152 lines
8.6 KiB
LLVM
152 lines
8.6 KiB
LLVM
|
; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 2
|
||
|
|
; RUN: opt -passes=loop-vectorize -force-vector-width=4 -force-vector-interleave=1 -lv-strided-pointer-ivs=true -S %s | FileCheck %s
|
||
|
|
|
||
|
|
define void @step_direction_unknown(i32 %arg, ptr %dst) {
|
||
|
|
; CHECK-LABEL: define void @step_direction_unknown
|
||
|
|
; CHECK-SAME: (i32 [[ARG:%.*]], ptr [[DST:%.*]]) {
|
||
|
|
; CHECK-NEXT: entry:
|
||
|
|
; CHECK-NEXT: [[ADD:%.*]] = add i32 [[ARG]], 1
|
||
|
|
; CHECK-NEXT: br i1 true, label [[SCALAR_PH:%.*]], label [[VECTOR_SCEVCHECK:%.*]]
|
||
|
|
; CHECK: vector.scevcheck:
|
||
|
|
; CHECK-NEXT: [[TMP0:%.*]] = sub i32 -1, [[ARG]]
|
||
|
|
; CHECK-NEXT: [[TMP1:%.*]] = icmp slt i32 [[ADD]], 0
|
||
|
|
; CHECK-NEXT: [[TMP2:%.*]] = select i1 [[TMP1]], i32 [[TMP0]], i32 [[ADD]]
|
||
|
|
; CHECK-NEXT: [[MUL1:%.*]] = call { i32, i1 } @llvm.umul.with.overflow.i32(i32 [[TMP2]], i32 -1)
|
||
|
|
; CHECK-NEXT: [[MUL_RESULT:%.*]] = extractvalue { i32, i1 } [[MUL1]], 0
|
||
|
|
; CHECK-NEXT: [[MUL_OVERFLOW:%.*]] = extractvalue { i32, i1 } [[MUL1]], 1
|
||
|
|
; CHECK-NEXT: [[TMP3:%.*]] = sub i32 0, [[MUL_RESULT]]
|
||
|
|
; CHECK-NEXT: [[TMP4:%.*]] = icmp ugt i32 [[TMP3]], 0
|
||
|
|
; CHECK-NEXT: [[TMP5:%.*]] = select i1 [[TMP1]], i1 [[TMP4]], i1 false
|
||
|
|
; CHECK-NEXT: [[TMP6:%.*]] = or i1 [[TMP5]], [[MUL_OVERFLOW]]
|
||
|
|
; CHECK-NEXT: [[TMP7:%.*]] = icmp ne i32 [[ADD]], 0
|
||
|
|
; CHECK-NEXT: [[TMP8:%.*]] = or i1 [[TMP6]], [[TMP7]]
|
||
|
|
; CHECK-NEXT: br i1 [[TMP8]], label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
|
||
|
|
; CHECK: vector.ph:
|
||
|
|
; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> poison, i32 [[ADD]], i64 0
|
||
|
|
; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT]], <4 x i32> 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 i32> [ <i32 0, i32 1, i32 2, i32 3>, [[VECTOR_PH]] ], [ [[VEC_IND_NEXT:%.*]], [[VECTOR_BODY]] ]
|
||
|
|
; CHECK-NEXT: [[TMP9:%.*]] = mul <4 x i32> [[BROADCAST_SPLAT]], [[VEC_IND]]
|
||
|
|
; CHECK-NEXT: [[TMP10:%.*]] = zext <4 x i32> [[TMP9]] to <4 x i64>
|
||
|
|
; CHECK-NEXT: [[TMP11:%.*]] = extractelement <4 x i64> [[TMP10]], i32 0
|
||
|
|
; CHECK-NEXT: [[TMP12:%.*]] = getelementptr double, ptr [[DST]], i64 [[TMP11]]
|
||
|
|
; CHECK-NEXT: [[TMP13:%.*]] = extractelement <4 x i64> [[TMP10]], i32 1
|
||
|
|
; CHECK-NEXT: [[TMP14:%.*]] = getelementptr double, ptr [[DST]], i64 [[TMP13]]
|
||
|
|
; CHECK-NEXT: [[TMP15:%.*]] = extractelement <4 x i64> [[TMP10]], i32 2
|
||
|
|
; CHECK-NEXT: [[TMP16:%.*]] = getelementptr double, ptr [[DST]], i64 [[TMP15]]
|
||
|
|
; CHECK-NEXT: [[TMP17:%.*]] = extractelement <4 x i64> [[TMP10]], i32 3
|
||
|
|
; CHECK-NEXT: [[TMP18:%.*]] = getelementptr double, ptr [[DST]], i64 [[TMP17]]
|
||
|
|
; CHECK-NEXT: store double 0.000000e+00, ptr [[TMP12]], align 8
|
||
|
|
; CHECK-NEXT: store double 0.000000e+00, ptr [[TMP14]], align 8
|
||
|
|
; CHECK-NEXT: store double 0.000000e+00, ptr [[TMP16]], align 8
|
||
|
|
; CHECK-NEXT: store double 0.000000e+00, ptr [[TMP18]], align 8
|
||
|
|
