Page MenuHomePhabricator

Please use GitHub pull requests for new patches. Phabricator shutdown timeline

Differential D142894 Diff 501488 llvm/test/Transforms/LoopVectorize/AArch64/first-order-recurrence-fold-tail.ll

Changeset View

Standalone View

View Options

llvm/test/Transforms/LoopVectorize/AArch64/first-order-recurrence-fold-tail.ll

Show First 20 Lines • Show All 57 Lines • ▼ Show 20 Lines
; CHECK: pred.load.continue6: ; CHECK: pred.load.continue6:
; CHECK-NEXT: [[TMP24]] = phi <4 x i16> [ [[TMP18]], [[PRED_LOAD_CONTINUE4]] ], [ [[TMP23]], [[PRED_LOAD_IF5]] ] ; CHECK-NEXT: [[TMP24]] = phi <4 x i16> [ [[TMP18]], [[PRED_LOAD_CONTINUE4]] ], [ [[TMP23]], [[PRED_LOAD_IF5]] ]
; CHECK-NEXT: [[TMP25:%.*]] = shufflevector <4 x i16> [[VECTOR_RECUR]], <4 x i16> [[TMP24]], <4 x i32> <i32 3, i32 4, i32 5, i32 6> ; CHECK-NEXT: [[TMP25:%.*]] = shufflevector <4 x i16> [[VECTOR_RECUR]], <4 x i16> [[TMP24]], <4 x i32> <i32 3, i32 4, i32 5, i32 6>
; CHECK-NEXT: [[TMP26:%.*]] = sext <4 x i16> [[TMP25]] to <4 x i32> ; CHECK-NEXT: [[TMP26:%.*]] = sext <4 x i16> [[TMP25]] to <4 x i32>
; CHECK-NEXT: [[TMP27:%.*]] = extractelement <4 x i64> [[TMP0]], i32 0 ; CHECK-NEXT: [[TMP27:%.*]] = extractelement <4 x i64> [[TMP0]], i32 0
; CHECK-NEXT: [[TMP28:%.*]] = getelementptr i32, ptr [[B:%.*]], i64 [[TMP27]] ; CHECK-NEXT: [[TMP28:%.*]] = getelementptr i32, ptr [[B:%.*]], i64 [[TMP27]]
; CHECK-NEXT: [[TMP29:%.*]] = getelementptr i32, ptr [[TMP28]], i32 0 ; CHECK-NEXT: [[TMP29:%.*]] = getelementptr i32, ptr [[TMP28]], i32 0
; CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[TMP26]], ptr [[TMP29]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]]) ; CHECK-NEXT: call void @llvm.masked.store.v4i32.p0(<4 x i32> [[TMP26]], ptr [[TMP29]], i32 4, <4 x i1> [[ACTIVE_LANE_MASK]])
; CHECK-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i64(i64 [[INDEX]], i64 998)
; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 4 ; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 4
; CHECK-NEXT: [[ACTIVE_LANE_MASK_NEXT]] = call <4 x i1> @llvm.get.active.lane.mask.v4i1.i64(i64 [[INDEX_NEXT]], i64 1002)
; CHECK-NEXT: [[TMP30:%.*]] = xor <4 x i1> [[ACTIVE_LANE_MASK_NEXT]], <i1 true, i1 true, i1 true, i1 true> ; CHECK-NEXT: [[TMP30:%.*]] = xor <4 x i1> [[ACTIVE_LANE_MASK_NEXT]], <i1 true, i1 true, i1 true, i1 true>
; CHECK-NEXT: [[VEC_IND_NEXT]] = add <4 x i64> [[VEC_IND]], <i64 4, i64 4, i64 4, i64 4> ; CHECK-NEXT: [[VEC_IND_NEXT]] = add <4 x i64> [[VEC_IND]], <i64 4, i64 4, i64 4, i64 4>
; CHECK-NEXT: [[TMP31:%.*]] = extractelement <4 x i1> [[TMP30]], i32 0 ; CHECK-NEXT: [[TMP31:%.*]] = extractelement <4 x i1> [[TMP30]], i32 0
; CHECK-NEXT: br i1 [[TMP31]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]] ; CHECK-NEXT: br i1 [[TMP31]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: middle.block: ; CHECK: middle.block:
; CHECK-NEXT: [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <4 x i16> [[TMP24]], i32 3 ; CHECK-NEXT: [[VECTOR_RECUR_EXTRACT:%.*]] = extractelement <4 x i16> [[TMP24]], i32 3
; CHECK-NEXT: [[VECTOR_RECUR_EXTRACT_FOR_PHI:%.*]] = extractelement <4 x i16> [[TMP24]], i32 2 ; CHECK-NEXT: [[VECTOR_RECUR_EXTRACT_FOR_PHI:%.*]] = extractelement <4 x i16> [[TMP24]], i32 2
; CHECK-NEXT: br i1 true, label [[EXIT:%.*]], label [[SCALAR_PH]] ; CHECK-NEXT: br i1 true, label [[EXIT:%.*]], label [[SCALAR_PH]]
Show All 36 Lines