llvm/test/Transforms/LoopVectorize/AArch64/sve-vfabi.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 2
; RUN: opt < %s -passes=loop-vectorize,simplifycfg,instcombine -force-vector-interleave=1 -prefer-predicate-over-epilogue=predicate-dont-vectorize -S | FileCheck %s

target triple = "aarch64-unknown-linux-gnu"

define void @test_big_little_params(ptr readonly %a, ptr readonly %b, ptr noalias %c) #0 {
; CHECK-LABEL: define void @test_big_little_params
; CHECK-SAME: (ptr readonly [[A:%.*]], ptr readonly [[B:%.*]], ptr noalias [[C:%.*]]) #[[ATTR0:[0-9]+]] {
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT:    [[TMP5:%.*]] = shl nuw nsw i64 [[TMP4]], 2
; CHECK-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 4 x i1> @llvm.get.active.lane.mask.nxv4i1.i64(i64 0, i64 1025)
; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
; CHECK:       vector.body:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 4 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr i32, ptr [[A]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 4 x i32> @llvm.masked.load.nxv4i32.p0(ptr [[TMP0]], i32 4, <vscale x 4 x i1> [[ACTIVE_LANE_MASK]], <vscale x 4 x i32> poison)
; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr i8, ptr [[B]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_MASKED_LOAD1:%.*]] = call <vscale x 4 x i8> @llvm.masked.load.nxv4i8.p0(ptr [[TMP1]], i32 1, <vscale x 4 x i1> [[ACTIVE_LANE_MASK]], <vscale x 4 x i8> poison)
; CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 4 x i32> @foo_vector(<vscale x 4 x i32> [[WIDE_MASKED_LOAD]], <vscale x 4 x i8> [[WIDE_MASKED_LOAD1]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds i32, ptr [[C]], i64 [[INDEX]]
; CHECK-NEXT:    call void @llvm.masked.store.nxv4i32.p0(<vscale x 4 x i32> [[TMP2]], ptr [[TMP3]], i32 4, <vscale x 4 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP5]]
; CHECK-NEXT:    [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 4 x i1> @llvm.get.active.lane.mask.nxv4i1.i64(i64 [[INDEX_NEXT]], i64 1025)
; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <vscale x 4 x i1> [[ACTIVE_LANE_MASK_NEXT]], i64 0
; CHECK-NEXT:    br i1 [[TMP6]], label [[VECTOR_BODY]], label [[EXIT:%.*]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK:       exit:
; CHECK-NEXT:    ret void
;
entry:
  br label %for.body

for.body:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
  %gep_s = getelementptr i32, ptr %a, i64 %iv
  %load_s = load i32, ptr %gep_s
  %gep_b = getelementptr i8, ptr %b, i64 %iv
  %load_b = load i8, ptr %gep_b
  %call = call i32 @foo_big_little(i32 %load_s, i8 %load_b) #1
  %arrayidx = getelementptr inbounds i32, ptr %c, i64 %iv
  store i32 %call, ptr %arrayidx
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv.next, 1025
  br i1 %exitcond, label %exit, label %for.body

exit:
  ret void
}

define void @test_little_big_params(ptr readonly %a, ptr readonly %b, ptr noalias %c) #0 {
; CHECK-LABEL: define void @test_little_big_params
; CHECK-SAME: (ptr readonly [[A:%.*]], ptr readonly [[B:%.*]], ptr noalias [[C:%.*]]) #[[ATTR0]] {
; CHECK-NEXT:  entry:
; CHECK-NEXT:    [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
; CHECK-NEXT:    [[TMP5:%.*]] = shl nuw nsw i64 [[TMP4]], 1
; CHECK-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)
; CHECK-NEXT:    br label [[VECTOR_BODY:%.*]]
; CHECK:       vector.body:
; CHECK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT:    [[TMP0:%.*]] = getelementptr float, ptr [[A]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x float> @llvm.masked.load.nxv2f32.p0(ptr [[TMP0]], i32 4, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x float> poison)
; CHECK-NEXT:    [[TMP1:%.*]] = getelementptr double, ptr [[B]], i64 [[INDEX]]
; CHECK-NEXT:    [[WIDE_MASKED_LOAD1:%.*]] = call <vscale x 2 x double> @llvm.masked.load.nxv2f64.p0(ptr [[TMP1]], i32 8, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x double> poison)
; CHECK-NEXT:    [[TMP2:%.*]] = call <vscale x 2 x double> @bar_vector(<vscale x 2 x float> [[WIDE_MASKED_LOAD]], <vscale x 2 x double> [[WIDE_MASKED_LOAD1]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT:    [[TMP3:%.*]] = getelementptr inbounds double, ptr [[C]], i64 [[INDEX]]
; CHECK-NEXT:    call void @llvm.masked.store.nxv2f64.p0(<vscale x 2 x double> [[TMP2]], ptr [[TMP3]], i32 8, <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], [[TMP5]]
; CHECK-NEXT:    [[ACTIVE_LANE_MASK_NEXT]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[INDEX_NEXT]], i64 1025)
; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <vscale x 2 x i1> [[ACTIVE_LANE_MASK_NEXT]], i64 0
; CHECK-NEXT:    br i1 [[TMP6]], label [[VECTOR_BODY]], label [[FOR_COND_CLEANUP:%.*]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK:       for.cond.cleanup:
; CHECK-NEXT:    ret void
;
entry:
  br label %for.body

for.body:
  %iv = phi i64 [ 0, %entry ], [ %iv.next, %for.body ]
  %gep_f = getelementptr float, ptr %a, i64 %iv
  %load_f = load float, ptr %gep_f
  %gep_d = getelementptr double, ptr %b, i64 %iv
  %load_d = load double, ptr %gep_d
  %call = call double @bar_little_big(float %load_f, double %load_d) #2
  %arrayidx = getelementptr inbounds double, ptr %c, i64 %iv
  store double %call, ptr %arrayidx
  %iv.next = add nuw nsw i64 %iv, 1
  %exitcond = icmp eq i64 %iv.next, 1025
  br i1 %exitcond, label %for.cond.cleanup, label %for.body

for.cond.cleanup:
  ret void
}

;; TODO: Test uniform and linear parameters when they are properly supported,
;;       especially a variant with no vector parameters so the return type
;;       must be used to find the VF.

;; Scalar functions
declare i32 @foo_big_little(i32, i8)
declare double @bar_little_big(float, double)

;; Vector function variants
declare <vscale x 4 x i32> @foo_vector(<vscale x 4 x i32>, <vscale x 4 x i8>, <vscale x 4 x i1>)
declare <vscale x 2 x double> @bar_vector(<vscale x 2 x float>, <vscale x 2 x double>, <vscale x 2 x i1>)

attributes #0 = { "target-features"="+sve" vscale_range(1,16) }
attributes #1 = { nounwind "vector-function-abi-variant"="_ZGVsMxvv_foo_big_little(foo_vector)" }
attributes #2 = { nounwind "vector-function-abi-variant"="_ZGVsMxvv_bar_little_big(bar_vector)" }