDEX_NEXT]], 32
; CHECK-NEXT:    br i1 [[TMP1]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %latch ]
  br i1 %c, label %if, label %latch

if:
  br label %latch

latch:
  %blend = phi i64 [ %iv, %loop ], [ poison, %if ]
  %gep = getelementptr i64, ptr %p, i64 %iv
  store i64 %blend, ptr %gep
  %iv.next = add nuw nsw i64 %iv, 1
  %ec = icmp ult i64 %iv, 31
  br i1 %ec, label %loop, label %exit

exit:
  ret void
}

define void @blend_all_poison(ptr %p, i1 %c) {
; CHECK-LABEL: define void @blend_all_poison(
; CHECK-SAME: ptr [[P:%.*]], i1 [[C:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    br label %[[VECTOR_BODY:.*]]
; CHECK:       [[VECTOR_BODY]]:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr i64, ptr [[P]], i64 [[INDEX]]
; CHECK-NEXT:    store <4 x i64> poison, ptr [[TMP0]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP1:%.*]] = icmp eq i64 [[INDEX_NEXT]], 32
; CHECK-NEXT:    br i1 [[TMP1]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
; CHECK:       [[MIDDLE_BLOCK]]:
; CHECK-NEXT:    br label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %latch ]
  br i1 %c, label %if, label %latch

if:
  br label %latch

latch:
  %blend = phi i64 [ poison, %loop ], [ poison, %if ]
  %gep = getelementptr i64, ptr %p, i64 %iv
  store i64 %blend, ptr %gep
  %iv.next = add nuw nsw i64 %iv, 1
  %ec = icmp ult i64 %iv, 31
  br i1 %ec, label %loop, label %exit

exit:
  ret void
}

;.
; CHECK: [[LOOP0]] = distinct !{[[LOOP0]], [[META1:![0-9]+]], [[META2:![0-9]+]]}
; CHECK: [[META1]] = !{!"llvm.loop.isvectorized", i32 1}
; CHECK: [[META2]] = !{!"llvm.loop.unroll.runtime.disable"}
; CHECK: [[LOOP3]] = distinct !{[[LOOP3]], [[META1]], [[META2]]}
; CHECK: [[LOOP4]] = distinct !{[[LOOP4]], [[META1]], [[META2]]}
; CHECK: [[LOOP5]] = distinct !{[[LOOP5]], [[META1]], [[META2]]}
; CHECK: [[LOOP6]] = distinct !{[[LOOP6]], [[META1]], [[META2]]}
; CHECK: [[LOOP7]] = distinct !{[[LOOP7]], [[META1]], [[META2]]}
;.