Unverified Commit 3d893f3e authored by David Sherwood's avatar David Sherwood Committed by GitHub
Browse files

[LV][NFC] Remove instsimplify pass run from all tests (#193722) 

The instsimplify pass was only giving minor incidental improvements that
aren't essential to what is being tested.
parent 6b25ae4f

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

+46 −25
Original line number Diff line number Diff line 
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --check-globals none --version 5

; 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 -tail-folding-policy=must-fold-tail -S | FileCheck %s --check-prefixes=TFCOMMON,TFALWAYS

; RUN: opt < %s -passes=loop-vectorize,instsimplify,simplifycfg -force-vector-interleave=1 -tail-folding-policy=prefer-fold-tail -S | FileCheck %s --check-prefixes=TFCOMMON,TFFALLBACK

; RUN: opt < %s -passes=loop-vectorize,instsimplify,simplifycfg -force-vector-interleave=2 -tail-folding-policy=must-fold-tail -S | FileCheck %s --check-prefixes=TFA_INTERLEAVE

; RUN: opt < %s -passes=loop-vectorize -force-vector-interleave=1 -S | FileCheck %s --check-prefixes=TFNONE

; RUN: opt < %s -passes=loop-vectorize,simplifycfg -force-vector-interleave=1 -tail-folding-policy=must-fold-tail -S | FileCheck %s --check-prefixes=TFCOMMON,TFALWAYS

; RUN: opt < %s -passes=loop-vectorize,simplifycfg -force-vector-interleave=1 -tail-folding-policy=prefer-fold-tail -S | FileCheck %s --check-prefixes=TFCOMMON,TFFALLBACK

; RUN: opt < %s -passes=loop-vectorize,simplifycfg -force-vector-interleave=2 -tail-folding-policy=must-fold-tail -S | FileCheck %s --check-prefixes=TFA_INTERLEAVE



target triple = "aarch64-unknown-linux-gnu"
@@ -30,7 +30,8 @@ define void @test_widen(ptr noalias %a, ptr readnone %b) #4 { 
; TFNONE-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]

; TFNONE-NEXT:    br i1 [[TMP9]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]

; TFNONE:       [[MIDDLE_BLOCK]]:

; TFNONE-NEXT:    br i1 false, label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1025, [[N_VEC]]

; TFNONE-NEXT:    br i1 [[CMP_N]], label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE:       [[SCALAR_PH]]:

; TFNONE-NEXT:    br label %[[FOR_BODY:.*]]

; TFNONE:       [[FOR_BODY]]:
@@ -51,10 +52,11 @@ define void @test_widen(ptr noalias %a, ptr readnone %b) #4 { 
; TFCOMMON-NEXT:  [[ENTRY:.*]]:

; TFCOMMON-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFCOMMON-NEXT:    [[TMP4:%.*]] = shl nuw i64 [[TMP3]], 1

; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFCOMMON-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFCOMMON:       [[VECTOR_BODY]]:

; TFCOMMON-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFCOMMON-NEXT:    [[TMP5:%.*]] = getelementptr i64, ptr [[B]], i64 [[INDEX]]

; TFCOMMON-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x i64> @llvm.masked.load.nxv2i64.p0(ptr align 8 [[TMP5]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x i64> poison)

; TFCOMMON-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x i64> @foo_vector(<vscale x 2 x i64> [[WIDE_MASKED_LOAD]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
@@ -74,11 +76,12 @@ define void @test_widen(ptr noalias %a, ptr readnone %b) #4 { 
; TFA_INTERLEAVE-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFA_INTERLEAVE-NEXT:    [[TMP5:%.*]] = shl nuw i64 [[TMP3]], 1

; TFA_INTERLEAVE-NEXT:    [[TMP6:%.*]] = shl nuw i64 [[TMP5]], 1

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY1:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[TMP5]], i64 1025)

; TFA_INTERLEAVE-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFA_INTERLEAVE:       [[VECTOR_BODY]]:

