shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions 6 Packages Projects Releases Wiki Activity
llvm-project/llvm/test/Transforms/LoopVectorize/X86/pr54634.ll
Florian Hahn 4277c21059
[VPlan] Introduce explicit broadcasts for live-ins. (#124644) 
Add a new VPInstruction::Broadcast opcode and use it to materialize
explicit broadcasts of live-ins. The initial patch only materlizes the
broadcasts if the vector preheader dominates all uses that need it.
Later patches will pick the best valid insert point, thus retiring
implicit hoisting of broadcasts from VPTransformsState::get().

PR: https://github.com/llvm/llvm-project/pull/124644
2025-02-26 13:57:51 +00:00

152 lines
13 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -S -passes=loop-vectorize < %s -mcpu=skylake | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128-ni:10:11:12:13"
target triple = "x86_64-unknown-linux-gnu"
@jlplt_ijl_alloc_array_1d_10294_got = external dso_local local_unnamed_addr global ptr
define ptr addrspace(10) @japi1_vect_42283(ptr nocapture readonly %0, i32 %1) local_unnamed_addr #0 {
; CHECK-LABEL: @japi1_vect_42283(
; CHECK-NEXT: iter.check:
; CHECK-NEXT: [[TMP2:%.*]] = sext i32 [[TMP1:%.*]] to i64
; CHECK-NEXT: [[TMP3:%.*]] = load atomic ptr, ptr @jlplt_ijl_alloc_array_1d_10294_got unordered, align 8
; CHECK-NEXT: [[TMP4:%.*]] = tail call ptr addrspace(10) [[TMP3]](ptr addrspace(10) null, i64 0)
; CHECK-NEXT: [[TMP5:%.*]] = load ptr addrspace(10), ptr [[TMP0:%.*]], align 8, !tbaa [[TBAA0:![0-9]+]]
; CHECK-NEXT: [[TMP6:%.*]] = addrspacecast ptr addrspace(10) [[TMP4]] to ptr addrspace(11)
; CHECK-NEXT: [[TMP7:%.*]] = load ptr addrspace(13), ptr addrspace(11) [[TMP6]], align 8, !tbaa [[TBAA5:![0-9]+]]
; CHECK-NEXT: [[DOTELT:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(10) [[TMP5]], i64 0, i32 0
; CHECK-NEXT: [[DOTUNPACK:%.*]] = load ptr addrspace(10), ptr addrspace(10) [[DOTELT]], align 8, !tbaa [[TBAA8:![0-9]+]]
; CHECK-NEXT: [[DOTELT1:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(10) [[TMP5]], i64 0, i32 1
; CHECK-NEXT: [[DOTUNPACK2:%.*]] = load i64, ptr addrspace(10) [[DOTELT1]], align 8, !tbaa [[TBAA8]]
; CHECK-NEXT: [[TMP8:%.*]] = add nsw i64 [[TMP2]], 1
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP8]], 4
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[TOP:%.*]]
; CHECK: vector.main.loop.iter.check:
; CHECK-NEXT: [[TMP17:%.*]] = icmp ult i64 [[TMP8]], 16
; CHECK-NEXT: br i1 [[TMP17]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[TMP8]], 16
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[TMP8]], [[N_MOD_VF]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x ptr addrspace(10)> poison, ptr addrspace(10) [[DOTUNPACK]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x ptr addrspace(10)> [[BROADCAST_SPLATINSERT]], <4 x ptr addrspace(10)> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: [[BROADCAST_SPLATINSERT7:%.*]] = insertelement <4 x i64> poison, i64 [[DOTUNPACK2]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT8:%.*]] = shufflevector <4 x i64> [[BROADCAST_SPLATINSERT7]], <4 x i64> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[VEC_IND:%.*]] = phi <4 x i64> [ <i64 0, i64 1, i64 2, i64 3>, [[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[STEP_ADD:%.*]] = add <4 x i64> [[VEC_IND]], splat (i64 4)
