llvm/test/Transforms/LoopVectorize/operand-bundles.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 6
; RUN: opt -p loop-vectorize -force-vector-width=4 -S %s | FileCheck %s

define void @call_loop_invariant_operand_bundle(ptr %dst, {float, float} %sv) {
; CHECK-LABEL: define void @call_loop_invariant_operand_bundle(
; CHECK-SAME: ptr [[DST:%.*]], { float, float } [[SV:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i32 [ 0, %[[ENTRY]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[A:%.*]] = extractvalue { float, float } [[SV]], 0
; CHECK-NEXT:    [[B:%.*]] = extractvalue { float, float } [[SV]], 1
; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr float, ptr [[DST]], i32 [[IV]]
; CHECK-NEXT:    [[P:%.*]] = call float @llvm.pow.f32(float [[A]], float [[B]]) [ "deopt"(float 1.000000e+01) ]
; CHECK-NEXT:    store float [[P]], ptr [[ADDR]], align 4
; CHECK-NEXT:    [[IV_NEXT]] = add nsw i32 [[IV]], 1
; CHECK-NEXT:    [[COND:%.*]] = icmp ne i32 [[IV_NEXT]], 1000
; CHECK-NEXT:    br i1 [[COND]], label %[[LOOP]], label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i32 [ 0, %entry ], [ %iv.next, %loop ]
  %a = extractvalue { float, float } %sv, 0
  %b = extractvalue { float, float } %sv, 1
  %addr = getelementptr float, ptr %dst, i32 %iv
  %p = call float @llvm.pow.f32(float %a, float %b) [ "deopt"(float 10.0) ]
  store float %p, ptr %addr
  %iv.next = add nsw i32 %iv, 1
  %cond = icmp ne i32 %iv.next, 1000
  br i1 %cond, label %loop, label %exit

exit:
  ret void
}

define void @call_unknown_operand_bundle(ptr %dst, {float, float} %sv) {
; CHECK-LABEL: define void @call_unknown_operand_bundle(
; CHECK-SAME: ptr [[DST:%.*]], { float, float } [[SV:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i32 [ 0, %[[ENTRY]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[A:%.*]] = extractvalue { float, float } [[SV]], 0
; CHECK-NEXT:    [[B:%.*]] = extractvalue { float, float } [[SV]], 1
; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr float, ptr [[DST]], i32 [[IV]]
; CHECK-NEXT:    [[P:%.*]] = call float @llvm.pow.f32(float [[A]], float [[B]]) [ "unknown"(ptr null) ]
; CHECK-NEXT:    store float [[P]], ptr [[ADDR]], align 4
; CHECK-NEXT:    [[IV_NEXT]] = add nsw i32 [[IV]], 1
; CHECK-NEXT:    [[COND:%.*]] = icmp ne i32 [[IV_NEXT]], 1000
; CHECK-NEXT:    br i1 [[COND]], label %[[LOOP]], label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i32 [ 0, %entry ], [ %iv.next, %loop ]
  %a = extractvalue { float, float } %sv, 0
  %b = extractvalue { float, float } %sv, 1
  %addr = getelementptr float, ptr %dst, i32 %iv
  %p = call float @llvm.pow.f32(float %a, float %b) [ "unknown"(ptr null) ]
  store float %p, ptr %addr
  %iv.next = add nsw i32 %iv, 1
  %cond = icmp ne i32 %iv.next, 1000
  br i1 %cond, label %loop, label %exit

exit:
  ret void
}

define void @call_cold_operand_bundle(ptr %dst, {float, float} %sv) {
; CHECK-LABEL: define void @call_cold_operand_bundle(
; CHECK-SAME: ptr [[DST:%.*]], { float, float } [[SV:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i32 [ 0, %[[ENTRY]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[A:%.*]] = extractvalue { float, float } [[SV]], 0
; CHECK-NEXT:    [[B:%.*]] = extractvalue { float, float } [[SV]], 1
; CHECK-NEXT:    [[ADDR:%.*]] = getelementptr float, ptr [[DST]], i32 [[IV]]
; CHECK-NEXT:    [[P:%.*]] = call float @llvm.pow.f32(float [[A]], float [[B]]) [ "cold"() ]
; CHECK-NEXT:    store float [[P]], ptr [[ADDR]], align 4
; CHECK-NEXT:    [[IV_NEXT]] = add nsw i32 [[IV]], 1
; CHECK-NEXT:    [[COND:%.*]] = icmp ne i32 [[IV_NEXT]], 1000
; CHECK-NEXT:    br i1 [[COND]], label %[[LOOP]], label %[[EXIT:.*]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i32 [ 0, %entry ], [ %iv.next, %loop ]
  %a = extractvalue { float, float } %sv, 0
  %b = extractvalue { float, float } %sv, 1
  %addr = getelementptr float, ptr %dst, i32 %iv
  %p = call float @llvm.pow.f32(float %a, float %b) [ "cold"() ]
  store float %p, ptr %addr
  %iv.next = add nsw i32 %iv, 1
  %cond = icmp ne i32 %iv.next, 1000
  br i1 %cond, label %loop, label %exit