; TFA_INTERLEAVE-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK2:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY1]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT4:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[TMP7:%.*]] = getelementptr i64, ptr [[B]], i64 [[INDEX]]

; TFA_INTERLEAVE-NEXT:    [[TMP10:%.*]] = getelementptr i64, ptr [[TMP7]], i64 [[TMP5]]
@@ -143,7 +146,8 @@ define void @test_if_then(ptr noalias %a, ptr readnone %b) #4 { 
; TFNONE-NEXT:    [[TMP11:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]

; TFNONE-NEXT:    br i1 [[TMP11]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]

; TFNONE:       [[MIDDLE_BLOCK]]:

; TFNONE-NEXT:    br i1 false, label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1025, [[N_VEC]]

; TFNONE-NEXT:    br i1 [[CMP_N]], label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE:       [[SCALAR_PH]]:

; TFNONE-NEXT:    br label %[[FOR_BODY:.*]]

; TFNONE:       [[FOR_BODY]]:
@@ -170,10 +174,11 @@ define void @test_if_then(ptr noalias %a, ptr readnone %b) #4 { 
; TFCOMMON-NEXT:  [[ENTRY:.*]]:

; TFCOMMON-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFCOMMON-NEXT:    [[TMP4:%.*]] = shl nuw i64 [[TMP3]], 1

; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFCOMMON-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFCOMMON:       [[VECTOR_BODY]]:

; TFCOMMON-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFCOMMON-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[INDEX]]

; TFCOMMON-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x i64> @llvm.masked.load.nxv2i64.p0(ptr align 8 [[TMP5]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x i64> poison)

; TFCOMMON-NEXT:    [[TMP6:%.*]] = icmp ugt <vscale x 2 x i64> [[WIDE_MASKED_LOAD]], splat (i64 50)
@@ -196,11 +201,12 @@ define void @test_if_then(ptr noalias %a, ptr readnone %b) #4 { 
; TFA_INTERLEAVE-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFA_INTERLEAVE-NEXT:    [[TMP5:%.*]] = shl nuw i64 [[TMP3]], 1

; TFA_INTERLEAVE-NEXT:    [[TMP6:%.*]] = shl nuw i64 [[TMP5]], 1

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY1:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[TMP5]], i64 1025)

; TFA_INTERLEAVE-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFA_INTERLEAVE:       [[VECTOR_BODY]]:

; TFA_INTERLEAVE-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK2:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY1]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT5:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[INDEX]]

; TFA_INTERLEAVE-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i64, ptr [[TMP7]], i64 [[TMP5]]
@@ -284,7 +290,8 @@ define void @test_widen_if_then_else(ptr noalias %a, ptr readnone %b) #4 { 
; TFNONE-NEXT:    [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]

; TFNONE-NEXT:    br i1 [[TMP12]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]

; TFNONE:       [[MIDDLE_BLOCK]]:

; TFNONE-NEXT:    br i1 false, label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1025, [[N_VEC]]

; TFNONE-NEXT:    br i1 [[CMP_N]], label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE:       [[SCALAR_PH]]:

; TFNONE-NEXT:    br label %[[FOR_BODY:.*]]

; TFNONE:       [[FOR_BODY]]:
@@ -314,10 +321,11 @@ define void @test_widen_if_then_else(ptr noalias %a, ptr readnone %b) #4 { 
; TFCOMMON-NEXT:  [[ENTRY:.*]]:

; TFCOMMON-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFCOMMON-NEXT:    [[TMP4:%.*]] = shl nuw i64 [[TMP3]], 1

; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFCOMMON-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFCOMMON:       [[VECTOR_BODY]]:

; TFCOMMON-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFCOMMON-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFCOMMON-NEXT:    [[TMP5:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[INDEX]]

; TFCOMMON-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x i64> @llvm.masked.load.nxv2i64.p0(ptr align 8 [[TMP5]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x i64> poison)

