shylie/llvm-project
1
0
Fork 0
Code Issues Pull Requests Actions 6 Packages Projects Releases Wiki Activity
llvm-project/llvm/test/Transforms/LoopVectorize/RISCV/select-cmp-reduction.ll
Arthur Eubanks c6e01627ac Revert "Reapply "[LV] Improve AnyOf reduction codegen. (#78304)"" 
This reverts commit c6e38b928c56f562aea68a8e90f02dbdf0eada85.

Causes miscompiles, see comments on #78304.
2024-04-16 20:40:21 +00:00

851 lines
57 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
; RUN: opt -passes=loop-vectorize -force-vector-interleave=1 -force-vector-width=4 -S \
; RUN: < %s | FileCheck %s
; RUN: opt -passes=loop-vectorize -force-vector-interleave=1 -force-vector-width=4 \
; RUN: -scalable-vectorization=on -S < %s | FileCheck %s -check-prefix=SCALABLE
target triple = "riscv64"
define i32 @select_icmp(i32 %x, i32 %y, ptr nocapture readonly %c, i64 %n) #0 {
; CHECK-LABEL: define i32 @select_icmp(
; CHECK-SAME: i32 [[X:%.*]], i32 [[Y:%.*]], ptr nocapture readonly [[C:%.*]], i64 [[N:%.*]]) #[[ATTR0:[0-9]+]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], 4
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], 4
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> poison, i32 [[X]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT]], <4 x i32> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: [[BROADCAST_SPLATINSERT1:%.*]] = insertelement <4 x i32> poison, i32 [[Y]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT2:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT1]], <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_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP4:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds i32, ptr [[C]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP2]], align 4
; CHECK-NEXT: [[TMP3:%.*]] = icmp slt <4 x i32> [[WIDE_LOAD]], [[BROADCAST_SPLAT]]
; CHECK-NEXT: [[TMP4]] = select <4 x i1> [[TMP3]], <4 x i32> [[VEC_PHI]], <4 x i32> [[BROADCAST_SPLAT2]]
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP5:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP5]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <4 x i32> [[TMP4]], zeroinitializer
; CHECK-NEXT: [[TMP6:%.*]] = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> [[RDX_SELECT_CMP]])
; CHECK-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP6]], i32 [[Y]], i32 0
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[A:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[COND:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[C]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP7:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
; CHECK-NEXT: [[CMP1:%.*]] = icmp slt i32 [[TMP7]], [[X]]
; CHECK-NEXT: [[COND]] = select i1 [[CMP1]], i32 [[A]], i32 [[Y]]
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[EXITCOND_NOT:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], [[N]]
; CHECK-NEXT: br i1 [[EXITCOND_NOT]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK: for.end:
; CHECK-NEXT: [[COND_LCSSA:%.*]] = phi i32 [ [[COND]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: ret i32 [[COND_LCSSA]]
;
; SCALABLE-LABEL: define i32 @select_icmp(
; SCALABLE-SAME: i32 [[X:%.*]], i32 [[Y:%.*]], ptr nocapture readonly [[C:%.*]], i64 [[N:%.*]]) #[[ATTR0:[0-9]+]] {
; SCALABLE-NEXT: entry:
; SCALABLE-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP1:%.*]] = mul i64 [[TMP0]], 4
; SCALABLE-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], [[TMP1]]
; SCALABLE-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; SCALABLE: vector.ph:
; SCALABLE-NEXT: [[TMP2:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP3:%.*]] = mul i64 [[TMP2]], 4
; SCALABLE-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP3]]
; SCALABLE-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; SCALABLE-NEXT: [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP5:%.*]] = mul i64 [[TMP4]], 4
; SCALABLE-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[X]], i64 0
; SCALABLE-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[BROADCAST_SPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
; SCALABLE-NEXT: [[BROADCAST_SPLATINSERT1:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[Y]], i64 0
; SCALABLE-NEXT: [[BROADCAST_SPLAT2:%.*]] = shufflevector <vscale x 4 x i32> [[BROADCAST_SPLATINSERT1]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
; SCALABLE-NEXT: br label [[VECTOR_BODY:%.*]]
; SCALABLE: vector.body:
; SCALABLE-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP10:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[TMP6:%.*]] = add i64 [[INDEX]], 0
; SCALABLE-NEXT: [[TMP7:%.*]] = getelementptr inbounds i32, ptr [[C]], i64 [[TMP6]]
; SCALABLE-NEXT: [[TMP8:%.*]] = getelementptr inbounds i32, ptr [[TMP7]], i32 0
; SCALABLE-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 4 x i32>, ptr [[TMP8]], align 4
; SCALABLE-NEXT: [[TMP9:%.*]] = icmp slt <vscale x 4 x i32> [[WIDE_LOAD]], [[BROADCAST_SPLAT]]
; SCALABLE-NEXT: [[TMP10]] = select <vscale x 4 x i1> [[TMP9]], <vscale x 4 x i32> [[VEC_PHI]], <vscale x 4 x i32> [[BROADCAST_SPLAT2]]
; SCALABLE-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP5]]
; SCALABLE-NEXT: [[TMP11:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[TMP11]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; SCALABLE: middle.block:
; SCALABLE-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <vscale x 4 x i32> [[TMP10]], zeroinitializer
; SCALABLE-NEXT: [[TMP12:%.*]] = call i1 @llvm.vector.reduce.or.nxv4i1(<vscale x 4 x i1> [[RDX_SELECT_CMP]])
; SCALABLE-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP12]], i32 [[Y]], i32 0
; SCALABLE-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
; SCALABLE: scalar.ph:
; SCALABLE-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; SCALABLE-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: br label [[FOR_BODY:%.*]]
; SCALABLE: for.body:
; SCALABLE-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[A:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[COND:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[C]], i64 [[INDVARS_IV]]
; SCALABLE-NEXT: [[TMP13:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
; SCALABLE-NEXT: [[CMP1:%.*]] = icmp slt i32 [[TMP13]], [[X]]
; SCALABLE-NEXT: [[COND]] = select i1 [[CMP1]], i32 [[A]], i32 [[Y]]
; SCALABLE-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; SCALABLE-NEXT: [[EXITCOND_NOT:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], [[N]]
; SCALABLE-NEXT: br i1 [[EXITCOND_NOT]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; SCALABLE: for.end:
; SCALABLE-NEXT: [[COND_LCSSA:%.*]] = phi i32 [ [[COND]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: ret i32 [[COND_LCSSA]]
;
entry:
br label %for.body
for.body:
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%a = phi i32 [ 0, %entry], [ %cond, %for.body ]
%arrayidx = getelementptr inbounds i32, ptr %c, i64 %indvars.iv
%0 = load i32, ptr %arrayidx, align 4
%cmp1 = icmp slt i32 %0, %x
%cond = select i1 %cmp1, i32 %a, i32 %y
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond.not = icmp eq i64 %indvars.iv.next, %n
br i1 %exitcond.not, label %for.end, label %for.body
for.end:
ret i32 %cond
}
define i32 @select_fcmp(float %x, i32 %y, ptr nocapture readonly %c, i64 %n) #0 {
; CHECK-LABEL: define i32 @select_fcmp(
; CHECK-SAME: float [[X:%.*]], i32 [[Y:%.*]], ptr nocapture readonly [[C:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], 4
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], 4
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x float> poison, float [[X]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x float> [[BROADCAST_SPLATINSERT]], <4 x float> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: [[BROADCAST_SPLATINSERT1:%.*]] = insertelement <4 x i32> poison, i32 [[Y]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT2:%.*]] = shufflevector <4 x i32> [[BROADCAST_SPLATINSERT1]], <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_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP4:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds float, ptr [[C]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds float, ptr [[TMP1]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP2]], align 4
; CHECK-NEXT: [[TMP3:%.*]] = fcmp fast olt <4 x float> [[WIDE_LOAD]], [[BROADCAST_SPLAT]]
; CHECK-NEXT: [[TMP4]] = select <4 x i1> [[TMP3]], <4 x i32> [[VEC_PHI]], <4 x i32> [[BROADCAST_SPLAT2]]
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP5:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP5]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <4 x i32> [[TMP4]], zeroinitializer
; CHECK-NEXT: [[TMP6:%.*]] = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> [[RDX_SELECT_CMP]])
; CHECK-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP6]], i32 [[Y]], i32 0
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[A:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[COND:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[C]], i64 [[INDVARS_IV]]
; CHECK-NEXT: [[TMP7:%.*]] = load float, ptr [[ARRAYIDX]], align 4
; CHECK-NEXT: [[CMP1:%.*]] = fcmp fast olt float [[TMP7]], [[X]]
; CHECK-NEXT: [[COND]] = select i1 [[CMP1]], i32 [[A]], i32 [[Y]]
; CHECK-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; CHECK-NEXT: [[EXITCOND_NOT:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], [[N]]
; CHECK-NEXT: br i1 [[EXITCOND_NOT]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK: for.end:
; CHECK-NEXT: [[COND_LCSSA:%.*]] = phi i32 [ [[COND]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: ret i32 [[COND_LCSSA]]
;
; SCALABLE-LABEL: define i32 @select_fcmp(
; SCALABLE-SAME: float [[X:%.*]], i32 [[Y:%.*]], ptr nocapture readonly [[C:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; SCALABLE-NEXT: entry:
; SCALABLE-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP1:%.*]] = mul i64 [[TMP0]], 4
; SCALABLE-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], [[TMP1]]
; SCALABLE-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; SCALABLE: vector.ph:
; SCALABLE-NEXT: [[TMP2:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP3:%.*]] = mul i64 [[TMP2]], 4
; SCALABLE-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP3]]
; SCALABLE-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; SCALABLE-NEXT: [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP5:%.*]] = mul i64 [[TMP4]], 4
; SCALABLE-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 4 x float> poison, float [[X]], i64 0
; SCALABLE-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 4 x float> [[BROADCAST_SPLATINSERT]], <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer
; SCALABLE-NEXT: [[BROADCAST_SPLATINSERT1:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[Y]], i64 0
; SCALABLE-NEXT: [[BROADCAST_SPLAT2:%.*]] = shufflevector <vscale x 4 x i32> [[BROADCAST_SPLATINSERT1]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
; SCALABLE-NEXT: br label [[VECTOR_BODY:%.*]]
; SCALABLE: vector.body:
; SCALABLE-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[TMP10:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[TMP6:%.*]] = add i64 [[INDEX]], 0
; SCALABLE-NEXT: [[TMP7:%.*]] = getelementptr inbounds float, ptr [[C]], i64 [[TMP6]]
; SCALABLE-NEXT: [[TMP8:%.*]] = getelementptr inbounds float, ptr [[TMP7]], i32 0
; SCALABLE-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 4 x float>, ptr [[TMP8]], align 4
; SCALABLE-NEXT: [[TMP9:%.*]] = fcmp fast olt <vscale x 4 x float> [[WIDE_LOAD]], [[BROADCAST_SPLAT]]
; SCALABLE-NEXT: [[TMP10]] = select <vscale x 4 x i1> [[TMP9]], <vscale x 4 x i32> [[VEC_PHI]], <vscale x 4 x i32> [[BROADCAST_SPLAT2]]
; SCALABLE-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP5]]
; SCALABLE-NEXT: [[TMP11:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[TMP11]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; SCALABLE: middle.block:
; SCALABLE-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <vscale x 4 x i32> [[TMP10]], zeroinitializer
; SCALABLE-NEXT: [[TMP12:%.*]] = call i1 @llvm.vector.reduce.or.nxv4i1(<vscale x 4 x i1> [[RDX_SELECT_CMP]])
; SCALABLE-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP12]], i32 [[Y]], i32 0
; SCALABLE-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[CMP_N]], label [[FOR_END:%.*]], label [[SCALAR_PH]]
; SCALABLE: scalar.ph:
; SCALABLE-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; SCALABLE-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: br label [[FOR_BODY:%.*]]
; SCALABLE: for.body:
; SCALABLE-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[INDVARS_IV_NEXT:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[A:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[COND:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, ptr [[C]], i64 [[INDVARS_IV]]
; SCALABLE-NEXT: [[TMP13:%.*]] = load float, ptr [[ARRAYIDX]], align 4
; SCALABLE-NEXT: [[CMP1:%.*]] = fcmp fast olt float [[TMP13]], [[X]]
; SCALABLE-NEXT: [[COND]] = select i1 [[CMP1]], i32 [[A]], i32 [[Y]]
; SCALABLE-NEXT: [[INDVARS_IV_NEXT]] = add nuw nsw i64 [[INDVARS_IV]], 1
; SCALABLE-NEXT: [[EXITCOND_NOT:%.*]] = icmp eq i64 [[INDVARS_IV_NEXT]], [[N]]
; SCALABLE-NEXT: br i1 [[EXITCOND_NOT]], label [[FOR_END]], label [[FOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
; SCALABLE: for.end:
; SCALABLE-NEXT: [[COND_LCSSA:%.*]] = phi i32 [ [[COND]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: ret i32 [[COND_LCSSA]]
;
entry:
br label %for.body
for.body:
%indvars.iv = phi i64 [ 0, %entry ], [ %indvars.iv.next, %for.body ]
%a = phi i32 [ 0, %entry], [ %cond, %for.body ]
%arrayidx = getelementptr inbounds float, ptr %c, i64 %indvars.iv
%0 = load float, ptr %arrayidx, align 4
%cmp1 = fcmp fast olt float %0, %x
%cond = select i1 %cmp1, i32 %a, i32 %y
%indvars.iv.next = add nuw nsw i64 %indvars.iv, 1
%exitcond.not = icmp eq i64 %indvars.iv.next, %n
br i1 %exitcond.not, label %for.end, label %for.body
for.end:
ret i32 %cond
}
define i32 @select_const_i32_from_icmp(ptr nocapture readonly %v, i64 %n) #0 {
; CHECK-LABEL: define i32 @select_const_i32_from_icmp(
; CHECK-SAME: ptr nocapture readonly [[V:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], 4
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], 4
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[VEC_PHI:%.*]] = phi <4 x i32> [ <i32 3, i32 3, i32 3, i32 3>, [[VECTOR_PH]] ], [ [[TMP5:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP2]], align 4
; CHECK-NEXT: [[TMP4:%.*]] = icmp eq <4 x i32> [[WIDE_LOAD]], <i32 3, i32 3, i32 3, i32 3>
; CHECK-NEXT: [[TMP5]] = select <4 x i1> [[TMP4]], <4 x i32> [[VEC_PHI]], <4 x i32> <i32 7, i32 7, i32 7, i32 7>
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <4 x i32> [[TMP5]], <i32 3, i32 3, i32 3, i32 3>
; CHECK-NEXT: [[TMP7:%.*]] = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> [[RDX_SELECT_CMP]])
; CHECK-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP7]], i32 7, i32 3
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 3, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[TMP15:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[TMP13:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[TMP8:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TMP12:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP15]]
; CHECK-NEXT: [[TMP10:%.*]] = load i32, ptr [[TMP9]], align 4
; CHECK-NEXT: [[TMP11:%.*]] = icmp eq i32 [[TMP10]], 3
; CHECK-NEXT: [[TMP12]] = select i1 [[TMP11]], i32 [[TMP8]], i32 7
; CHECK-NEXT: [[TMP13]] = add nuw nsw i64 [[TMP15]], 1
; CHECK-NEXT: [[TMP14:%.*]] = icmp eq i64 [[TMP13]], [[N]]
; CHECK-NEXT: br i1 [[TMP14]], label [[EXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
; CHECK: exit:
; CHECK-NEXT: [[DOTLCSSA:%.*]] = phi i32 [ [[TMP12]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: ret i32 [[DOTLCSSA]]
;
; SCALABLE-LABEL: define i32 @select_const_i32_from_icmp(
; SCALABLE-SAME: ptr nocapture readonly [[V:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; SCALABLE-NEXT: entry:
; SCALABLE-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP1:%.*]] = mul i64 [[TMP0]], 4
; SCALABLE-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], [[TMP1]]
; SCALABLE-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; SCALABLE: vector.ph:
; SCALABLE-NEXT: [[TMP2:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP3:%.*]] = mul i64 [[TMP2]], 4
; SCALABLE-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP3]]
; SCALABLE-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; SCALABLE-NEXT: [[TMP10:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP11:%.*]] = mul i64 [[TMP10]], 4
; SCALABLE-NEXT: br label [[VECTOR_BODY:%.*]]
; SCALABLE: vector.body:
; SCALABLE-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 3, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer), [[VECTOR_PH]] ], [ [[TMP9:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[TMP4:%.*]] = add i64 [[INDEX]], 0
; SCALABLE-NEXT: [[TMP5:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP4]]
; SCALABLE-NEXT: [[TMP6:%.*]] = getelementptr inbounds i32, ptr [[TMP5]], i32 0
; SCALABLE-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 4 x i32>, ptr [[TMP6]], align 4
; SCALABLE-NEXT: [[TMP8:%.*]] = icmp eq <vscale x 4 x i32> [[WIDE_LOAD]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 3, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
; SCALABLE-NEXT: [[TMP9]] = select <vscale x 4 x i1> [[TMP8]], <vscale x 4 x i32> [[VEC_PHI]], <vscale x 4 x i32> shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 7, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
; SCALABLE-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP11]]
; SCALABLE-NEXT: [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[TMP12]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
; SCALABLE: middle.block:
; SCALABLE-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <vscale x 4 x i32> [[TMP9]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 3, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
; SCALABLE-NEXT: [[TMP13:%.*]] = call i1 @llvm.vector.reduce.or.nxv4i1(<vscale x 4 x i1> [[RDX_SELECT_CMP]])
; SCALABLE-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP13]], i32 7, i32 3
; SCALABLE-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
; SCALABLE: scalar.ph:
; SCALABLE-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; SCALABLE-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 3, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: br label [[FOR_BODY:%.*]]
; SCALABLE: for.body:
; SCALABLE-NEXT: [[TMP21:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[TMP19:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[TMP14:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TMP18:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[TMP15:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP21]]
; SCALABLE-NEXT: [[TMP16:%.*]] = load i32, ptr [[TMP15]], align 4
; SCALABLE-NEXT: [[TMP17:%.*]] = icmp eq i32 [[TMP16]], 3
; SCALABLE-NEXT: [[TMP18]] = select i1 [[TMP17]], i32 [[TMP14]], i32 7
; SCALABLE-NEXT: [[TMP19]] = add nuw nsw i64 [[TMP21]], 1
; SCALABLE-NEXT: [[TMP20:%.*]] = icmp eq i64 [[TMP19]], [[N]]
; SCALABLE-NEXT: br i1 [[TMP20]], label [[EXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
; SCALABLE: exit:
; SCALABLE-NEXT: [[DOTLCSSA:%.*]] = phi i32 [ [[TMP18]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: ret i32 [[DOTLCSSA]]
;
entry:
br label %for.body
for.body: ; preds = %entry, %for.body
%0 = phi i64 [ 0, %entry ], [ %6, %for.body ]
%1 = phi i32 [ 3, %entry ], [ %5, %for.body ]
%2 = getelementptr inbounds i32, ptr %v, i64 %0
%3 = load i32, ptr %2, align 4
%4 = icmp eq i32 %3, 3
%5 = select i1 %4, i32 %1, i32 7
%6 = add nuw nsw i64 %0, 1
%7 = icmp eq i64 %6, %n
br i1 %7, label %exit, label %for.body
exit: ; preds = %for.body
ret i32 %5
}
define i32 @select_i32_from_icmp(ptr nocapture readonly %v, i32 %a, i32 %b, i64 %n) #0 {
; CHECK-LABEL: define i32 @select_i32_from_icmp(
; CHECK-SAME: ptr nocapture readonly [[V:%.*]], i32 [[A:%.*]], i32 [[B:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], 4
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], 4
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-NEXT: [[MINMAX_IDENT_SPLATINSERT:%.*]] = insertelement <4 x i32> poison, i32 [[A]], i64 0
; CHECK-NEXT: [[MINMAX_IDENT_SPLAT:%.*]] = shufflevector <4 x i32> [[MINMAX_IDENT_SPLATINSERT]], <4 x i32> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x i32> poison, i32 [[B]], 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_PHI:%.*]] = phi <4 x i32> [ [[MINMAX_IDENT_SPLAT]], [[VECTOR_PH]] ], [ [[TMP5:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP2]], align 4
; CHECK-NEXT: [[TMP4:%.*]] = icmp eq <4 x i32> [[WIDE_LOAD]], <i32 3, i32 3, i32 3, i32 3>
; CHECK-NEXT: [[TMP5]] = select <4 x i1> [[TMP4]], <4 x i32> [[VEC_PHI]], <4 x i32> [[BROADCAST_SPLAT]]
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <4 x i32> poison, i32 [[A]], i64 0
; CHECK-NEXT: [[DOTSPLAT:%.*]] = shufflevector <4 x i32> [[DOTSPLATINSERT]], <4 x i32> poison, <4 x i32> zeroinitializer
; CHECK-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <4 x i32> [[TMP5]], [[DOTSPLAT]]
; CHECK-NEXT: [[TMP7:%.*]] = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> [[RDX_SELECT_CMP]])
; CHECK-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP7]], i32 [[B]], i32 [[A]]
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ [[A]], [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[TMP15:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[TMP13:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[TMP8:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TMP12:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP15]]
; CHECK-NEXT: [[TMP10:%.*]] = load i32, ptr [[TMP9]], align 4
; CHECK-NEXT: [[TMP11:%.*]] = icmp eq i32 [[TMP10]], 3
; CHECK-NEXT: [[TMP12]] = select i1 [[TMP11]], i32 [[TMP8]], i32 [[B]]
; CHECK-NEXT: [[TMP13]] = add nuw nsw i64 [[TMP15]], 1
; CHECK-NEXT: [[TMP14:%.*]] = icmp eq i64 [[TMP13]], [[N]]
; CHECK-NEXT: br i1 [[TMP14]], label [[EXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP9:![0-9]+]]
; CHECK: exit:
; CHECK-NEXT: [[DOTLCSSA:%.*]] = phi i32 [ [[TMP12]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: ret i32 [[DOTLCSSA]]
;
; SCALABLE-LABEL: define i32 @select_i32_from_icmp(
; SCALABLE-SAME: ptr nocapture readonly [[V:%.*]], i32 [[A:%.*]], i32 [[B:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; SCALABLE-NEXT: entry:
; SCALABLE-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP1:%.*]] = mul i64 [[TMP0]], 4
; SCALABLE-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], [[TMP1]]
; SCALABLE-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; SCALABLE: vector.ph:
; SCALABLE-NEXT: [[TMP2:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP3:%.*]] = mul i64 [[TMP2]], 4
; SCALABLE-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP3]]
; SCALABLE-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; SCALABLE-NEXT: [[TMP10:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP11:%.*]] = mul i64 [[TMP10]], 4
; SCALABLE-NEXT: [[MINMAX_IDENT_SPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[A]], i64 0
; SCALABLE-NEXT: [[MINMAX_IDENT_SPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[MINMAX_IDENT_SPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
; SCALABLE-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[B]], i64 0
; SCALABLE-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[BROADCAST_SPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
; SCALABLE-NEXT: br label [[VECTOR_BODY:%.*]]
; SCALABLE: vector.body:
; SCALABLE-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ [[MINMAX_IDENT_SPLAT]], [[VECTOR_PH]] ], [ [[TMP9:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[TMP4:%.*]] = add i64 [[INDEX]], 0
