Address review comment

Author: david-arm

llvm/test/Transforms/LoopVectorize/AArch64/masked-call-scalarize.ll

@@ -0,0 +1,211 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -passes=loop-vectorize,instsimplify -force-vector-interleave=1 -S | FileCheck %s --check-prefixes=TFNONE
+; RUN: opt < %s -passes=loop-vectorize,instsimplify,simplifycfg -force-vector-interleave=1 -prefer-predicate-over-epilogue=predicate-dont-vectorize -S | FileCheck %s --check-prefixes=TFCOMMON,TFALWAYS
+; RUN: opt < %s -passes=loop-vectorize,instsimplify,simplifycfg -force-vector-interleave=1 -prefer-predicate-over-epilogue=predicate-else-scalar-epilogue -S | FileCheck %s --check-prefixes=TFCOMMON,TFFALLBACK
+; RUN: opt < %s -passes=loop-vectorize,instsimplify,simplifycfg -force-vector-interleave=2 -prefer-predicate-over-epilogue=predicate-dont-vectorize -S | FileCheck %s --check-prefixes=TFA_INTERLEAVE
+
+target triple = "aarch64-unknown-linux-gnu"
+
+define void @test_widen_exp_v2(ptr noalias %p2, ptr noalias %p, i64 %n) #5 {
+; TFNONE-LABEL: @test_widen_exp_v2(
+; TFNONE-NEXT:  entry:
+; TFNONE-NEXT:    [[TMP0:%.*]] = add i64 [[N:%.*]], 1
+; TFNONE-NEXT:    [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP0]], 2
+; TFNONE-NEXT:    br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
+; TFNONE:       vector.ph:
+; TFNONE-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[TMP0]], 2
+; TFNONE-NEXT:    [[N_VEC:%.*]] = sub i64 [[TMP0]], [[N_MOD_VF]]
+; TFNONE-NEXT:    br label [[VECTOR_BODY:%.*]]
+; TFNONE:       vector.body:
+; TFNONE-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
+; TFNONE-NEXT:    [[TMP7:%.*]] = load double, ptr [[P2:%.*]], align 8
+; TFNONE-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <2 x double> poison, double [[TMP7]], i64 0
+; TFNONE-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <2 x double> [[BROADCAST_SPLATINSERT]], <2 x double> poison, <2 x i32> zeroinitializer
+; TFNONE-NEXT:    [[TMP2:%.*]] = call <2 x double> @exp_fixed(<2 x double> [[BROADCAST_SPLAT]])
+; TFNONE-NEXT:    [[TMP3:%.*]] = fcmp ogt <2 x double> [[TMP2]], zeroinitializer
+; TFNONE-NEXT:    [[PREDPHI:%.*]] = select <2 x i1> [[TMP3]], <2 x double> zeroinitializer, <2 x double> splat (double 1.000000e+00)
+; TFNONE-NEXT:    [[TMP14:%.*]] = extractelement <2 x double> [[PREDPHI]], i32 1
+; TFNONE-NEXT:    store double [[TMP14]], ptr [[P:%.*]], align 8
+; TFNONE-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 2
+; TFNONE-NEXT:    [[TMP15:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; TFNONE-NEXT:    br i1 [[TMP15]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; TFNONE:       middle.block:
+; TFNONE-NEXT:    [[CMP_N:%.*]] = icmp eq i64 [[TMP0]], [[N_VEC]]
+; TFNONE-NEXT:    br i1 [[CMP_N]], label [[END:%.*]], label [[SCALAR_PH]]
+; TFNONE:       scalar.ph:
+; TFNONE-NEXT:    [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