; TFCOMMON-NEXT:    [[TMP6:%.*]] = icmp ugt <vscale x 2 x i64> [[WIDE_MASKED_LOAD]], splat (i64 50)
@@ -343,11 +351,12 @@ define void @test_widen_if_then_else(ptr noalias %a, ptr readnone %b) #4 { 
; TFA_INTERLEAVE-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFA_INTERLEAVE-NEXT:    [[TMP5:%.*]] = shl nuw i64 [[TMP3]], 1

; TFA_INTERLEAVE-NEXT:    [[TMP6:%.*]] = shl nuw i64 [[TMP5]], 1

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY1:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[TMP5]], i64 1025)

; TFA_INTERLEAVE-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFA_INTERLEAVE:       [[VECTOR_BODY]]:

; TFA_INTERLEAVE-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK2:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY1]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT5:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[TMP7:%.*]] = getelementptr inbounds i64, ptr [[A]], i64 [[INDEX]]

; TFA_INTERLEAVE-NEXT:    [[TMP10:%.*]] = getelementptr inbounds i64, ptr [[TMP7]], i64 [[TMP5]]
@@ -436,7 +445,8 @@ define void @test_widen_nomask(ptr noalias %a, ptr readnone %b) #4 { 
; TFNONE-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]

; TFNONE-NEXT:    br i1 [[TMP9]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]

; TFNONE:       [[MIDDLE_BLOCK]]:

; TFNONE-NEXT:    br i1 false, label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1025, [[N_VEC]]

; TFNONE-NEXT:    br i1 [[CMP_N]], label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE:       [[SCALAR_PH]]:

; TFNONE-NEXT:    br label %[[FOR_BODY:.*]]

; TFNONE:       [[FOR_BODY]]:
@@ -486,9 +496,12 @@ define void @test_widen_nomask(ptr noalias %a, ptr readnone %b) #4 { 
; TFFALLBACK-NEXT:    store <vscale x 2 x i64> [[TMP7]], ptr [[TMP8]], align 8

; TFFALLBACK-NEXT:    [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP3]]

; TFFALLBACK-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]

; TFFALLBACK-NEXT:    br i1 [[TMP9]], label %[[FOR_BODY:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]

; TFFALLBACK-NEXT:    br i1 [[TMP9]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]

; TFFALLBACK:       [[MIDDLE_BLOCK]]:

; TFFALLBACK-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1025, [[N_VEC]]

; TFFALLBACK-NEXT:    br i1 [[CMP_N]], label %[[FOR_COND_CLEANUP:.*]], label %[[FOR_BODY:.*]]

; TFFALLBACK:       [[FOR_BODY]]:

; TFFALLBACK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], %[[FOR_BODY]] ], [ [[N_VEC]], %[[VECTOR_BODY]] ]

; TFFALLBACK-NEXT:    [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], %[[FOR_BODY]] ], [ [[N_VEC]], %[[MIDDLE_BLOCK]] ]

; TFFALLBACK-NEXT:    [[GEP:%.*]] = getelementptr i64, ptr [[B]], i64 [[INDVARS_IV]]

; TFFALLBACK-NEXT:    [[LOAD:%.*]] = load i64, ptr [[GEP]], align 8

; TFFALLBACK-NEXT:    [[CALL:%.*]] = call i64 @foo(i64 [[LOAD]]) #[[ATTR6:[0-9]+]]
@@ -496,7 +509,7 @@ define void @test_widen_nomask(ptr noalias %a, ptr readnone %b) #4 { 
; TFFALLBACK-NEXT:    store i64 [[CALL]], ptr [[ARRAYIDX]], align 8

; TFFALLBACK-NEXT:    [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1

; TFFALLBACK-NEXT:    [[EXITCOND:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], 1025

; TFFALLBACK-NEXT:    br i1 [[EXITCOND]], label %[[FOR_COND_CLEANUP:.*]], label %[[FOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]

; TFFALLBACK-NEXT:    br i1 [[EXITCOND]], label %[[FOR_COND_CLEANUP]], label %[[FOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]

; TFFALLBACK:       [[FOR_COND_CLEANUP]]:

; TFFALLBACK-NEXT:    ret void

;
@@ -560,7 +573,8 @@ define void @test_widen_optmask(ptr noalias %a, ptr readnone %b) #4 { 
; TFNONE-NEXT:    [[TMP9:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]

; TFNONE-NEXT:    br i1 [[TMP9]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]

; TFNONE:       [[MIDDLE_BLOCK]]:

; TFNONE-NEXT:    br i1 false, label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1025, [[N_VEC]]

; TFNONE-NEXT:    br i1 [[CMP_N]], label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE:       [[SCALAR_PH]]:

; TFNONE-NEXT:    br label %[[FOR_BODY:.*]]

; TFNONE:       [[FOR_BODY]]:
@@ -581,10 +595,11 @@ define void @test_widen_optmask(ptr noalias %a, ptr readnone %b) #4 { 
; TFALWAYS-NEXT:  [[ENTRY:.*]]:

; TFALWAYS-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFALWAYS-NEXT:    [[TMP4:%.*]] = shl nuw i64 [[TMP3]], 1

; TFALWAYS-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFALWAYS-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFALWAYS:       [[VECTOR_BODY]]:

; TFALWAYS-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFALWAYS-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFALWAYS-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFALWAYS-NEXT:    [[TMP5:%.*]] = getelementptr i64, ptr [[B]], i64 [[INDEX]]

; TFALWAYS-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x i64> @llvm.masked.load.nxv2i64.p0(ptr align 8 [[TMP5]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x i64> poison)

; TFALWAYS-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x i64> @foo_vector(<vscale x 2 x i64> [[WIDE_MASKED_LOAD]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
@@ -603,10 +618,11 @@ define void @test_widen_optmask(ptr noalias %a, ptr readnone %b) #4 { 
; TFFALLBACK-NEXT:  [[ENTRY:.*]]:

; TFFALLBACK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFFALLBACK-NEXT:    [[TMP4:%.*]] = shl nuw i64 [[TMP3]], 1

; TFFALLBACK-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFFALLBACK-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFFALLBACK:       [[VECTOR_BODY]]:

; TFFALLBACK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFFALLBACK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFFALLBACK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFFALLBACK-NEXT:    [[TMP5:%.*]] = getelementptr i64, ptr [[B]], i64 [[INDEX]]

; TFFALLBACK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x i64> @llvm.masked.load.nxv2i64.p0(ptr align 8 [[TMP5]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x i64> poison)

; TFFALLBACK-NEXT:    [[TMP6:%.*]] = call <vscale x 2 x i64> @foo_vector(<vscale x 2 x i64> [[WIDE_MASKED_LOAD]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]])
@@ -626,11 +642,12 @@ define void @test_widen_optmask(ptr noalias %a, ptr readnone %b) #4 { 
; TFA_INTERLEAVE-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFA_INTERLEAVE-NEXT:    [[TMP5:%.*]] = shl nuw i64 [[TMP3]], 1

; TFA_INTERLEAVE-NEXT:    [[TMP6:%.*]] = shl nuw i64 [[TMP5]], 1

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY1:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[TMP5]], i64 1025)

; TFA_INTERLEAVE-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFA_INTERLEAVE:       [[VECTOR_BODY]]:

; TFA_INTERLEAVE-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK2:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY1]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT4:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[TMP7:%.*]] = getelementptr i64, ptr [[B]], i64 [[INDEX]]

; TFA_INTERLEAVE-NEXT:    [[TMP10:%.*]] = getelementptr i64, ptr [[TMP7]], i64 [[TMP5]]
@@ -701,7 +718,8 @@ define double @test_widen_fmuladd_and_call(ptr noalias %a, ptr readnone %b, doub 
; TFNONE-NEXT:    [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]

; TFNONE-NEXT:    br i1 [[TMP12]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]

; TFNONE:       [[MIDDLE_BLOCK]]:

; TFNONE-NEXT:    br i1 false, label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1025, [[N_VEC]]