; SCALABLE-NEXT: [[TMP5:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP4]]
; SCALABLE-NEXT: [[TMP6:%.*]] = getelementptr inbounds i32, ptr [[TMP5]], i32 0
; SCALABLE-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 4 x i32>, ptr [[TMP6]], align 4
; SCALABLE-NEXT: [[TMP8:%.*]] = icmp eq <vscale x 4 x i32> [[WIDE_LOAD]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 3, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
; SCALABLE-NEXT: [[TMP9]] = select <vscale x 4 x i1> [[TMP8]], <vscale x 4 x i32> [[VEC_PHI]], <vscale x 4 x i32> [[BROADCAST_SPLAT]]
; SCALABLE-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP11]]
; SCALABLE-NEXT: [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[TMP12]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP8:![0-9]+]]
; SCALABLE: middle.block:
; SCALABLE-NEXT: [[DOTSPLATINSERT:%.*]] = insertelement <vscale x 4 x i32> poison, i32 [[A]], i64 0
; SCALABLE-NEXT: [[DOTSPLAT:%.*]] = shufflevector <vscale x 4 x i32> [[DOTSPLATINSERT]], <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer
; SCALABLE-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <vscale x 4 x i32> [[TMP9]], [[DOTSPLAT]]
; SCALABLE-NEXT: [[TMP13:%.*]] = call i1 @llvm.vector.reduce.or.nxv4i1(<vscale x 4 x i1> [[RDX_SELECT_CMP]])
; SCALABLE-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP13]], i32 [[B]], i32 [[A]]
; SCALABLE-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
; SCALABLE: scalar.ph:
; SCALABLE-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; SCALABLE-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ [[A]], [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: br label [[FOR_BODY:%.*]]
; SCALABLE: for.body:
; SCALABLE-NEXT: [[TMP21:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[TMP19:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[TMP14:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TMP18:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[TMP15:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP21]]
; SCALABLE-NEXT: [[TMP16:%.*]] = load i32, ptr [[TMP15]], align 4
; SCALABLE-NEXT: [[TMP17:%.*]] = icmp eq i32 [[TMP16]], 3
; SCALABLE-NEXT: [[TMP18]] = select i1 [[TMP17]], i32 [[TMP14]], i32 [[B]]
; SCALABLE-NEXT: [[TMP19]] = add nuw nsw i64 [[TMP21]], 1
; SCALABLE-NEXT: [[TMP20:%.*]] = icmp eq i64 [[TMP19]], [[N]]
; SCALABLE-NEXT: br i1 [[TMP20]], label [[EXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP9:![0-9]+]]
; SCALABLE: exit:
; SCALABLE-NEXT: [[DOTLCSSA:%.*]] = phi i32 [ [[TMP18]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: ret i32 [[DOTLCSSA]]
;
entry:
br label %for.body
for.body: ; preds = %entry, %for.body
%0 = phi i64 [ 0, %entry ], [ %6, %for.body ]
%1 = phi i32 [ %a, %entry ], [ %5, %for.body ]
%2 = getelementptr inbounds i32, ptr %v, i64 %0
%3 = load i32, ptr %2, align 4
%4 = icmp eq i32 %3, 3
%5 = select i1 %4, i32 %1, i32 %b
%6 = add nuw nsw i64 %0, 1
%7 = icmp eq i64 %6, %n
br i1 %7, label %exit, label %for.body
exit: ; preds = %for.body
ret i32 %5
}
define i32 @select_const_i32_from_fcmp(ptr nocapture readonly %v, i64 %n) #0 {
; CHECK-LABEL: define i32 @select_const_i32_from_fcmp(
; CHECK-SAME: ptr nocapture readonly [[V:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], 4
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], 4
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[VEC_PHI:%.*]] = phi <4 x i32> [ <i32 2, i32 2, i32 2, i32 2>, [[VECTOR_PH]] ], [ [[TMP5:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds float, ptr [[V]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds float, ptr [[TMP1]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x float>, ptr [[TMP2]], align 4
; CHECK-NEXT: [[TMP4:%.*]] = fcmp fast ueq <4 x float> [[WIDE_LOAD]], <float 3.000000e+00, float 3.000000e+00, float 3.000000e+00, float 3.000000e+00>
; CHECK-NEXT: [[TMP5]] = select <4 x i1> [[TMP4]], <4 x i32> [[VEC_PHI]], <4 x i32> <i32 1, i32 1, i32 1, i32 1>
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP6:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP6]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <4 x i32> [[TMP5]], <i32 2, i32 2, i32 2, i32 2>
; CHECK-NEXT: [[TMP7:%.*]] = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> [[RDX_SELECT_CMP]])
; CHECK-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP7]], i32 1, i32 2
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 2, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[TMP15:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[TMP13:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[TMP8:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TMP12:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[TMP9:%.*]] = getelementptr inbounds float, ptr [[V]], i64 [[TMP15]]
; CHECK-NEXT: [[TMP10:%.*]] = load float, ptr [[TMP9]], align 4
; CHECK-NEXT: [[TMP11:%.*]] = fcmp fast ueq float [[TMP10]], 3.000000e+00
; CHECK-NEXT: [[TMP12]] = select i1 [[TMP11]], i32 [[TMP8]], i32 1
; CHECK-NEXT: [[TMP13]] = add nuw nsw i64 [[TMP15]], 1
; CHECK-NEXT: [[TMP14:%.*]] = icmp eq i64 [[TMP13]], [[N]]
; CHECK-NEXT: br i1 [[TMP14]], label [[EXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP11:![0-9]+]]
; CHECK: exit:
; CHECK-NEXT: [[DOTLCSSA:%.*]] = phi i32 [ [[TMP12]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: ret i32 [[DOTLCSSA]]
;
; SCALABLE-LABEL: define i32 @select_const_i32_from_fcmp(
; SCALABLE-SAME: ptr nocapture readonly [[V:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; SCALABLE-NEXT: entry:
; SCALABLE-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP1:%.*]] = mul i64 [[TMP0]], 4
; SCALABLE-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], [[TMP1]]
; SCALABLE-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; SCALABLE: vector.ph:
; SCALABLE-NEXT: [[TMP2:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP3:%.*]] = mul i64 [[TMP2]], 4
