; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt -mcpu=bdver1 -passes=loop-vectorize,instcombine,simplifycfg < %s -S -o - | FileCheck %s --check-prefix=CHECK
; RUN: opt -mcpu=bdver1 -passes=loop-vectorize -debug-only=loop-vectorize -disable-output < %s 2>&1 | FileCheck %s --check-prefix=CHECK-COST
; REQUIRES: asserts
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"
; CHECK-COST-LABEL: uaddsat
define void @uaddsat(ptr nocapture readonly %pSrc, i16 signext %offset, ptr nocapture noalias %pDst, i32 %blockSize) #0 {
; CHECK-LABEL: @uaddsat(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP_NOT6:%.*]] = icmp eq i32 [[BLOCKSIZE:%.*]], 0
; CHECK-NEXT: br i1 [[CMP_NOT6]], label [[WHILE_END:%.*]], label [[ITER_CHECK:%.*]]
; CHECK: iter.check:
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[BLOCKSIZE]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[BLOCKSIZE]], 8
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
; CHECK: vector.main.loop.iter.check:
; CHECK-NEXT: [[MIN_ITERS_CHECK1:%.*]] = icmp ult i32 [[BLOCKSIZE]], 64
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_VEC:%.*]] = and i64 [[TMP0]], 4294967232
; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <16 x i16> poison, i16 [[OFFSET:%.*]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <16 x i16> [[BROADCAST_SPLATINSERT]], <16 x i16> poison, <16 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: [[TMP1:%.*]] = shl i64 [[INDEX]], 1
; CHECK-NEXT: [[NEXT_GEP:%.*]] = getelementptr i8, ptr [[PSRC:%.*]], i64 [[TMP1]]
; CHECK-NEXT: [[TMP2:%.*]] = shl i64 [[INDEX]], 1
; CHECK-NEXT: [[NEXT_GEP5:%.*]] = getelementptr i8, ptr [[PDST:%.*]], i64 [[TMP2]]
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr i8, ptr [[NEXT_GEP]], i64 32
; CHECK-NEXT: [[TMP4:%.*]] = getelementptr i8, ptr [[NEXT_GEP]], i64 64
; CHECK-NEXT: [[TMP5:%.*]] = getelementptr i8, ptr [[NEXT_GEP]], i64 96
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <16 x i16>, ptr [[NEXT_GEP]], align 2
; CHECK-NEXT: [[WIDE_LOAD9:%.*]] = load <16 x i16>, ptr [[TMP3]], align 2
; CHECK-NEXT: [[WIDE_LOAD10:%.*]] = load <16 x i16>, ptr [[TMP4]], align 2
; CHECK-NEXT: [[WIDE_LOAD11:%.*]] = load <16 x i16>, ptr [[TMP5]], align 2
; CHECK-NEXT: [[TMP6:%.*]] = call <16 x i16> @llvm.uadd.sat.v16i16(<16 x i16> [[WIDE_LOAD]], <16 x i16> [[BROADCAST_SPLAT]])
; CHECK-NEXT: [[TMP7:%.*]] = call <16 x i16> @llvm.uadd.sat.v16i16(<16 x i16> [[WIDE_LOAD9]], <16 x i16> [[BROADCAST_SPLAT]])
; CHECK-NEXT: [[TMP8:%.*]] = call <16 x i16> @llvm.uadd.sat.v16i16(<16 x i16> [[WIDE_LOAD10]], <16 x i16> [[BROADCAST_SPLAT]])
; CHECK-NEXT: [[TMP9:%.*]] = call <16 x i16> @llvm.uadd.sat.v16i16(<16 x i16> [[WIDE_LOAD11]], <16 x i16> [[BROADCAST_SPLAT]])