+; TFNONE-NEXT:    br label [[LOOP:%.*]]
+; TFNONE:       loop:
+; TFNONE-NEXT:    [[IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[IV_NEXT:%.*]], [[LOOP_END:%.*]] ]
+; TFNONE-NEXT:    [[LD:%.*]] = load double, ptr [[P2]], align 8
+; TFNONE-NEXT:    [[EXP:%.*]] = tail call double @llvm.exp.f64(double [[LD]]) #[[ATTR2:[0-9]+]]
+; TFNONE-NEXT:    [[COND1:%.*]] = fcmp ogt double [[EXP]], 0.000000e+00
+; TFNONE-NEXT:    br i1 [[COND1]], label [[LOOP_MIDDLE:%.*]], label [[LOOP_END]]
+; TFNONE:       loop.middle:
+; TFNONE-NEXT:    br label [[LOOP_END]]
+; TFNONE:       loop.end:
+; TFNONE-NEXT:    [[SINK:%.*]] = phi double [ 0.000000e+00, [[LOOP_MIDDLE]] ], [ 1.000000e+00, [[LOOP]] ]
+; TFNONE-NEXT:    store double [[SINK]], ptr [[P]], align 8
+; TFNONE-NEXT:    [[IV_NEXT]] = add i64 [[IV]], 1
+; TFNONE-NEXT:    [[COND2:%.*]] = icmp eq i64 [[IV]], [[N]]
+; TFNONE-NEXT:    br i1 [[COND2]], label [[END]], label [[LOOP]], !llvm.loop [[LOOP3:![0-9]+]]
+; TFNONE:       end:
+; TFNONE-NEXT:    ret void
+;
+; TFCOMMON-LABEL: @test_widen_exp_v2(
+; TFCOMMON-NEXT:  entry:
+; TFCOMMON-NEXT:    [[TMP0:%.*]] = add i64 [[N:%.*]], 1
+; TFCOMMON-NEXT:    [[N_RND_UP:%.*]] = add i64 [[TMP0]], 1
+; TFCOMMON-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP]], 2
+; TFCOMMON-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP]], [[N_MOD_VF]]
+; TFCOMMON-NEXT:    [[TMP1:%.*]] = sub i64 [[TMP0]], 2
+; TFCOMMON-NEXT:    [[TMP2:%.*]] = icmp ugt i64 [[TMP0]], 2
+; TFCOMMON-NEXT:    [[TMP3:%.*]] = select i1 [[TMP2]], i64 [[TMP1]], i64 0
+; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 0, i64 [[TMP0]])
+; TFCOMMON-NEXT:    br label [[LOOP:%.*]]
+; TFCOMMON:       vector.body:
+; TFCOMMON-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDEX_NEXT:%.*]], [[PRED_STORE_CONTINUE6:%.*]] ]
+; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[PRED_STORE_CONTINUE6]] ]
+; TFCOMMON-NEXT:    [[LD:%.*]] = load double, ptr [[P2:%.*]], align 8
+; TFCOMMON-NEXT:    [[TMP5:%.*]] = tail call double @llvm.exp.f64(double [[LD]]) #[[ATTR3:[0-9]+]]
+; TFCOMMON-NEXT:    [[TMP6:%.*]] = tail call double @llvm.exp.f64(double [[LD]]) #[[ATTR3]]
+; TFCOMMON-NEXT:    [[TMP7:%.*]] = insertelement <2 x double> poison, double [[TMP5]], i32 0
+; TFCOMMON-NEXT:    [[TMP8:%.*]] = insertelement <2 x double> [[TMP7]], double [[TMP6]], i32 1
+; TFCOMMON-NEXT:    [[TMP9:%.*]] = fcmp ogt <2 x double> [[TMP8]], zeroinitializer
+; TFCOMMON-NEXT:    [[TMP10:%.*]] = xor <2 x i1> [[TMP9]], splat (i1 true)
+; TFCOMMON-NEXT:    [[TMP11:%.*]] = select <2 x i1> [[ACTIVE_LANE_MASK]], <2 x i1> [[TMP10]], <2 x i1> zeroinitializer
+; TFCOMMON-NEXT:    [[PREDPHI:%.*]] = select <2 x i1> [[TMP11]], <2 x double> splat (double 1.000000e+00), <2 x double> zeroinitializer
+; TFCOMMON-NEXT:    [[TMP16:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK]], i32 0