; SCALABLE-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP3]]
; SCALABLE-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; SCALABLE-NEXT: [[TMP10:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP11:%.*]] = mul i64 [[TMP10]], 4
; SCALABLE-NEXT: br label [[VECTOR_BODY:%.*]]
; SCALABLE: vector.body:
; SCALABLE-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 2, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer), [[VECTOR_PH]] ], [ [[TMP9:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[TMP4:%.*]] = add i64 [[INDEX]], 0
; SCALABLE-NEXT: [[TMP5:%.*]] = getelementptr inbounds float, ptr [[V]], i64 [[TMP4]]
; SCALABLE-NEXT: [[TMP6:%.*]] = getelementptr inbounds float, ptr [[TMP5]], i32 0
; SCALABLE-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 4 x float>, ptr [[TMP6]], align 4
; SCALABLE-NEXT: [[TMP8:%.*]] = fcmp fast ueq <vscale x 4 x float> [[WIDE_LOAD]], shufflevector (<vscale x 4 x float> insertelement (<vscale x 4 x float> poison, float 3.000000e+00, i64 0), <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer)
; SCALABLE-NEXT: [[TMP9]] = select <vscale x 4 x i1> [[TMP8]], <vscale x 4 x i32> [[VEC_PHI]], <vscale x 4 x i32> shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 1, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
; SCALABLE-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP11]]
; SCALABLE-NEXT: [[TMP12:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[TMP12]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP10:![0-9]+]]
; SCALABLE: middle.block:
; SCALABLE-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <vscale x 4 x i32> [[TMP9]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 2, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
; SCALABLE-NEXT: [[TMP13:%.*]] = call i1 @llvm.vector.reduce.or.nxv4i1(<vscale x 4 x i1> [[RDX_SELECT_CMP]])
; SCALABLE-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP13]], i32 1, i32 2
; SCALABLE-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[CMP_N]], label [[EXIT:%.*]], label [[SCALAR_PH]]
; SCALABLE: scalar.ph:
; SCALABLE-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; SCALABLE-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 2, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: br label [[FOR_BODY:%.*]]
; SCALABLE: for.body:
; SCALABLE-NEXT: [[TMP21:%.*]] = phi i64 [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ], [ [[TMP19:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[TMP14:%.*]] = phi i32 [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ], [ [[TMP18:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[TMP15:%.*]] = getelementptr inbounds float, ptr [[V]], i64 [[TMP21]]
; SCALABLE-NEXT: [[TMP16:%.*]] = load float, ptr [[TMP15]], align 4
; SCALABLE-NEXT: [[TMP17:%.*]] = fcmp fast ueq float [[TMP16]], 3.000000e+00
; SCALABLE-NEXT: [[TMP18]] = select i1 [[TMP17]], i32 [[TMP14]], i32 1
; SCALABLE-NEXT: [[TMP19]] = add nuw nsw i64 [[TMP21]], 1
; SCALABLE-NEXT: [[TMP20:%.*]] = icmp eq i64 [[TMP19]], [[N]]
; SCALABLE-NEXT: br i1 [[TMP20]], label [[EXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP11:![0-9]+]]
; SCALABLE: exit:
; SCALABLE-NEXT: [[DOTLCSSA:%.*]] = phi i32 [ [[TMP18]], [[FOR_BODY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: ret i32 [[DOTLCSSA]]
;
entry:
br label %for.body
for.body: ; preds = %entry, %for.body
%0 = phi i64 [ 0, %entry ], [ %6, %for.body ]
%1 = phi i32 [ 2, %entry ], [ %5, %for.body ]
%2 = getelementptr inbounds float, ptr %v, i64 %0
%3 = load float, ptr %2, align 4
%4 = fcmp fast ueq float %3, 3.0
%5 = select i1 %4, i32 %1, i32 1
%6 = add nuw nsw i64 %0, 1
%7 = icmp eq i64 %6, %n
br i1 %7, label %exit, label %for.body
exit: ; preds = %for.body
ret i32 %5
}
define float @select_const_f32_from_icmp(ptr nocapture readonly %v, i64 %n) #0 {
; CHECK-LABEL: define float @select_const_f32_from_icmp(
; CHECK-SAME: ptr nocapture readonly [[V:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[TMP0:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[TMP6:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[TMP1:%.*]] = phi fast float [ 3.000000e+00, [[ENTRY]] ], [ [[TMP5:%.*]], [[FOR_BODY]] ]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
; CHECK-NEXT: [[TMP4:%.*]] = icmp eq i32 [[TMP3]], 3
; CHECK-NEXT: [[TMP5]] = select fast i1 [[TMP4]], float [[TMP1]], float 7.000000e+00
; CHECK-NEXT: [[TMP6]] = add nuw nsw i64 [[TMP0]], 1
; CHECK-NEXT: [[TMP7:%.*]] = icmp eq i64 [[TMP6]], [[N]]
; CHECK-NEXT: br i1 [[TMP7]], label [[EXIT:%.*]], label [[FOR_BODY]]
; CHECK: exit:
; CHECK-NEXT: [[DOTLCSSA:%.*]] = phi float [ [[TMP5]], [[FOR_BODY]] ]
; CHECK-NEXT: ret float [[DOTLCSSA]]
;
; SCALABLE-LABEL: define float @select_const_f32_from_icmp(
; SCALABLE-SAME: ptr nocapture readonly [[V:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; SCALABLE-NEXT: entry:
; SCALABLE-NEXT: br label [[FOR_BODY:%.*]]
; SCALABLE: for.body:
; SCALABLE-NEXT: [[TMP0:%.*]] = phi i64 [ 0, [[ENTRY:%.*]] ], [ [[TMP6:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[TMP1:%.*]] = phi fast float [ 3.000000e+00, [[ENTRY]] ], [ [[TMP5:%.*]], [[FOR_BODY]] ]
; SCALABLE-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[V]], i64 [[TMP0]]
; SCALABLE-NEXT: [[TMP3:%.*]] = load i32, ptr [[TMP2]], align 4
; SCALABLE-NEXT: [[TMP4:%.*]] = icmp eq i32 [[TMP3]], 3
; SCALABLE-NEXT: [[TMP5]] = select fast i1 [[TMP4]], float [[TMP1]], float 7.000000e+00
; SCALABLE-NEXT: [[TMP6]] = add nuw nsw i64 [[TMP0]], 1
; SCALABLE-NEXT: [[TMP7:%.*]] = icmp eq i64 [[TMP6]], [[N]]
; SCALABLE-NEXT: br i1 [[TMP7]], label [[EXIT:%.*]], label [[FOR_BODY]]
; SCALABLE: exit:
; SCALABLE-NEXT: [[DOTLCSSA:%.*]] = phi float [ [[TMP5]], [[FOR_BODY]] ]
; SCALABLE-NEXT: ret float [[DOTLCSSA]]
;
entry:
br label %for.body
for.body: ; preds = %entry, %for.body
%0 = phi i64 [ 0, %entry ], [ %6, %for.body ]
%1 = phi fast float [ 3.0, %entry ], [ %5, %for.body ]
%2 = getelementptr inbounds i32, ptr %v, i64 %0
%3 = load i32, ptr %2, align 4