; CHECK-NEXT: [[TMP10:%.*]] = getelementptr i8, ptr [[NEXT_GEP5]], i64 32
; CHECK-NEXT: [[TMP11:%.*]] = getelementptr i8, ptr [[NEXT_GEP5]], i64 64
; CHECK-NEXT: [[TMP12:%.*]] = getelementptr i8, ptr [[NEXT_GEP5]], i64 96
; CHECK-NEXT: store <16 x i16> [[TMP6]], ptr [[NEXT_GEP5]], align 2
; CHECK-NEXT: store <16 x i16> [[TMP7]], ptr [[TMP10]], align 2
; CHECK-NEXT: store <16 x i16> [[TMP8]], ptr [[TMP11]], align 2
; CHECK-NEXT: store <16 x i16> [[TMP9]], ptr [[TMP12]], align 2
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 64
; CHECK-NEXT: [[TMP13:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP13]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N_VEC]], [[TMP0]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[WHILE_END]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
; CHECK: vec.epilog.iter.check:
; CHECK-NEXT: [[TMP14:%.*]] = shl nuw nsw i64 [[N_VEC]], 1
; CHECK-NEXT: [[IND_END24:%.*]] = getelementptr i8, ptr [[PDST]], i64 [[TMP14]]
; CHECK-NEXT: [[TMP15:%.*]] = shl nuw nsw i64 [[N_VEC]], 1
; CHECK-NEXT: [[IND_END21:%.*]] = getelementptr i8, ptr [[PSRC]], i64 [[TMP15]]
; CHECK-NEXT: [[DOTCAST17:%.*]] = trunc nuw i64 [[N_VEC]] to i32
; CHECK-NEXT: [[IND_END18:%.*]] = sub i32 [[BLOCKSIZE]], [[DOTCAST17]]
; CHECK-NEXT: [[N_VEC_REMAINING:%.*]] = and i64 [[TMP0]], 56
; CHECK-NEXT: [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp eq i64 [[N_VEC_REMAINING]], 0
; CHECK-NEXT: br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
; CHECK: vec.epilog.ph:
; CHECK-NEXT: [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT: [[N_VEC15:%.*]] = and i64 [[TMP0]], 4294967288
; CHECK-NEXT: [[DOTCAST:%.*]] = trunc nuw i64 [[N_VEC15]] to i32
; CHECK-NEXT: [[IND_END16:%.*]] = sub i32 [[BLOCKSIZE]], [[DOTCAST]]
; CHECK-NEXT: [[TMP16:%.*]] = shl nuw nsw i64 [[N_VEC15]], 1
; CHECK-NEXT: [[IND_END20:%.*]] = getelementptr i8, ptr [[PSRC]], i64 [[TMP16]]
; CHECK-NEXT: [[TMP17:%.*]] = shl nuw nsw i64 [[N_VEC15]], 1
; CHECK-NEXT: [[IND_END23:%.*]] = getelementptr i8, ptr [[PDST]], i64 [[TMP17]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT31:%.*]] = insertelement <8 x i16> poison, i16 [[OFFSET]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT32:%.*]] = shufflevector <8 x i16> [[BROADCAST_SPLATINSERT31]], <8 x i16> poison, <8 x i32> zeroinitializer
; CHECK-NEXT: br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
; CHECK: vec.epilog.vector.body:
; CHECK-NEXT: [[INDEX27:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT33:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT: [[TMP18:%.*]] = shl i64 [[INDEX27]], 1
; CHECK-NEXT: [[NEXT_GEP28:%.*]] = getelementptr i8, ptr [[PSRC]], i64 [[TMP18]]
; CHECK-NEXT: [[TMP19:%.*]] = shl i64 [[INDEX27]], 1