; TFNONE-NEXT:    br i1 [[CMP_N]], label %[[FOR_COND_CLEANUP:.*]], label %[[SCALAR_PH:.*]]

; TFNONE:       [[SCALAR_PH]]:

; TFNONE-NEXT:    br label %[[FOR_BODY:.*]]

; TFNONE:       [[FOR_BODY]]:
@@ -726,12 +744,13 @@ define double @test_widen_fmuladd_and_call(ptr noalias %a, ptr readnone %b, doub 
; TFALWAYS-NEXT:  [[ENTRY:.*]]:

; TFALWAYS-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFALWAYS-NEXT:    [[TMP4:%.*]] = shl nuw i64 [[TMP3]], 1

; TFALWAYS-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFALWAYS-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 2 x double> poison, double [[M]], i64 0

; TFALWAYS-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 2 x double> [[BROADCAST_SPLATINSERT]], <vscale x 2 x double> poison, <vscale x 2 x i32> zeroinitializer

; TFALWAYS-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFALWAYS:       [[VECTOR_BODY]]:

; TFALWAYS-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFALWAYS-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFALWAYS-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFALWAYS-NEXT:    [[VEC_PHI:%.*]] = phi double [ 0.000000e+00, %[[ENTRY]] ], [ [[TMP11:%.*]], %[[VECTOR_BODY]] ]

; TFALWAYS-NEXT:    [[TMP5:%.*]] = getelementptr double, ptr [[B]], i64 [[INDEX]]

; TFALWAYS-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x double> @llvm.masked.load.nxv2f64.p0(ptr align 8 [[TMP5]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x double> poison)
@@ -755,12 +774,13 @@ define double @test_widen_fmuladd_and_call(ptr noalias %a, ptr readnone %b, doub 
; TFFALLBACK-NEXT:  [[ENTRY:.*]]:

; TFFALLBACK-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFFALLBACK-NEXT:    [[TMP4:%.*]] = shl nuw i64 [[TMP3]], 1

; TFFALLBACK-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFFALLBACK-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 2 x double> poison, double [[M]], i64 0

; TFFALLBACK-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 2 x double> [[BROADCAST_SPLATINSERT]], <vscale x 2 x double> poison, <vscale x 2 x i32> zeroinitializer

; TFFALLBACK-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFFALLBACK:       [[VECTOR_BODY]]:

; TFFALLBACK-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFFALLBACK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFFALLBACK-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFFALLBACK-NEXT:    [[VEC_PHI:%.*]] = phi double [ 0.000000e+00, %[[ENTRY]] ], [ [[TMP11:%.*]], %[[VECTOR_BODY]] ]

; TFFALLBACK-NEXT:    [[TMP5:%.*]] = getelementptr double, ptr [[B]], i64 [[INDEX]]

; TFFALLBACK-NEXT:    [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 2 x double> @llvm.masked.load.nxv2f64.p0(ptr align 8 [[TMP5]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK]], <vscale x 2 x double> poison)
@@ -785,13 +805,14 @@ define double @test_widen_fmuladd_and_call(ptr noalias %a, ptr readnone %b, doub 
; TFA_INTERLEAVE-NEXT:    [[TMP3:%.*]] = call i64 @llvm.vscale.i64()

; TFA_INTERLEAVE-NEXT:    [[TMP5:%.*]] = shl nuw i64 [[TMP3]], 1

; TFA_INTERLEAVE-NEXT:    [[TMP6:%.*]] = shl nuw i64 [[TMP5]], 1

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 0, i64 1025)

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK_ENTRY1:%.*]] = call <vscale x 2 x i1> @llvm.get.active.lane.mask.nxv2i1.i64(i64 [[TMP5]], i64 1025)

; TFA_INTERLEAVE-NEXT:    [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 2 x double> poison, double [[M]], i64 0

; TFA_INTERLEAVE-NEXT:    [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 2 x double> [[BROADCAST_SPLATINSERT]], <vscale x 2 x double> poison, <vscale x 2 x i32> zeroinitializer

