From 3a06e9fa87f19af2d485a291fd8058ba3e2ff1f2 Mon Sep 17 00:00:00 2001 From: Sanjay Patel Date: Thu, 21 Jun 2018 16:54:32 +0000 Subject: [PATCH] [LoopVectorize] regenerate full checks; NFC llvm-svn: 335257 --- .../Transforms/LoopVectorize/runtime-check.ll | 68 +++++++++++++++++-- 1 file changed, 61 insertions(+), 7 deletions(-) diff --git a/test/Transforms/LoopVectorize/runtime-check.ll b/test/Transforms/LoopVectorize/runtime-check.ll index a569f9bc6ae..88489faa831 100644 --- a/test/Transforms/LoopVectorize/runtime-check.ll +++ b/test/Transforms/LoopVectorize/runtime-check.ll @@ -1,3 +1,4 @@ +; NOTE: Assertions have been autogenerated by utils/update_test_checks.py ; RUN: opt < %s -loop-vectorize -force-vector-interleave=1 -force-vector-width=4 -dce -instcombine -S | FileCheck %s target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64-S128" @@ -8,13 +9,66 @@ target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f3 ; a[i] = b[i] * 3; ; } -;CHECK-LABEL: define i32 @foo -;CHECK: for.body.preheader: -;CHECK: br i1 %min.iters.check, label %scalar.ph, label %vector.memcheck, !dbg [[BODY_LOC:![0-9]+]] -;CHECK: vector.memcheck: -;CHECK: br i1 %memcheck.conflict, label %scalar.ph, label %vector.ph, !dbg [[BODY_LOC]] -;CHECK: load <4 x float> define i32 @foo(float* nocapture %a, float* nocapture %b, i32 %n) nounwind uwtable ssp { +; CHECK-LABEL: @foo( +; CHECK-NEXT: entry: +; CHECK-NEXT: [[CMP6:%.*]] = icmp sgt i32 [[N:%.*]], 0, !dbg !4 +; CHECK-NEXT: br i1 [[CMP6]], label [[FOR_BODY_PREHEADER:%.*]], label [[FOR_END:%.*]], !dbg !4 +; CHECK: for.body.preheader: +; CHECK-NEXT: [[TMP0:%.*]] = add i32 [[N]], -1, !dbg !9 +; CHECK-NEXT: [[TMP1:%.*]] = zext i32 [[TMP0]] to i64, !dbg !9 +; CHECK-NEXT: [[TMP2:%.*]] = add nuw nsw i64 [[TMP1]], 1, !dbg !9 +; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[TMP2]], 4, !dbg !9 +; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_MEMCHECK:%.*]], !dbg !9 +; CHECK: vector.memcheck: +; CHECK-NEXT: [[TMP3:%.*]] = add i32 [[N]], -1, !dbg !9 +; CHECK-NEXT: [[TMP4:%.*]] = zext i32 [[TMP3]] to i64, !dbg !9 +; CHECK-NEXT: [[TMP5:%.*]] = add nuw nsw i64 [[TMP4]], 1, !dbg !9 +; CHECK-NEXT: [[SCEVGEP:%.*]] = getelementptr float, float* [[A:%.*]], i64 [[TMP5]], !dbg !9 +; CHECK-NEXT: [[SCEVGEP4:%.*]] = getelementptr float, float* [[B:%.*]], i64 [[TMP5]], !dbg !9 +; CHECK-NEXT: [[BOUND0:%.*]] = icmp ugt float* [[SCEVGEP4]], [[A]], !dbg !9 +; CHECK-NEXT: [[BOUND1:%.*]] = icmp ugt float* [[SCEVGEP]], [[B]], !dbg !9 +; CHECK-NEXT: [[MEMCHECK_CONFLICT:%.*]] = and i1 [[BOUND0]], [[BOUND1]], !dbg !9 +; CHECK-NEXT: br i1 [[MEMCHECK_CONFLICT]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]], !dbg !9 +; CHECK: vector.ph: +; CHECK-NEXT: [[TMP6:%.*]] = and i32 [[N]], 3, !dbg !9 +; CHECK-NEXT: [[N_MOD_VF:%.*]] = zext i32 [[TMP6]] to i64, !dbg !9 +; CHECK-NEXT: [[N_VEC:%.