exit:
  ret void
}

define void @assume_loop_variant_operand_bundle(ptr noalias %a, ptr noalias %b) {
; CHECK-LABEL: define void @assume_loop_variant_operand_bundle(
; CHECK-SAME: ptr noalias [[A:%.*]], ptr noalias [[B:%.*]]) {
; 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:    [[TMP1:%.*]] = add i64 [[INDEX]], 1
; CHECK-NEXT:    [[TMP2:%.*]] = add i64 [[INDEX]], 2
; CHECK-NEXT:    [[TMP3:%.*]] = add i64 [[INDEX]], 3
; CHECK-NEXT:    [[TMP8:%.*]] = getelementptr inbounds float, ptr [[B]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP8]], align 4
; CHECK-NEXT:    tail call void @llvm.assume(i1 true) [ "align"(ptr [[A]], i64 [[INDEX]]) ]
; CHECK-NEXT:    tail call void @llvm.assume(i1 true) [ "align"(ptr [[A]], i64 [[TMP1]]) ]
; CHECK-NEXT:    tail call void @llvm.assume(i1 true) [ "align"(ptr [[A]], i64 [[TMP2]]) ]
; CHECK-NEXT:    tail call void @llvm.assume(i1 true) [ "align"(ptr [[A]], i64 [[TMP3]]) ]
; CHECK-NEXT:    [[TMP5:%.*]] = fadd <4 x float> [[WIDE_LOAD]], splat (float 1.000000e+00)
; CHECK-NEXT:    [[TMP10:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[INDEX]]
; CHECK-NEXT:    store <4 x float> [[TMP5]], ptr [[TMP10]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP11:%.*]] = icmp eq i64 [[INDEX_NEXT]], 1600
; CHECK-NEXT:    br i1 [[TMP11]], 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

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %arrayidx = getelementptr inbounds float, ptr %b, i64 %iv
  %0 = load float, ptr %arrayidx, align 4
  %cmp1 = fcmp ogt float %0, 1.000000e+02
  tail call void @llvm.assume(i1 true) [ "align"(ptr %a, i64 %iv) ]
  %add = fadd float %0, 1.000000e+00
  %arrayidx5 = getelementptr inbounds float, ptr %a, i64 %iv
  store float %add, ptr %arrayidx5, align 4
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv, 1599
  br i1 %exitcond, label %exit, label %loop

exit:
  ret void
}

define void @assume_cold_operand_bundle(ptr noalias %a, ptr noalias %b) {
; CHECK-LABEL: define void @assume_cold_operand_bundle(
; CHECK-SAME: ptr noalias [[A:%.*]], ptr noalias [[B:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    tail call void @llvm.assume(i1 true) [ "cold"() ]
; 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 inbounds float, ptr [[B]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP0]], align 4
; CHECK-NEXT:    [[TMP1:%.*]] = fadd <4 x float> [[WIDE_LOAD]], splat (float 1.000000e+00)
; CHECK-NEXT:    [[TMP2:%.*]] = getelementptr inbounds float, ptr [[A]], i64 [[INDEX]]
; CHECK-NEXT:    store <4 x float> [[TMP1]], ptr [[TMP2]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP3:%.*]] = icmp eq i64 [[INDEX_NEXT]], 1600
; 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