; CHECK-NEXT: [[STEP_ADD4:%.*]] = add <4 x i64> [[STEP_ADD]], splat (i64 4)
; CHECK-NEXT: [[STEP_ADD5:%.*]] = add <4 x i64> [[STEP_ADD4]], splat (i64 4)
; CHECK-NEXT: [[TMP18:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[VEC_IND]], i32 0
; CHECK-NEXT: [[TMP19:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[STEP_ADD]], i32 0
; CHECK-NEXT: [[TMP20:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[STEP_ADD4]], i32 0
; CHECK-NEXT: [[TMP21:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[STEP_ADD5]], i32 0
; CHECK-NEXT: call void @llvm.masked.scatter.v4p10.v4p13(<4 x ptr addrspace(10)> [[BROADCAST_SPLAT]], <4 x ptr addrspace(13)> [[TMP18]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10:![0-9]+]]
; CHECK-NEXT: call void @llvm.masked.scatter.v4p10.v4p13(<4 x ptr addrspace(10)> [[BROADCAST_SPLAT]], <4 x ptr addrspace(13)> [[TMP19]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10]]
; CHECK-NEXT: call void @llvm.masked.scatter.v4p10.v4p13(<4 x ptr addrspace(10)> [[BROADCAST_SPLAT]], <4 x ptr addrspace(13)> [[TMP20]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10]]
; CHECK-NEXT: call void @llvm.masked.scatter.v4p10.v4p13(<4 x ptr addrspace(10)> [[BROADCAST_SPLAT]], <4 x ptr addrspace(13)> [[TMP21]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10]]
; CHECK-NEXT: [[TMP22:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[VEC_IND]], i32 1
; CHECK-NEXT: [[TMP23:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[STEP_ADD]], i32 1
; CHECK-NEXT: [[TMP24:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[STEP_ADD4]], i32 1
; CHECK-NEXT: [[TMP25:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[STEP_ADD5]], i32 1
; CHECK-NEXT: call void @llvm.masked.scatter.v4i64.v4p13(<4 x i64> [[BROADCAST_SPLAT8]], <4 x ptr addrspace(13)> [[TMP22]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10]]
; CHECK-NEXT: call void @llvm.masked.scatter.v4i64.v4p13(<4 x i64> [[BROADCAST_SPLAT8]], <4 x ptr addrspace(13)> [[TMP23]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10]]
; CHECK-NEXT: call void @llvm.masked.scatter.v4i64.v4p13(<4 x i64> [[BROADCAST_SPLAT8]], <4 x ptr addrspace(13)> [[TMP24]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10]]
; CHECK-NEXT: call void @llvm.masked.scatter.v4i64.v4p13(<4 x i64> [[BROADCAST_SPLAT8]], <4 x ptr addrspace(13)> [[TMP25]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10]]
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 16
; CHECK-NEXT: [[VEC_IND_NEXT]] = add <4 x i64> [[STEP_ADD5]], splat (i64 4)
; CHECK-NEXT: [[TMP26:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP26]], label [[MIDDLE_BLOCK1:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[TMP8]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[L44:%.*]], label [[MIDDLE_BLOCK:%.*]]
; CHECK: vec.epilog.iter.check:
; CHECK-NEXT: [[N_VEC_REMAINING:%.*]] = sub i64 [[TMP8]], [[N_VEC]]