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
|
||
|
|
; CHECK-NEXT: [[VEC_IND_NEXT]] = add <4 x i32> [[VEC_IND]], <i32 4, i32 4, i32 4, i32 4>
|
||
|
|
; CHECK-NEXT: [[TMP19:%.*]] = icmp eq i64 [[INDEX_NEXT]], 0
|
||
|
|
; CHECK-NEXT: br i1 [[TMP19]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
|
||
|
|
; CHECK: middle.block:
|
||
|
|
; CHECK-NEXT: br i1 true, label [[EXIT:%.*]], label [[SCALAR_PH]]
|
||
|
|
; CHECK: scalar.ph:
|
||
|
|
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ 0, [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ], [ 0, [[VECTOR_SCEVCHECK]] ]
|
||
|
|
; CHECK-NEXT: br label [[LOOP:%.*]]
|
||
|
|
; CHECK: loop:
|
||
|
|
; CHECK-NEXT: [[PHI:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[ADD2:%.*]], [[LOOP]] ]
|
||
|
|
; CHECK-NEXT: [[TRUNC:%.*]] = trunc i64 [[PHI]] to i32
|
||
|
|
; CHECK-NEXT: [[MUL:%.*]] = mul i32 [[ADD]], [[TRUNC]]
|
||
|
|
; CHECK-NEXT: [[ZEXT:%.*]] = zext i32 [[MUL]] to i64
|
||
|
|
; CHECK-NEXT: [[GETELEMENTPTR:%.*]] = getelementptr double, ptr [[DST]], i64 [[ZEXT]]
|
||
|
|
; CHECK-NEXT: store double 0.000000e+00, ptr [[GETELEMENTPTR]], align 8
|
||
|
|
; CHECK-NEXT: [[ADD2]] = add i64 [[PHI]], 1
|
||
|
|
; CHECK-NEXT: [[ICMP:%.*]] = icmp eq i64 [[ADD2]], 0
|
||
|
|
; CHECK-NEXT: br i1 [[ICMP]], label [[EXIT]], label [[LOOP]], !llvm.loop [[LOOP3:![0-9]+]]
|
||
|
|
; CHECK: exit:
|
||
|
|
; CHECK-NEXT: ret void
|
||
|
|
;
|
||
|
|
entry:
|
||
|
|
%add = add i32 %arg, 1
|
||
|
|
br label %loop
|
||
|
|
|
||
|
|
loop:
|
||
|
|
%phi = phi i64 [ 0, %entry ], [ %add2, %loop ]
|
||
|
|
%trunc = trunc i64 %phi to i32
|
||
|
|
%mul = mul i32 %add, %trunc
|
||
|
|
%zext = zext i32 %mul to i64
|
||
|
|
%getelementptr = getelementptr double, ptr %dst, i64 %zext
|
||
|
|
store double 0.000000e+00, ptr %getelementptr, align 8
|
||
|
|
%add2 = add i64 %phi, 1
|
||
|
|
%icmp = icmp eq i64 %add2, 0
|
||
|
|
br i1 %icmp, label %exit, label %loop
|
||
|
|
|
||
|
|
exit:
|
||
|
|
ret void
|
||
|
|
}
|
||
|
|
|
||
|
|
define void @integer_induction_wraps_scev_predicate_known(i32 %x, ptr %call, ptr %start) {
|
||
|
|
; CHECK-LABEL: define void @integer_induction_wraps_scev_predicate_known
|
||
|
|
; CHECK-SAME: (i32 [[X:%.*]], ptr [[CALL:%.*]], ptr [[START:%.*]]) {
|
||
|
|
; CHECK-NEXT: entry:
|
||
|
|
; CHECK-NEXT: [[MUL:%.*]] = shl i32 [[X]], 1
|
||
|
|
; CHECK-NEXT: [[TMP0:%.*]] = sext i32 [[MUL]] to i64
|
||
|
|
; CHECK-NEXT: br i1 false, label [[SCALAR_PH:%.*]], label [[VECTOR_SCEVCHECK:%.*]]
|
||
|
|
; CHECK: vector.scevcheck:
|
||
|
|
; CHECK-NEXT: br i1 true, label [[SCALAR_PH]], label [[VECTOR_PH:%.*]]
|
||
|
|
; CHECK: vector.ph:
|
||
|
|
; CHECK-NEXT: [[TMP1:%.*]] = mul i64 4294967264, [[TMP0]]
|
||
|
|
; CHECK-NEXT: [[IND_END:%.*]] = getelementptr i8, ptr [[START]], i64 [[TMP1]]
|
||
|
|
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
|
||
|
|
; CHECK: vector.body:
|
||
|
|
; CHECK-NEXT: [[POINTER_PHI:%.*]] = phi ptr [ [[START]], [[VECTOR_PH]] ], [ [[PTR_IND:%.*]], [[VECTOR_BODY]] ]
|
||
|
|
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
|
||
|
|
; CHECK-NEXT: [[TMP2:%.*]] = mul i64 [[TMP0]], 4
|
||
|
|