+; TFCOMMON-NEXT:    br i1 [[TMP16]], label [[PRED_STORE_IF:%.*]], label [[PRED_STORE_CONTINUE:%.*]]
+; TFCOMMON:       pred.store.if:
+; TFCOMMON-NEXT:    [[SINK:%.*]] = extractelement <2 x double> [[PREDPHI]], i32 0
+; TFCOMMON-NEXT:    store double [[SINK]], ptr [[P:%.*]], align 8
+; TFCOMMON-NEXT:    br label [[PRED_STORE_CONTINUE]]
+; TFCOMMON:       pred.store.continue:
+; TFCOMMON-NEXT:    [[TMP14:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK]], i32 1
+; TFCOMMON-NEXT:    br i1 [[TMP14]], label [[PRED_STORE_IF1:%.*]], label [[PRED_STORE_CONTINUE6]]
+; TFCOMMON:       pred.store.if1:
+; TFCOMMON-NEXT:    [[TMP19:%.*]] = extractelement <2 x double> [[PREDPHI]], i32 1
+; TFCOMMON-NEXT:    store double [[TMP19]], ptr [[P]], align 8
+; TFCOMMON-NEXT:    br label [[PRED_STORE_CONTINUE6]]
+; TFCOMMON:       pred.store.continue2:
+; TFCOMMON-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 2
+; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK_NEXT]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 [[INDEX]], i64 [[TMP3]])
+; TFCOMMON-NEXT:    [[TMP18:%.*]] = xor <2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
+; TFCOMMON-NEXT:    [[TMP17:%.*]] = extractelement <2 x i1> [[TMP18]], i32 0
+; TFCOMMON-NEXT:    br i1 [[TMP17]], label [[END:%.*]], label [[LOOP]], !llvm.loop [[LOOP0:![0-9]+]]
+; TFCOMMON:       end:
+; TFCOMMON-NEXT:    ret void
+;
+; TFA_INTERLEAVE-LABEL: @test_widen_exp_v2(
+; TFA_INTERLEAVE-NEXT:  entry:
+; TFA_INTERLEAVE-NEXT:    [[N_RND_UP:%.*]] = add i64 [[TMP0:%.*]], 1
+; TFA_INTERLEAVE-NEXT:    [[N_RND_UP1:%.*]] = add i64 [[N_RND_UP]], 3
+; TFA_INTERLEAVE-NEXT:    [[N_MOD_VF:%.*]] = urem i64 [[N_RND_UP1]], 4
+; TFA_INTERLEAVE-NEXT:    [[N_VEC:%.*]] = sub i64 [[N_RND_UP1]], [[N_MOD_VF]]
+; TFA_INTERLEAVE-NEXT:    [[TMP1:%.*]] = sub i64 [[N_RND_UP]], 4
+; TFA_INTERLEAVE-NEXT:    [[TMP2:%.*]] = icmp ugt i64 [[N_RND_UP]], 4
+; TFA_INTERLEAVE-NEXT:    [[TMP3:%.*]] = select i1 [[TMP2]], i64 [[TMP1]], i64 0
+; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 0, i64 [[N_RND_UP]])
+; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY1:%.*]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 2, i64 [[N_RND_UP]])
+; TFA_INTERLEAVE-NEXT:    br label [[VECTOR_BODY:%.*]]
+; TFA_INTERLEAVE:       vector.body:
+; TFA_INTERLEAVE-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[INDEX_NEXT:%.*]], [[PRED_STORE_CONTINUE9:%.*]] ]
+; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], [[PRED_STORE_CONTINUE9]] ]
+; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK2:%.*]] = phi <2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY1]], [[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT10:%.*]], [[PRED_STORE_CONTINUE9]] ]
+; TFA_INTERLEAVE-NEXT:    [[TMP4:%.*]] = load double, ptr [[P2:%.*]], align 8