; CHECK-NEXT: [[NEXT_GEP29:%.*]] = getelementptr i8, ptr [[PDST]], i64 [[TMP19]]
; CHECK-NEXT: [[WIDE_LOAD30:%.*]] = load <8 x i16>, ptr [[NEXT_GEP28]], align 2
; CHECK-NEXT: [[TMP20:%.*]] = call <8 x i16> @llvm.uadd.sat.v8i16(<8 x i16> [[WIDE_LOAD30]], <8 x i16> [[BROADCAST_SPLAT32]])
; CHECK-NEXT: store <8 x i16> [[TMP20]], ptr [[NEXT_GEP29]], align 2
; CHECK-NEXT: [[INDEX_NEXT33]] = add nuw i64 [[INDEX27]], 8
; CHECK-NEXT: [[TMP21:%.*]] = icmp eq i64 [[INDEX_NEXT33]], [[N_VEC15]]
; CHECK-NEXT: br i1 [[TMP21]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP3:![0-9]+]]
; CHECK: vec.epilog.middle.block:
; CHECK-NEXT: [[CMP_N26:%.*]] = icmp eq i64 [[N_VEC15]], [[TMP0]]
; CHECK-NEXT: br i1 [[CMP_N26]], label [[WHILE_END]], label [[VEC_EPILOG_SCALAR_PH]]
; CHECK: vec.epilog.scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL19:%.*]] = phi i32 [ [[IND_END16]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END18]], [[VEC_EPILOG_ITER_CHECK]] ], [ [[BLOCKSIZE]], [[ITER_CHECK]] ]
; CHECK-NEXT: [[BC_RESUME_VAL22:%.*]] = phi ptr [ [[IND_END20]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END21]], [[VEC_EPILOG_ITER_CHECK]] ], [ [[PSRC]], [[ITER_CHECK]] ]
; CHECK-NEXT: [[BC_RESUME_VAL25:%.*]] = phi ptr [ [[IND_END23]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END24]], [[VEC_EPILOG_ITER_CHECK]] ], [ [[PDST]], [[ITER_CHECK]] ]
; CHECK-NEXT: br label [[WHILE_BODY:%.*]]
; CHECK: while.body:
; CHECK-NEXT: [[BLKCNT_09:%.*]] = phi i32 [ [[DEC:%.*]], [[WHILE_BODY]] ], [ [[BC_RESUME_VAL19]], [[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT: [[PSRC_ADDR_08:%.*]] = phi ptr [ [[INCDEC_PTR:%.*]], [[WHILE_BODY]] ], [ [[BC_RESUME_VAL22]], [[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT: [[PDST_ADDR_07:%.*]] = phi ptr [ [[INCDEC_PTR3:%.*]], [[WHILE_BODY]] ], [ [[BC_RESUME_VAL25]], [[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT: [[INCDEC_PTR]] = getelementptr inbounds i8, ptr [[PSRC_ADDR_08]], i64 2
; CHECK-NEXT: [[TMP22:%.*]] = load i16, ptr [[PSRC_ADDR_08]], align 2
; CHECK-NEXT: [[TMP23:%.*]] = tail call i16 @llvm.uadd.sat.i16(i16 [[TMP22]], i16 [[OFFSET]])
; CHECK-NEXT: [[INCDEC_PTR3]] = getelementptr inbounds i8, ptr [[PDST_ADDR_07]], i64 2
; CHECK-NEXT: store i16 [[TMP23]], ptr [[PDST_ADDR_07]], align 2
; CHECK-NEXT: [[DEC]] = add i32 [[BLKCNT_09]], -1
; CHECK-NEXT: [[CMP_NOT:%.*]] = icmp eq i32 [[DEC]], 0
; CHECK-NEXT: br i1 [[CMP_NOT]], label [[WHILE_END]], label [[WHILE_BODY]], !llvm.loop [[LOOP4:![0-9]+]]
; CHECK: while.end:
; CHECK-NEXT: ret void
;
entry:
%cmp.not6 = icmp eq i32 %blockSize, 0
br i1 %cmp.not6, label %while.end, label %while.body
while.body: ; preds = %entry, %while.body