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %arrayidx = getelementptr inbounds float, ptr %b, i64 %iv
  %0 = load float, ptr %arrayidx, align 4
  %cmp1 = fcmp ogt float %0, 1.000000e+02
  tail call void @llvm.assume(i1 true) [ "cold"() ]
  %add = fadd float %0, 1.000000e+00
  %arrayidx5 = getelementptr inbounds float, ptr %a, i64 %iv
  store float %add, ptr %arrayidx5, align 4
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv, 1599
  br i1 %exitcond, label %exit, label %loop

exit:
  ret void
}

define void @assume_loop_invariant_operand_bundle(ptr noalias %a, ptr %q) {
; CHECK-LABEL: define void @assume_loop_invariant_operand_bundle(
; CHECK-SAME: ptr noalias [[A:%.*]], ptr [[Q:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*:]]
; CHECK-NEXT:    br label %[[VECTOR_PH:.*]]
; CHECK:       [[VECTOR_PH]]:
; CHECK-NEXT:    call void @llvm.assume(i1 true) [ "nonnull"(ptr [[Q]]) ]
; 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 inbounds float, ptr [[A]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP0]], align 4
; CHECK-NEXT:    [[TMP1:%.*]] = fadd <4 x float> [[WIDE_LOAD]], splat (float 1.000000e+00)
; CHECK-NEXT:    store <4 x float> [[TMP1]], ptr [[TMP0]], align 4
; CHECK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT:    [[TMP2:%.*]] = icmp eq i64 [[INDEX_NEXT]], 1600
; CHECK-NEXT:    br i1 [[TMP2]], 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

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %gep = getelementptr inbounds float, ptr %a, i64 %iv
  call void @llvm.assume(i1 true) [ "nonnull"(ptr %q) ]
  %l = load float, ptr %gep, align 4
  %add = fadd float %l, 1.000000e+00
  store float %add, ptr %gep, align 4
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv.next, 1600
  br i1 %exitcond, label %exit, label %loop

exit:
  ret void
}

define void @call_loop_invariant_unknown_operand_bundle(ptr %dst, float %x) {
; CHECK-LABEL: define void @call_loop_invariant_unknown_operand_bundle(
; CHECK-SAME: ptr [[DST:%.*]], float [[X:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[GEP:%.*]] = getelementptr float, ptr [[DST]], i64 [[IV]]
; CHECK-NEXT:    [[S:%.*]] = call float @foo(float [[X]]) [ "unknown"() ]
; CHECK-NEXT:    store float [[S]], ptr [[GEP]], align 4
; CHECK-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[IV_NEXT]], 1000
; CHECK-NEXT:    br i1 [[EXITCOND]], label %[[EXIT:.*]], label %[[LOOP]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %gep = getelementptr float, ptr %dst, i64 %iv
  %s = call float @foo(float %x) [ "unknown"() ]
  store float %s, ptr %gep
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv.next, 1000
  br i1 %exitcond, label %exit, label %loop

exit:
  ret void
}

define void @call_unused_unknown_operand_bundle(ptr %dst) {
; CHECK-LABEL: define void @call_unused_unknown_operand_bundle(
; CHECK-SAME: ptr [[DST:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[GEP:%.*]] = getelementptr float, ptr [[DST]], i64 [[IV]]
; CHECK-NEXT:    [[L:%.*]] = load float, ptr [[GEP]], align 4
; CHECK-NEXT:    [[S:%.*]] = call float @foo(float [[L]]) [ "unknown"() ]
; CHECK-NEXT:    [[ADD:%.*]] = fadd float [[L]], 1.000000e+00
; CHECK-NEXT:    store float [[ADD]], ptr [[GEP]], align 4
; CHECK-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[IV_NEXT]], 1000
; CHECK-NEXT:    br i1 [[EXITCOND]], label %[[EXIT:.*]], label %[[LOOP]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %gep = getelementptr float, ptr %dst, i64 %iv
  %l = load float, ptr %gep
  %s = call float @foo(float %l) [ "unknown"() ]
  %add = fadd float %l, 1.000000e+00
  store float %add, ptr %gep
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv.next, 1000
  br i1 %exitcond, label %exit, label %loop