*]] = sub nsw i64 [[TMP2]], [[N_MOD_VF]], !dbg !9 +; CHECK-NEXT: br label [[VECTOR_BODY:%.*]], !dbg !9 +; CHECK: vector.body: +; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ], !dbg !9 +; CHECK-NEXT: [[TMP7:%.*]] = getelementptr inbounds float, float* [[B]], i64 [[INDEX]], !dbg !9 +; CHECK-NEXT: [[TMP8:%.*]] = bitcast float* [[TMP7]] to <4 x float>*, !dbg !9 +; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x float>, <4 x float>* [[TMP8]], align 4, !dbg !9, !alias.scope !10 +; CHECK-NEXT: [[TMP9:%.*]] = fmul <4 x float> [[WIDE_LOAD]], , !dbg !9 +; CHECK-NEXT: [[TMP10:%.*]] = getelementptr inbounds float, float* [[A]], i64 [[INDEX]], !dbg !9 +; CHECK-NEXT: [[TMP11:%.*]] = bitcast float* [[TMP10]] to <4 x float>*, !dbg !9 +; CHECK-NEXT: store <4 x float> [[TMP9]], <4 x float>* [[TMP11]], align 4, !dbg !9, !alias.scope !13, !noalias !10 +; CHECK-NEXT: [[INDEX_NEXT]] = add i64 [[INDEX]], 4, !dbg !9 +; CHECK-NEXT: [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]], !dbg !9 +; CHECK-NEXT: br i1 [[TMP12]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !dbg !9, !llvm.loop !15 +; CHECK: middle.block: +; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i32 [[TMP6]], 0 +; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_END_LOOPEXIT:%.*]], label [[SCALAR_PH]], !dbg !9 +; CHECK: scalar.ph: +; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[FOR_BODY_PREHEADER]] ], [ 0, [[VECTOR_MEMCHECK]] ] +; CHECK-NEXT: br label [[FOR_BODY:%.*]], !dbg !9 +; CHECK: for.body: +; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], !dbg !9 +; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[B]], i64 [[INDVARS_IV]], !dbg !9 +; CHECK-NEXT: [[TMP13:%.*]] = load float, float* [[ARRAYIDX]], align 4, !dbg !9 +; CHECK-NEXT: [[MUL:%.*]] = fmul float [[TMP13]], 3.000000e+00, !dbg !9 +; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds float, float* [[A]], i64 [[INDVARS_IV]], !dbg !9 +; CHECK-NEXT: store float [[MUL]], float* [[ARRAYIDX2]], align 4, !dbg !9 +; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add i64 [[INDVARS_IV]], 1, !dbg !9 +; CHECK-NEXT: [[LFTR_WIDEIV:%.*]] = trunc i64 [[INDVARS_IV_NEXT]] to i32, !dbg !9 +; CHECK-NEXT: [[EXITCOND:%.*]] = icmp eq i32 [[LFTR_WIDEIV]], [[N]], !dbg !9 +; CHECK-NEXT: br i1 [[EXITCOND]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !dbg !9, !llvm.loop !17 +; CHECK: for.end.loopexit: +; CHECK-NEXT: br label [[FOR_END]], !dbg !18 +; CHECK: for.end: +; CHECK-NEXT: ret i32 undef, !dbg !18 +; entry: %cmp6 = icmp sgt i32 %n, 0, !dbg !6 br i1 %cmp6, label %for.body, label %for.end, !dbg !6 @@ -63,7 +117,7 @@ loopexit: ret void } -; CHECK: [[BODY_LOC]] = !DILocation(line: 101, column: 1, scope: !{{.*}}) +; CHECK: !9 = !DILocation(line: 101, column: 1, scope: !{{.*}}) !llvm.module.flags = !{!0, !1} !llvm.dbg.cu = !{!9}