; TFA_INTERLEAVE-NEXT:    br label %[[VECTOR_BODY:.*]]

; TFA_INTERLEAVE:       [[VECTOR_BODY]]:

; TFA_INTERLEAVE-NEXT:    [[INDEX:%.*]] = phi i64 [ 0, %[[ENTRY]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ splat (i1 true), %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[ACTIVE_LANE_MASK2:%.*]] = phi <vscale x 2 x i1> [ [[ACTIVE_LANE_MASK_ENTRY1]], %[[ENTRY]] ], [ [[ACTIVE_LANE_MASK_NEXT4:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[VEC_PHI:%.*]] = phi double [ 0.000000e+00, %[[ENTRY]] ], [ [[TMP24:%.*]], %[[VECTOR_BODY]] ]

; TFA_INTERLEAVE-NEXT:    [[TMP7:%.*]] = getelementptr double, ptr [[B]], i64 [[INDEX]]

llvm/test/Transforms/LoopVectorize/AArch64/wider-VF-for-callinst.ll

+6 −4
Original line number Diff line number Diff line 
; 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=WIDE

; RUN: opt < %s -passes=loop-vectorize,instsimplify -force-vector-interleave=1 -vectorizer-maximize-bandwidth-for-vector-calls=false -S | FileCheck %s --check-prefixes=NARROW

; RUN: opt < %s -passes=loop-vectorize -force-vector-interleave=1 -S | FileCheck %s --check-prefixes=WIDE

; RUN: opt < %s -passes=loop-vectorize -force-vector-interleave=1 -vectorizer-maximize-bandwidth-for-vector-calls=false -S | FileCheck %s --check-prefixes=NARROW



target triple = "aarch64-unknown-linux-gnu"
@@ -26,7 +26,8 @@ define void @test_widen(ptr noalias %a, ptr readnone %b) #1 { 
; WIDE-NEXT:    [[TMP10:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]

; WIDE-NEXT:    br i1 [[TMP10]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]

; WIDE:       middle.block:

; WIDE-NEXT:    br i1 false, label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH:%.*]]

; WIDE-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1025, [[N_VEC]]

; WIDE-NEXT:    br i1 [[CMP_N]], label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH:%.*]]

; WIDE:       scalar.ph:

; WIDE-NEXT:    br label [[FOR_BODY:%.*]]

; WIDE:       for.body:
@@ -64,7 +65,8 @@ define void @test_widen(ptr noalias %a, ptr readnone %b) #1 { 
; NARROW-NEXT:    [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]

; NARROW-NEXT:    br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]

; NARROW:       middle.block:

; NARROW-NEXT:    br i1 false, label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH:%.*]]

; NARROW-NEXT:    [[CMP_N:%.*]] = icmp eq i64 1025, [[N_VEC]]

; NARROW-NEXT:    br i1 [[CMP_N]], label [[FOR_COND_CLEANUP:%.*]], label [[SCALAR_PH:%.*]]

; NARROW:       scalar.ph:

; NARROW-NEXT:    br label [[FOR_BODY:%.*]]

; NARROW:       for.body:

llvm/test/Transforms/LoopVectorize/hints-trans.ll

+1 −1
Original line number Diff line number Diff line 
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5

; RUN: opt -S -passes=loop-vectorize,simplifycfg -force-vector-interleave=1 -force-vector-width=4 -simplifycfg-require-and-preserve-domtree=1 < %s | FileCheck %s



; Note: -passes=instsimplify,simplifycfg -simplifycfg-require-and-preserve-domtree=1 remove the (now dead) original loop, making

; Note: -passes=simplifycfg -simplifycfg-require-and-preserve-domtree=1 remove the (now dead) original loop, making

; it easy to test that the llvm.loop.unroll.disable hint is still present.



define void @scalar_loop_dead_after_vectorization(ptr nocapture %b) {

llvm/test/Transforms/LoopVectorize/scalable-trunc-min-bitwidth.ll

+1 −3

File changed.

Preview size limit exceeded, changes collapsed.