+; TFA_INTERLEAVE-NEXT:    [[TMP5:%.*]] = tail call double @llvm.exp.f64(double [[TMP4]]) #[[ATTR3:[0-9]+]]
+; TFA_INTERLEAVE-NEXT:    [[TMP6:%.*]] = tail call double @llvm.exp.f64(double [[TMP4]]) #[[ATTR3]]
+; TFA_INTERLEAVE-NEXT:    [[TMP7:%.*]] = insertelement <2 x double> poison, double [[TMP5]], i32 0
+; TFA_INTERLEAVE-NEXT:    [[TMP8:%.*]] = insertelement <2 x double> [[TMP7]], double [[TMP6]], i32 1
+; TFA_INTERLEAVE-NEXT:    [[TMP9:%.*]] = tail call double @llvm.exp.f64(double [[TMP4]]) #[[ATTR3]]
+; TFA_INTERLEAVE-NEXT:    [[TMP10:%.*]] = tail call double @llvm.exp.f64(double [[TMP4]]) #[[ATTR3]]
+; TFA_INTERLEAVE-NEXT:    [[TMP11:%.*]] = insertelement <2 x double> poison, double [[TMP9]], i32 0
+; TFA_INTERLEAVE-NEXT:    [[TMP12:%.*]] = insertelement <2 x double> [[TMP11]], double [[TMP10]], i32 1
+; TFA_INTERLEAVE-NEXT:    [[TMP13:%.*]] = fcmp ogt <2 x double> [[TMP8]], zeroinitializer
+; TFA_INTERLEAVE-NEXT:    [[TMP14:%.*]] = fcmp ogt <2 x double> [[TMP12]], zeroinitializer
+; TFA_INTERLEAVE-NEXT:    [[TMP15:%.*]] = xor <2 x i1> [[TMP13]], splat (i1 true)
+; TFA_INTERLEAVE-NEXT:    [[TMP16:%.*]] = xor <2 x i1> [[TMP14]], splat (i1 true)
+; TFA_INTERLEAVE-NEXT:    [[TMP17:%.*]] = select <2 x i1> [[ACTIVE_LANE_MASK]], <2 x i1> [[TMP15]], <2 x i1> zeroinitializer
+; TFA_INTERLEAVE-NEXT:    [[TMP18:%.*]] = select <2 x i1> [[ACTIVE_LANE_MASK2]], <2 x i1> [[TMP16]], <2 x i1> zeroinitializer
+; TFA_INTERLEAVE-NEXT:    [[PREDPHI:%.*]] = select <2 x i1> [[TMP17]], <2 x double> splat (double 1.000000e+00), <2 x double> zeroinitializer
+; TFA_INTERLEAVE-NEXT:    [[PREDPHI3:%.*]] = select <2 x i1> [[TMP18]], <2 x double> splat (double 1.000000e+00), <2 x double> zeroinitializer
+; TFA_INTERLEAVE-NEXT:    [[TMP19:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK]], i32 0
+; TFA_INTERLEAVE-NEXT:    br i1 [[TMP19]], label [[PRED_STORE_IF:%.*]], label [[PRED_STORE_CONTINUE:%.*]]
+; TFA_INTERLEAVE:       pred.store.if:
+; TFA_INTERLEAVE-NEXT:    [[TMP20:%.*]] = extractelement <2 x double> [[PREDPHI]], i32 0
+; TFA_INTERLEAVE-NEXT:    store double [[TMP20]], ptr [[P:%.*]], align 8
+; TFA_INTERLEAVE-NEXT:    br label [[PRED_STORE_CONTINUE]]
+; TFA_INTERLEAVE:       pred.store.continue:
+; TFA_INTERLEAVE-NEXT:    [[TMP29:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK]], i32 1
+; TFA_INTERLEAVE-NEXT:    br i1 [[TMP29]], label [[PRED_STORE_IF4:%.*]], label [[PRED_STORE_CONTINUE5:%.*]]
+; TFA_INTERLEAVE:       pred.store.if4:
+; TFA_INTERLEAVE-NEXT:    [[TMP22:%.*]] = extractelement <2 x double> [[PREDPHI]], i32 1
+; TFA_INTERLEAVE-NEXT:    store double [[TMP22]], ptr [[P]], align 8
+; TFA_INTERLEAVE-NEXT:    br label [[PRED_STORE_CONTINUE5]]
+; TFA_INTERLEAVE:       pred.store.continue5:
+; TFA_INTERLEAVE-NEXT:    [[TMP31:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK2]], i32 0
+; TFA_INTERLEAVE-NEXT:    br i1 [[TMP31]], label [[PRED_STORE_IF6:%.*]], label [[PRED_STORE_CONTINUE7:%.*]]