%blkCnt.09 = phi i32 [ %dec, %while.body ], [ %blockSize, %entry ]
%pSrc.addr.08 = phi ptr [ %incdec.ptr, %while.body ], [ %pSrc, %entry ]
%pDst.addr.07 = phi ptr [ %incdec.ptr3, %while.body ], [ %pDst, %entry ]
%incdec.ptr = getelementptr inbounds i16, ptr %pSrc.addr.08, i32 1
%0 = load i16, ptr %pSrc.addr.08, align 2
%1 = tail call i16 @llvm.uadd.sat.i16(i16 %0, i16 %offset)
%incdec.ptr3 = getelementptr inbounds i16, ptr %pDst.addr.07, i32 1
store i16 %1, ptr %pDst.addr.07, align 2
%dec = add i32 %blkCnt.09, -1
%cmp.not = icmp eq i32 %dec, 0
br i1 %cmp.not, label %while.end, label %while.body
while.end: ; preds = %while.body, %entry
ret void
}
; CHECK-COST-LABEL: fshl
define void @fshl(ptr nocapture readonly %pSrc, i8 signext %offset, ptr nocapture noalias %pDst, i32 %blockSize) #0 {
; CHECK-LABEL: @fshl(
; CHECK-NEXT: entry:
; CHECK-NEXT: [[CMP_NOT6:%.*]] = icmp eq i32 [[BLOCKSIZE:%.*]], 0
; CHECK-NEXT: br i1 [[CMP_NOT6]], label [[WHILE_END:%.*]], label [[ITER_CHECK:%.*]]
; CHECK: iter.check:
; CHECK-NEXT: [[TMP0:%.*]] = zext i32 [[BLOCKSIZE]] to i64
; CHECK-NEXT: [[MIN_ITERS_CHECK:%.*]] = icmp ult i32 [[BLOCKSIZE]], 16
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH:%.*]], label [[VECTOR_MAIN_LOOP_ITER_CHECK:%.*]]
; CHECK: vector.main.loop.iter.check:
; CHECK-NEXT: [[MIN_ITERS_CHECK1:%.*]] = icmp ult i32 [[BLOCKSIZE]], 128
; CHECK-NEXT: br i1 [[MIN_ITERS_CHECK1]], label [[VEC_EPILOG_PH:%.*]], label [[VECTOR_PH:%.*]]
; CHECK: vector.ph:
; CHECK-NEXT: [[N_VEC:%.*]] = and i64 [[TMP0]], 4294967168
; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <32 x i8> poison, i8 [[OFFSET:%.*]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <32 x i8> [[BROADCAST_SPLATINSERT]], <32 x i8> poison, <32 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: [[NEXT_GEP:%.*]] = getelementptr i8, ptr [[PSRC:%.*]], i64 [[INDEX]]
; CHECK-NEXT: [[NEXT_GEP5:%.*]] = getelementptr i8, ptr [[PDST:%.*]], i64 [[INDEX]]
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr i8, ptr [[NEXT_GEP]], i64 32
; CHECK-NEXT: [[TMP2:%.*]] = getelementptr i8, ptr [[NEXT_GEP]], i64 64
; CHECK-NEXT: [[TMP3:%.*]] = getelementptr i8, ptr [[NEXT_GEP]], i64 96
; CHECK-NEXT: [[WIDE_LOAD:%.*]] = load <32 x i8>, ptr [[NEXT_GEP]], align 2
; CHECK-NEXT: [[WIDE_LOAD9:%.*]] = load <32 x i8>, ptr [[TMP1]], align 2
; CHECK-NEXT: [[WIDE_LOAD10:%.*]] = load <32 x i8>, ptr [[TMP2]], align 2
; CHECK-NEXT: [[WIDE_LOAD11:%.*]] = load <32 x i8>, ptr [[TMP3]], align 2
; CHECK-NEXT: [[TMP4:%.*]] = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> [[WIDE_LOAD]], <32 x i8> [[WIDE_LOAD]], <32 x i8> [[BROADCAST_SPLAT]])
; CHECK-NEXT: [[TMP5:%.*]] = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> [[WIDE_LOAD9]], <32 x i8> [[WIDE_LOAD9]], <32 x i8> [[BROADCAST_SPLAT]])