%4 = icmp eq i32 %3, 3
%5 = select fast i1 %4, float %1, float 7.0
%6 = add nuw nsw i64 %0, 1
%7 = icmp eq i64 %6, %n
br i1 %7, label %exit, label %for.body
exit: ; preds = %for.body
ret float %5
}
define i32 @pred_select_const_i32_from_icmp(ptr noalias nocapture readonly %src1, ptr noalias nocapture readonly %src2, i64 %n) #0 {
; CHECK-LABEL: define i32 @pred_select_const_i32_from_icmp(
; CHECK-SAME: ptr noalias nocapture readonly [[SRC1:%.*]], ptr noalias nocapture readonly [[SRC2:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; CHECK-NEXT: entry:
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], 4
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], 4
; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; CHECK-NEXT: br label [[VECTOR_BODY:%.*]]
; CHECK: vector.body:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[VEC_PHI:%.*]] = phi <4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[PREDPHI:%.*]], [[VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP0:%.*]] = add i64 [[INDEX]], 0
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr inbounds i32, ptr [[SRC1]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr inbounds i32, ptr [[TMP1]], i32 0
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <4 x i32>, ptr [[TMP2]], align 4
; CHECK-NEXT: [[TMP3:%.*]] = icmp sgt <4 x i32> [[WIDE_LOAD]], <i32 35, i32 35, i32 35, i32 35>
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr i32, ptr [[SRC2]], i64 [[TMP0]]
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr i32, ptr [[TMP4]], i32 0
; CHECK-NEXT: [[WIDE_MASKED_LOAD:%.*]] = call <4 x i32> @llvm.masked.load.v4i32.p0(ptr [[TMP5]], i32 4, <4 x i1> [[TMP3]], <4 x i32> poison)
; CHECK-NEXT: [[TMP6:%.*]] = icmp eq <4 x i32> [[WIDE_MASKED_LOAD]], <i32 2, i32 2, i32 2, i32 2>
; CHECK-NEXT: [[TMP7:%.*]] = select <4 x i1> [[TMP6]], <4 x i32> <i32 1, i32 1, i32 1, i32 1>, <4 x i32> [[VEC_PHI]]
; CHECK-NEXT: [[PREDPHI]] = select <4 x i1> [[TMP3]], <4 x i32> [[TMP7]], <4 x i32> [[VEC_PHI]]
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 4
; CHECK-NEXT: [[TMP8:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP8]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <4 x i32> [[PREDPHI]], zeroinitializer
; CHECK-NEXT: [[TMP9:%.*]] = call i1 @llvm.vector.reduce.or.v4i1(<4 x i1> [[RDX_SELECT_CMP]])
; CHECK-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP9]], i32 1, i32 0
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[FOR_END_LOOPEXIT:%.*]], label [[SCALAR_PH]]
; CHECK: scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; CHECK-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: br label [[FOR_BODY:%.*]]
; CHECK: for.body:
; CHECK-NEXT: [[I_013:%.*]] = phi i64 [ [[INC:%.*]], [[FOR_INC:%.*]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[R_012:%.*]] = phi i32 [ [[R_1:%.*]], [[FOR_INC]] ], [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ]
; CHECK-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[SRC1]], i64 [[I_013]]
; CHECK-NEXT: [[TMP10:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
; CHECK-NEXT: [[CMP1:%.*]] = icmp sgt i32 [[TMP10]], 35
; CHECK-NEXT: br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_INC]]
; CHECK: if.then:
; CHECK-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds i32, ptr [[SRC2]], i64 [[I_013]]
; CHECK-NEXT: [[TMP11:%.*]] = load i32, ptr [[ARRAYIDX2]], align 4
; CHECK-NEXT: [[CMP3:%.*]] = icmp eq i32 [[TMP11]], 2
; CHECK-NEXT: [[SPEC_SELECT:%.*]] = select i1 [[CMP3]], i32 1, i32 [[R_012]]
; CHECK-NEXT: br label [[FOR_INC]]
; CHECK: for.inc:
; CHECK-NEXT: [[R_1]] = phi i32 [ [[R_012]], [[FOR_BODY]] ], [ [[SPEC_SELECT]], [[IF_THEN]] ]
; CHECK-NEXT: [[INC]] = add nuw nsw i64 [[I_013]], 1
; CHECK-NEXT: [[EXITCOND_NOT:%.*]] = icmp eq i64 [[INC]], [[N]]
; CHECK-NEXT: br i1 [[EXITCOND_NOT]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP13:![0-9]+]]
; CHECK: for.end.loopexit:
; CHECK-NEXT: [[R_1_LCSSA:%.*]] = phi i32 [ [[R_1]], [[FOR_INC]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; CHECK-NEXT: ret i32 [[R_1_LCSSA]]
;
; SCALABLE-LABEL: define i32 @pred_select_const_i32_from_icmp(
; SCALABLE-SAME: ptr noalias nocapture readonly [[SRC1:%.*]], ptr noalias nocapture readonly [[SRC2:%.*]], i64 [[N:%.*]]) #[[ATTR0]] {
; SCALABLE-NEXT: entry:
; SCALABLE-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP1:%.*]] = mul i64 [[TMP0]], 4
; SCALABLE-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i64 [[N]], [[TMP1]]
; SCALABLE-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[SCALAR_PH:%.*]], label [[VECTOR_PH:%.*]]
; SCALABLE: vector.ph:
; SCALABLE-NEXT: [[TMP2:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP3:%.*]] = mul i64 [[TMP2]], 4
; SCALABLE-NEXT: [[N_MOD_VF:%.*]] = urem i64 [[N]], [[TMP3]]
; SCALABLE-NEXT: [[N_VEC:%.*]] = sub i64 [[N]], [[N_MOD_VF]]
; SCALABLE-NEXT: [[TMP4:%.*]] = call i64 @llvm.vscale.i64()
; SCALABLE-NEXT: [[TMP5:%.*]] = mul i64 [[TMP4]], 4
; SCALABLE-NEXT: br label [[VECTOR_BODY:%.*]]
; SCALABLE: vector.body:
; SCALABLE-NEXT: [[INDEX:%.*]] = phi i64 [ 0, [[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[VEC_PHI:%.*]] = phi <vscale x 4 x i32> [ zeroinitializer, [[VECTOR_PH]] ], [ [[PREDPHI:%.*]], [[VECTOR_BODY]] ]
; SCALABLE-NEXT: [[TMP6:%.*]] = add i64 [[INDEX]], 0
; SCALABLE-NEXT: [[TMP7:%.*]] = getelementptr inbounds i32, ptr [[SRC1]], i64 [[TMP6]]
; SCALABLE-NEXT: [[TMP8:%.*]] = getelementptr inbounds i32, ptr [[TMP7]], i32 0
; SCALABLE-NEXT: [[WIDE_LOAD:%.*]] = load <vscale x 4 x i32>, ptr [[TMP8]], align 4
; SCALABLE-NEXT: [[TMP9:%.*]] = icmp sgt <vscale x 4 x i32> [[WIDE_LOAD]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 35, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
; SCALABLE-NEXT: [[TMP10:%.*]] = getelementptr i32, ptr [[SRC2]], i64 [[TMP6]]
; SCALABLE-NEXT: [[TMP11:%.*]] = getelementptr i32, ptr [[TMP10]], i32 0
; SCALABLE-NEXT: [[WIDE_MASKED_LOAD:%.*]] = call <vscale x 4 x i32> @llvm.masked.load.nxv4i32.p0(ptr [[TMP11]], i32 4, <vscale x 4 x i1> [[TMP9]], <vscale x 4 x i32> poison)