; CHECK-NEXT: [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp ult i64 [[N_VEC_REMAINING]], 4
; CHECK-NEXT: br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[SCALAR_PH]]
; CHECK: vec.epilog.ph:
; CHECK-NEXT: [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[TOP]] ]
; CHECK-NEXT: [[N_MOD_VF4:%.*]] = urem i64 [[TMP8]], 4
; CHECK-NEXT: [[N_VEC5:%.*]] = sub i64 [[TMP8]], [[N_MOD_VF4]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT10:%.*]] = insertelement <4 x ptr addrspace(10)> poison, ptr addrspace(10) [[DOTUNPACK]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT11:%.*]] = shufflevector <4 x ptr addrspace(10)> [[BROADCAST_SPLATINSERT10]], <4 x ptr addrspace(10)> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: [[BROADCAST_SPLATINSERT12:%.*]] = insertelement <4 x i64> poison, i64 [[DOTUNPACK2]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT13:%.*]] = shufflevector <4 x i64> [[BROADCAST_SPLATINSERT12]], <4 x i64> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <4 x i64> poison, i64 [[VEC_EPILOG_RESUME_VAL]], i64 0
; CHECK-NEXT: [[DOTSPLAT:%.*]] = shufflevector <4 x i64> [[DOTSPLATINSERT]], <4 x i64> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: [[INDUCTION:%.*]] = add <4 x i64> [[DOTSPLAT]], <i64 0, i64 1, i64 2, i64 3>
; CHECK-NEXT: br label [[L26:%.*]]
; CHECK: vec.epilog.vector.body:
; CHECK-NEXT: [[INDEX7:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDEX_NEXT14:%.*]], [[L26]] ]
; CHECK-NEXT: [[VEC_IND8:%.*]] = phi <4 x i64> [ [[INDUCTION]], [[SCALAR_PH]] ], [ [[VEC_IND_NEXT9:%.*]], [[L26]] ]
; CHECK-NEXT: [[TMP28:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[VEC_IND8]], i32 0
; CHECK-NEXT: call void @llvm.masked.scatter.v4p10.v4p13(<4 x ptr addrspace(10)> [[BROADCAST_SPLAT11]], <4 x ptr addrspace(13)> [[TMP28]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10]]
; CHECK-NEXT: [[TMP29:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], <4 x i64> [[VEC_IND8]], i32 1
; CHECK-NEXT: call void @llvm.masked.scatter.v4i64.v4p13(<4 x i64> [[BROADCAST_SPLAT13]], <4 x ptr addrspace(13)> [[TMP29]], i32 8, <4 x i1> splat (i1 true)), !tbaa [[TBAA10]]
; CHECK-NEXT: [[INDEX_NEXT14]] = add nuw i64 [[INDEX7]], 4
; CHECK-NEXT: [[VEC_IND_NEXT9]] = add <4 x i64> [[VEC_IND8]], splat (i64 4)
; CHECK-NEXT: [[TMP30:%.*]] = icmp eq i64 [[INDEX_NEXT14]], [[N_VEC5]]
; CHECK-NEXT: br i1 [[TMP30]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[L26]], !llvm.loop [[LOOP15:![0-9]+]]
; CHECK: vec.epilog.middle.block:
; CHECK-NEXT: [[CMP_N15:%.*]] = icmp eq i64 [[TMP8]], [[N_VEC5]]
; CHECK-NEXT: br i1 [[CMP_N15]], label [[L44]], label [[VEC_EPILOG_SCALAR_PH]]
; CHECK: vec.epilog.scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL6:%.*]] = phi i64 [ [[N_VEC5]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ITER_CHECK:%.*]] ]
; CHECK-NEXT: br label [[L27:%.*]]
; CHECK: L26:
; CHECK-NEXT: [[VALUE_PHI5:%.*]] = phi i64 [ [[BC_RESUME_VAL6]], [[VEC_EPILOG_SCALAR_PH]] ], [ [[TMP27:%.*]], [[L27]] ]
; CHECK-NEXT: [[DOTREPACK:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], i64 [[VALUE_PHI5]], i32 0
; CHECK-NEXT: store ptr addrspace(10) [[DOTUNPACK]], ptr addrspace(13) [[DOTREPACK]], align 8, !tbaa [[TBAA10]]