; CHECK-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <4 x i64> poison, i64 [[TMP0]], i64 0
|
||
|
|
; CHECK-NEXT: [[DOTSPLAT:%.*]] = shufflevector <4 x i64> [[DOTSPLATINSERT]], <4 x i64> poison, <4 x i32> zeroinitializer
|
||
|
|
; CHECK-NEXT: [[VECTOR_GEP:%.*]] = mul <4 x i64> <i64 0, i64 1, i64 2, i64 3>, [[DOTSPLAT]]
|
||
|
|
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr i8, ptr [[POINTER_PHI]], <4 x i64> [[VECTOR_GEP]]
|
||
|
|
; CHECK-NEXT: [[DOTCAST:%.*]] = trunc i64 [[INDEX]] to i32
|
||
|
|
; CHECK-NEXT: [[OFFSET_IDX:%.*]] = add i32 30, [[DOTCAST]]
|
||
|
|
; CHECK-NEXT: [[TMP4:%.*]] = add i32 [[OFFSET_IDX]], 0
|
||
|
|
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr ptr, ptr [[CALL]], i32 [[TMP4]]
|
||
|
|
; CHECK-NEXT: [[TMP6:%.*]] = getelementptr ptr, ptr [[TMP5]], i32 0
|
||
|
|
; CHECK-NEXT: store <4 x ptr> [[TMP3]], ptr [[TMP6]], align 4
|
||
|
|
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
|
||
|
|
; CHECK-NEXT: [[PTR_IND]] = getelementptr i8, ptr [[POINTER_PHI]], i64 [[TMP2]]
|
||
|
|
; CHECK-NEXT: [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], 4294967264
|
||
|
|
; CHECK-NEXT: br i1 [[TMP7]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
|
||
|
|
; CHECK: middle.block:
|
||
|
|
; CHECK-NEXT: br i1 false, label [[FOR_END:%.*]], label [[SCALAR_PH]]
|
||
|
|
; CHECK: scalar.ph:
|
||
|
|
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i32 [ -2, [[MIDDLE_BLOCK]] ], [ 30, [[ENTRY:%.*]] ], [ 30, [[VECTOR_SCEVCHECK]] ]
|
||
|
|
; CHECK-NEXT: [[BC_RESUME_VAL1:%.*]] = phi ptr [ [[IND_END]], [[MIDDLE_BLOCK]] ], [ [[START]], [[ENTRY]] ], [ [[START]], [[VECTOR_SCEVCHECK]] ]
|
||
|
|
; CHECK-NEXT: br label [[FOR_COND:%.*]]
|
||
|
|
; CHECK: for.cond:
|
||
|
|
; CHECK-NEXT: [[IV:%.*]] = phi i32 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INC:%.*]], [[FOR_COND]] ]
|
||
|
|
; CHECK-NEXT: [[P_0:%.*]] = phi ptr [ [[BC_RESUME_VAL1]], [[SCALAR_PH]] ], [ [[ADD_PTR:%.*]], [[FOR_COND]] ]
|
||
|
|
; CHECK-NEXT: [[ADD_PTR]] = getelementptr i8, ptr [[P_0]], i32 [[MUL]]
|
||
|
|
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr ptr, ptr [[CALL]], i32 [[IV]]
|
||
|
|
; CHECK-NEXT: store ptr [[P_0]], ptr [[ARRAYIDX]], align 4
|
||
|
|
; CHECK-NEXT: [[INC]] = add i32 [[IV]], 1
|
||
|
|
; CHECK-NEXT: [[TOBOOL_NOT:%.*]] = icmp eq i32 [[IV]], 0
|
||
|
|
; CHECK-NEXT: br i1 [[TOBOOL_NOT]], label [[FOR_END]], label [[FOR_COND]], !llvm.loop [[LOOP5:![0-9]+]]
|
||
|
|
; CHECK: for.end:
|
||
|
|
; CHECK-NEXT: ret void
|
||
|
|
;
|
||
|
|
entry:
|
||
|
|
%mul = shl i32 %x, 1
|
||
|
|
br label %for.cond
|
||
|
|
|
||
|
|
for.cond: ; preds = %for.body, %entry
|
||
|
|
%iv = phi i32 [ 30, %entry ], [ %inc, %for.cond ]
|
||
|
|
%p.0 = phi ptr [ %start, %entry ], [ %add.ptr, %for.cond ]
|
||
|
|
%add.ptr = getelementptr i8, ptr %p.0, i32 %mul
|
||
|
|
%arrayidx = getelementptr ptr, ptr %call, i32 %iv
|
||
|
|
store ptr %p.0, ptr %arrayidx, align 4
|
||
|
|
%inc = add i32 %iv, 1
|
||
|
|
%tobool.not = icmp eq i32 %iv, 0
|
||
|
|
br i1 %tobool.not, label %for.end, label %for.cond
|
||
|
|
|
||
|
|
for.end: ; preds = %for.cond
|
||
|
|
ret void
|
||
|
|
}
|