; SCALABLE-NEXT: [[TMP12:%.*]] = icmp eq <vscale x 4 x i32> [[WIDE_MASKED_LOAD]], shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 2, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer)
; SCALABLE-NEXT: [[TMP13:%.*]] = select <vscale x 4 x i1> [[TMP12]], <vscale x 4 x i32> shufflevector (<vscale x 4 x i32> insertelement (<vscale x 4 x i32> poison, i32 1, i64 0), <vscale x 4 x i32> poison, <vscale x 4 x i32> zeroinitializer), <vscale x 4 x i32> [[VEC_PHI]]
; SCALABLE-NEXT: [[PREDPHI]] = select <vscale x 4 x i1> [[TMP9]], <vscale x 4 x i32> [[TMP13]], <vscale x 4 x i32> [[VEC_PHI]]
; SCALABLE-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP5]]
; SCALABLE-NEXT: [[TMP14:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[TMP14]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP12:![0-9]+]]
; SCALABLE: middle.block:
; SCALABLE-NEXT: [[RDX_SELECT_CMP:%.*]] = icmp ne <vscale x 4 x i32> [[PREDPHI]], zeroinitializer
; SCALABLE-NEXT: [[TMP15:%.*]] = call i1 @llvm.vector.reduce.or.nxv4i1(<vscale x 4 x i1> [[RDX_SELECT_CMP]])
; SCALABLE-NEXT: [[RDX_SELECT:%.*]] = select i1 [[TMP15]], i32 1, i32 0
; SCALABLE-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N]], [[N_VEC]]
; SCALABLE-NEXT: br i1 [[CMP_N]], label [[FOR_END_LOOPEXIT:%.*]], label [[SCALAR_PH]]
; SCALABLE: scalar.ph:
; SCALABLE-NEXT: [[BC_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[MIDDLE_BLOCK]] ], [ 0, [[ENTRY:%.*]] ]
; SCALABLE-NEXT: [[BC_MERGE_RDX:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: br label [[FOR_BODY:%.*]]
; SCALABLE: for.body:
; SCALABLE-NEXT: [[I_013:%.*]] = phi i64 [ [[INC:%.*]], [[FOR_INC:%.*]] ], [ [[BC_RESUME_VAL]], [[SCALAR_PH]] ]
; SCALABLE-NEXT: [[R_012:%.*]] = phi i32 [ [[R_1:%.*]], [[FOR_INC]] ], [ [[BC_MERGE_RDX]], [[SCALAR_PH]] ]
; SCALABLE-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds i32, ptr [[SRC1]], i64 [[I_013]]
; SCALABLE-NEXT: [[TMP16:%.*]] = load i32, ptr [[ARRAYIDX]], align 4
; SCALABLE-NEXT: [[CMP1:%.*]] = icmp sgt i32 [[TMP16]], 35
; SCALABLE-NEXT: br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_INC]]
; SCALABLE: if.then:
; SCALABLE-NEXT: [[ARRAYIDX2:%.*]] = getelementptr inbounds i32, ptr [[SRC2]], i64 [[I_013]]
; SCALABLE-NEXT: [[TMP17:%.*]] = load i32, ptr [[ARRAYIDX2]], align 4
; SCALABLE-NEXT: [[CMP3:%.*]] = icmp eq i32 [[TMP17]], 2
; SCALABLE-NEXT: [[SPEC_SELECT:%.*]] = select i1 [[CMP3]], i32 1, i32 [[R_012]]
; SCALABLE-NEXT: br label [[FOR_INC]]
; SCALABLE: for.inc:
; SCALABLE-NEXT: [[R_1]] = phi i32 [ [[R_012]], [[FOR_BODY]] ], [ [[SPEC_SELECT]], [[IF_THEN]] ]
; SCALABLE-NEXT: [[INC]] = add nuw nsw i64 [[I_013]], 1
; SCALABLE-NEXT: [[EXITCOND_NOT:%.*]] = icmp eq i64 [[INC]], [[N]]
; SCALABLE-NEXT: br i1 [[EXITCOND_NOT]], label [[FOR_END_LOOPEXIT]], label [[FOR_BODY]], !llvm.loop [[LOOP13:![0-9]+]]
; SCALABLE: for.end.loopexit:
; SCALABLE-NEXT: [[R_1_LCSSA:%.*]] = phi i32 [ [[R_1]], [[FOR_INC]] ], [ [[RDX_SELECT]], [[MIDDLE_BLOCK]] ]
; SCALABLE-NEXT: ret i32 [[R_1_LCSSA]]
;
entry:
br label %for.body
for.body: ; preds = %entry, %for.inc
%i.013 = phi i64 [ %inc, %for.inc ], [ 0, %entry ]
%r.012 = phi i32 [ %r.1, %for.inc ], [ 0, %entry ]
%arrayidx = getelementptr inbounds i32, ptr %src1, i64 %i.013
%0 = load i32, ptr %arrayidx, align 4
%cmp1 = icmp sgt i32 %0, 35
br i1 %cmp1, label %if.then, label %for.inc
if.then: ; preds = %for.body
%arrayidx2 = getelementptr inbounds i32, ptr %src2, i64 %i.013
%1 = load i32, ptr %arrayidx2, align 4
%cmp3 = icmp eq i32 %1, 2
%spec.select = select i1 %cmp3, i32 1, i32 %r.012
br label %for.inc
for.inc: ; preds = %if.then, %for.body
%r.1 = phi i32 [ %r.012, %for.body ], [ %spec.select, %if.then ]
%inc = add nuw nsw i64 %i.013, 1
%exitcond.not = icmp eq i64 %inc, %n
br i1 %exitcond.not, label %for.end.loopexit, label %for.body
for.end.loopexit: ; preds = %for.inc
%r.1.lcssa = phi i32 [ %r.1, %for.inc ]
ret i32 %r.1.lcssa
}
attributes #0 = { "target-features"="+f,+v" }
;.
; CHECK: [[LOOP0]] = distinct !{[[LOOP0]], [[META1:![0-9]+]], [[META2:![0-9]+]]}
; CHECK: [[META1]] = !{!"llvm.loop.isvectorized", i32 1}
; CHECK: [[META2]] = !{!"llvm.loop.unroll.runtime.disable"}
; CHECK: [[LOOP3]] = distinct !{[[LOOP3]], [[META2]], [[META1]]}
; CHECK: [[LOOP4]] = distinct !{[[LOOP4]], [[META1]], [[META2]]}
; CHECK: [[LOOP5]] = distinct !{[[LOOP5]], [[META2]], [[META1]]}
; CHECK: [[LOOP6]] = distinct !{[[LOOP6]], [[META1]], [[META2]]}
; CHECK: [[LOOP7]] = distinct !{[[LOOP7]], [[META2]], [[META1]]}
; CHECK: [[LOOP8]] = distinct !{[[LOOP8]], [[META1]], [[META2]]}
; CHECK: [[LOOP9]] = distinct !{[[LOOP9]], [[META2]], [[META1]]}
; CHECK: [[LOOP10]] = distinct !{[[LOOP10]], [[META1]], [[META2]]}
; CHECK: [[LOOP11]] = distinct !{[[LOOP11]], [[META2]], [[META1]]}
; CHECK: [[LOOP12]] = distinct !{[[LOOP12]], [[META1]], [[META2]]}
; CHECK: [[LOOP13]] = distinct !{[[LOOP13]], [[META2]], [[META1]]}
;.
; SCALABLE: [[LOOP0]] = distinct !{[[LOOP0]], [[META1:![0-9]+]], [[META2:![0-9]+]]}
; SCALABLE: [[META1]] = !{!"llvm.loop.isvectorized", i32 1}
; SCALABLE: [[META2]] = !{!"llvm.loop.unroll.runtime.disable"}
; SCALABLE: [[LOOP3]] = distinct !{[[LOOP3]], [[META2]], [[META1]]}
; SCALABLE: [[LOOP4]] = distinct !{[[LOOP4]], [[META1]], [[META2]]}
; SCALABLE: [[LOOP5]] = distinct !{[[LOOP5]], [[META2]], [[META1]]}
; SCALABLE: [[LOOP6]] = distinct !{[[LOOP6]], [[META1]], [[META2]]}
; SCALABLE: [[LOOP7]] = distinct !{[[LOOP7]], [[META2]], [[META1]]}
; SCALABLE: [[LOOP8]] = distinct !{[[LOOP8]], [[META1]], [[META2]]}
; SCALABLE: [[LOOP9]] = distinct !{[[LOOP9]], [[META2]], [[META1]]}
; SCALABLE: [[LOOP10]] = distinct !{[[LOOP10]], [[META1]], [[META2]]}
; SCALABLE: [[LOOP11]] = distinct !{[[LOOP11]], [[META2]], [[META1]]}
; SCALABLE: [[LOOP12]] = distinct !{[[LOOP12]], [[META1]], [[META2]]}
; SCALABLE: [[LOOP13]] = distinct !{[[LOOP13]], [[META2]], [[META1]]}
;.
Reference in New Issue View Git Blame Copy Permalink