+; TFA_INTERLEAVE:       pred.store.if6:
+; TFA_INTERLEAVE-NEXT:    [[TMP32:%.*]] = extractelement <2 x double> [[PREDPHI3]], i32 0
+; TFA_INTERLEAVE-NEXT:    store double [[TMP32]], ptr [[P]], align 8
+; TFA_INTERLEAVE-NEXT:    br label [[PRED_STORE_CONTINUE7]]
+; TFA_INTERLEAVE:       pred.store.continue7:
+; TFA_INTERLEAVE-NEXT:    [[TMP25:%.*]] = extractelement <2 x i1> [[ACTIVE_LANE_MASK2]], i32 1
+; TFA_INTERLEAVE-NEXT:    br i1 [[TMP25]], label [[PRED_STORE_IF8:%.*]], label [[PRED_STORE_CONTINUE9]]
+; TFA_INTERLEAVE:       pred.store.if8:
+; TFA_INTERLEAVE-NEXT:    [[TMP34:%.*]] = extractelement <2 x double> [[PREDPHI3]], i32 1
+; TFA_INTERLEAVE-NEXT:    store double [[TMP34]], ptr [[P]], align 8
+; TFA_INTERLEAVE-NEXT:    br label [[PRED_STORE_CONTINUE9]]
+; TFA_INTERLEAVE:       pred.store.continue9:
+; TFA_INTERLEAVE-NEXT:    [[INDEX_NEXT]] = add i64 [[INDEX]], 4
+; TFA_INTERLEAVE-NEXT:    [[TMP27:%.*]] = add i64 [[INDEX]], 2
+; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_NEXT]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 [[INDEX]], i64 [[TMP3]])
+; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_NEXT10]] = call <2 x i1> @llvm.get.active.lane.mask.v2i1.i64(i64 [[TMP27]], i64 [[TMP3]])
+; TFA_INTERLEAVE-NEXT:    [[TMP28:%.*]] = xor <2 x i1> [[ACTIVE_LANE_MASK_NEXT]], splat (i1 true)
+; TFA_INTERLEAVE-NEXT:    [[TMP30:%.*]] = extractelement <2 x i1> [[TMP28]], i32 0
+; TFA_INTERLEAVE-NEXT:    br i1 [[TMP30]], label [[END:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; TFA_INTERLEAVE:       end:
+; TFA_INTERLEAVE-NEXT:    ret void
+;
+entry:
+  br label %loop
+
+loop:
+  %iv = phi i64 [ 0, %entry ], [ %iv.next, %loop.end ]
+  %ld = load double, ptr %p2, align 8
+  %exp = tail call double @llvm.exp.f64(double %ld) #6
+  %cond1 = fcmp ogt double %exp, 0.000000e+00
+  br i1 %cond1, label %loop.middle, label %loop.end
+
+loop.middle:
+  br label %loop.end
+
+loop.end:
+  %sink = phi double [ 0.000000e+00, %loop.middle ], [ 1.000000e+00, %loop ]
+  store double %sink, ptr %p, align 8
+  %iv.next = add i64 %iv, 1
+  %cond2 = icmp eq i64 %iv, %n
+  br i1 %cond2, label %end, label %loop, !llvm.loop !0
+
+end:
+  ret void
+}
+
+
+declare double @llvm.exp.f64(double)
+
+; fixed-width variant of exp
+declare <2 x double> @exp_fixed(<2 x double>)
+
+;; scalable vector variant of exp
+declare <vscale x 2 x double> @exp_masked_scalable(<vscale x 2 x double>, <vscale x 2 x i1>)
+
+attributes #5 = { "target-cpu"="neoverse-v2" vscale_range(1,16) }
+attributes #6 = { "vector-function-abi-variant"="_ZGV_LLVM_N2v_llvm.exp.f64(exp_fixed),_ZGVsMxv_llvm.exp.f64(exp_masked_scalable)" }
+
+!0 = distinct !{!0, !1}
+!1 = !{!"llvm.loop.vectorize.width", i32 2}
+;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
+; TFALWAYS: {{.*}}
+; TFFALLBACK: {{.*}}