; CHECK-NEXT: [[DOTREPACK4:%.*]] = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) [[TMP7]], i64 [[VALUE_PHI5]], i32 1
; CHECK-NEXT: store i64 [[DOTUNPACK2]], ptr addrspace(13) [[DOTREPACK4]], align 8, !tbaa [[TBAA10]]
; CHECK-NEXT: [[TMP27]] = add i64 [[VALUE_PHI5]], 1
; CHECK-NEXT: [[DOTNOT:%.*]] = icmp eq i64 [[VALUE_PHI5]], [[TMP2]]
; CHECK-NEXT: br i1 [[DOTNOT]], label [[L44]], label [[L27]], !llvm.loop [[LOOP16:![0-9]+]]
; CHECK: L44:
; CHECK-NEXT: ret ptr addrspace(10) null
;
top:
%2 = sext i32 %1 to i64
%3 = load atomic ptr addrspace(10) (ptr addrspace(10), i64)*, ptr addrspace(10) (ptr addrspace(10), i64)** bitcast (ptr @jlplt_ijl_alloc_array_1d_10294_got to ptr addrspace(10) (ptr addrspace(10), i64)**) unordered, align 8
%4 = tail call ptr addrspace(10) %3(ptr addrspace(10) null, i64 0)
%5 = load ptr addrspace(10), ptr %0, align 8, !tbaa !0
%6 = addrspacecast ptr addrspace(10) %4 to ptr addrspace(11)
%7 = load ptr addrspace(13), ptr addrspace(11) %6, align 8, !tbaa !5
%.elt = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(10) %5, i64 0, i32 0
%.unpack = load ptr addrspace(10), ptr addrspace(10) %.elt, align 8, !tbaa !8
%.elt1 = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(10) %5, i64 0, i32 1
%.unpack2 = load i64, ptr addrspace(10) %.elt1, align 8, !tbaa !8
br label %L26
L26: ; preds = %L26, %top
%value_phi5 = phi i64 [ 0, %top ], [ %8, %L26 ]
%.repack = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) %7, i64 %value_phi5, i32 0
store ptr addrspace(10) %.unpack, ptr addrspace(13) %.repack, align 8, !tbaa !10
%.repack4 = getelementptr inbounds { ptr addrspace(10), i64 }, ptr addrspace(13) %7, i64 %value_phi5, i32 1
store i64 %.unpack2, ptr addrspace(13) %.repack4, align 8, !tbaa !10
%8 = add i64 %value_phi5, 1
%.not = icmp eq i64 %value_phi5, %2
br i1 %.not, label %L44, label %L26
L44: ; preds = %L26
ret ptr addrspace(10) null
}
attributes #0 = { "target-cpu"="skylake-avx512" "target-features"="+xsaves,+xsavec,+prfchw,+lzcnt,+sahf,+pku,+avx512vl,+avx512bw,+avx512cd,+clwb,+clflushopt,+adx,+avx512dq,+avx512f,+bmi2,+avx2,+bmi,+fsgsbase,+f16c,+avx,+xsave,+aes,+popcnt,+movbe,+sse4.2,+sse4.1,+cx16,+fma,+ssse3,+pclmul,+sse3,-rdrnd,-rtm,-rdseed,-avx512ifma,-avx512pf,-sha,-avx512vbmi,-waitpkg,-avx512vbmi2,-shstk,-gfni,-vaes,-vpclmulqdq,-avx512vnni,-avx512bitalg,-avx512vpopcntdq,-rdpid,-cldemote,-movdiri,-movdir64b,-enqcmd,-avx512vp2intersect,-serialize,-tsxldtrk,-pconfig,-amx-bf16,-amx-tile,-amx-int8,-sse4a,-xop,-lwp,-fma4,-tbm,-mwaitx,-xsaveopt,-clzero,-wbnoinvd,-avx512bf16,-ptwrite,+sse2,+mmx,+fxsr,+64bit,+cx8" }
attributes #1 = { inaccessiblemem_or_argmemonly }
attributes #2 = { allocsize(1) }
!0 = !{!1, !1, i64 0}
!1 = !{!"jtbaa_value", !2, i64 0}
!2 = !{!"jtbaa_data", !3, i64 0}
!3 = !{!"jtbaa", !4, i64 0}
!4 = !{!"jtbaa"}
!5 = !{!6, !6, i64 0}
!6 = !{!"jtbaa_arrayptr", !7, i64 0}
!7 = !{!"jtbaa_array", !3, i64 0}
!8 = !{!9, !9, i64 0}
!9 = !{!"jtbaa_immut", !1, i64 0}
!10 = !{!11, !11, i64 0}
!11 = !{!"jtbaa_arraybuf", !2, i64 0}
Reference in New Issue View Git Blame Copy Permalink