exit:
  ret void
}

define void @call_deactivation_symbol_operand_bundle(ptr %dst) {
; CHECK-LABEL: define void @call_deactivation_symbol_operand_bundle(
; CHECK-SAME: ptr [[DST:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[GEP:%.*]] = getelementptr ptr, ptr [[DST]], i64 [[IV]]
; CHECK-NEXT:    [[L:%.*]] = load ptr, ptr [[GEP]], align 8
; CHECK-NEXT:    [[S:%.*]] = notail call ptr @bar(ptr [[L]]) [ "deactivation-symbol"(ptr @sym) ]
; CHECK-NEXT:    store ptr [[S]], ptr [[GEP]], align 8
; CHECK-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[IV_NEXT]], 1000
; CHECK-NEXT:    br i1 [[EXITCOND]], label %[[EXIT:.*]], label %[[LOOP]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %gep = getelementptr ptr, ptr %dst, i64 %iv
  %l = load ptr, ptr %gep
  %s = notail call ptr @bar(ptr %l) [ "deactivation-symbol"(ptr @sym) ]
  store ptr %s, ptr %gep
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv.next, 1000
  br i1 %exitcond, label %exit, label %loop

exit:
  ret void
}

; Calls with loop-variant bundle operands must not be widened.
define void @intrinsic_loop_variant_deopt_operand_bundle(ptr %dst, i64 %n) {
; CHECK-LABEL: define void @intrinsic_loop_variant_deopt_operand_bundle(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[N:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[GEP:%.*]] = getelementptr float, ptr [[DST]], i64 [[IV]]
; CHECK-NEXT:    [[L:%.*]] = load float, ptr [[GEP]], align 4
; CHECK-NEXT:    [[S:%.*]] = call float @llvm.sin.f32(float [[L]]) [ "deopt"(i64 [[IV]]) ]
; CHECK-NEXT:    store float [[S]], ptr [[GEP]], align 4
; CHECK-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[IV_NEXT]], [[N]]
; CHECK-NEXT:    br i1 [[EXITCOND]], label %[[EXIT:.*]], label %[[LOOP]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %gep = getelementptr float, ptr %dst, i64 %iv
  %l = load float, ptr %gep
  %s = call float @llvm.sin.f32(float %l) [ "deopt"(i64 %iv) ]
  store float %s, ptr %gep
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv.next, %n
  br i1 %exitcond, label %exit, label %loop

exit:
  ret void
}

define void @call_loop_variant_unknown_operand_bundle(ptr %dst, i64 %n) {
; CHECK-LABEL: define void @call_loop_variant_unknown_operand_bundle(
; CHECK-SAME: ptr [[DST:%.*]], i64 [[N:%.*]]) {
; CHECK-NEXT:  [[ENTRY:.*]]:
; CHECK-NEXT:    br label %[[LOOP:.*]]
; CHECK:       [[LOOP]]:
; CHECK-NEXT:    [[IV:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[IV_NEXT:%.*]], %[[LOOP]] ]
; CHECK-NEXT:    [[GEP:%.*]] = getelementptr float, ptr [[DST]], i64 [[IV]]
; CHECK-NEXT:    [[L:%.*]] = load float, ptr [[GEP]], align 4
; CHECK-NEXT:    [[S:%.*]] = call float @foo(float [[L]]) [ "unknown"(i64 [[IV]]) ]
; CHECK-NEXT:    store float [[S]], ptr [[GEP]], align 4
; CHECK-NEXT:    [[IV_NEXT]] = add nuw nsw i64 [[IV]], 1
; CHECK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[IV_NEXT]], [[N]]
; CHECK-NEXT:    br i1 [[EXITCOND]], label %[[EXIT:.*]], label %[[LOOP]]
; CHECK:       [[EXIT]]:
; CHECK-NEXT:    ret void
;
entry:
  br label %loop

loop:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop ]
  %gep = getelementptr float, ptr %dst, i64 %iv
  %l = load float, ptr %gep
  %s = call float @foo(float %l) [ "unknown"(i64 %iv) ]
  store float %s, ptr %gep
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv.next, %n
  br i1 %exitcond, label %exit, label %loop

exit:
  ret void
}

@sym = external global i8

declare float @foo(float) memory(none) nounwind willreturn #0
declare <4 x float> @foo_v4(<4 x float>) memory(none) nounwind willreturn
declare ptr @bar(ptr) memory(none) nounwind willreturn #1
declare <4 x ptr> @bar_v4(<4 x ptr>) memory(none) nounwind willreturn

attributes #0 = { "vector-function-abi-variant"="_ZGV_LLVM_N4v_foo(foo_v4)" }
attributes #1 = { "vector-function-abi-variant"="_ZGV_LLVM_N4v_bar(bar_v4)" }