; CHECK-NEXT: [[TMP6:%.*]] = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> [[WIDE_LOAD10]], <32 x i8> [[WIDE_LOAD10]], <32 x i8> [[BROADCAST_SPLAT]])
; CHECK-NEXT: [[TMP7:%.*]] = call <32 x i8> @llvm.fshl.v32i8(<32 x i8> [[WIDE_LOAD11]], <32 x i8> [[WIDE_LOAD11]], <32 x i8> [[BROADCAST_SPLAT]])
; CHECK-NEXT: [[TMP8:%.*]] = getelementptr i8, ptr [[NEXT_GEP5]], i64 32
; CHECK-NEXT: [[TMP9:%.*]] = getelementptr i8, ptr [[NEXT_GEP5]], i64 64
; CHECK-NEXT: [[TMP10:%.*]] = getelementptr i8, ptr [[NEXT_GEP5]], i64 96
; CHECK-NEXT: store <32 x i8> [[TMP4]], ptr [[NEXT_GEP5]], align 2
; CHECK-NEXT: store <32 x i8> [[TMP5]], ptr [[TMP8]], align 2
; CHECK-NEXT: store <32 x i8> [[TMP6]], ptr [[TMP9]], align 2
; CHECK-NEXT: store <32 x i8> [[TMP7]], ptr [[TMP10]], align 2
; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 128
; CHECK-NEXT: [[TMP11:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
; CHECK-NEXT: br i1 [[TMP11]], label [[MIDDLE_BLOCK:%.*]], label [[VECTOR_BODY]], !llvm.loop [[LOOP5:![0-9]+]]
; CHECK: middle.block:
; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 [[N_VEC]], [[TMP0]]
; CHECK-NEXT: br i1 [[CMP_N]], label [[WHILE_END]], label [[VEC_EPILOG_ITER_CHECK:%.*]]
; CHECK: vec.epilog.iter.check:
; CHECK-NEXT: [[IND_END24:%.*]] = getelementptr i8, ptr [[PDST]], i64 [[N_VEC]]
; CHECK-NEXT: [[IND_END21:%.*]] = getelementptr i8, ptr [[PSRC]], i64 [[N_VEC]]
; CHECK-NEXT: [[DOTCAST17:%.*]] = trunc nuw i64 [[N_VEC]] to i32
; CHECK-NEXT: [[IND_END18:%.*]] = sub i32 [[BLOCKSIZE]], [[DOTCAST17]]
; CHECK-NEXT: [[N_VEC_REMAINING:%.*]] = and i64 [[TMP0]], 112
; CHECK-NEXT: [[MIN_EPILOG_ITERS_CHECK:%.*]] = icmp eq i64 [[N_VEC_REMAINING]], 0
; CHECK-NEXT: br i1 [[MIN_EPILOG_ITERS_CHECK]], label [[VEC_EPILOG_SCALAR_PH]], label [[VEC_EPILOG_PH]]
; CHECK: vec.epilog.ph:
; CHECK-NEXT: [[VEC_EPILOG_RESUME_VAL:%.*]] = phi i64 [ [[N_VEC]], [[VEC_EPILOG_ITER_CHECK]] ], [ 0, [[VECTOR_MAIN_LOOP_ITER_CHECK]] ]
; CHECK-NEXT: [[N_VEC15:%.*]] = and i64 [[TMP0]], 4294967280
; CHECK-NEXT: [[DOTCAST:%.*]] = trunc nuw i64 [[N_VEC15]] to i32
; CHECK-NEXT: [[IND_END16:%.*]] = sub i32 [[BLOCKSIZE]], [[DOTCAST]]
; CHECK-NEXT: [[IND_END20:%.*]] = getelementptr i8, ptr [[PSRC]], i64 [[N_VEC15]]
; CHECK-NEXT: [[IND_END23:%.*]] = getelementptr i8, ptr [[PDST]], i64 [[N_VEC15]]
; CHECK-NEXT: [[BROADCAST_SPLATINSERT31:%.*]] = insertelement <16 x i8> poison, i8 [[OFFSET]], i64 0
; CHECK-NEXT: [[BROADCAST_SPLAT32:%.*]] = shufflevector <16 x i8> [[BROADCAST_SPLATINSERT31]], <16 x i8> poison, <16 x i32> zeroinitializer
; CHECK-NEXT: br label [[VEC_EPILOG_VECTOR_BODY:%.*]]
; CHECK: vec.epilog.vector.body:
; CHECK-NEXT: [[INDEX27:%.*]] = phi i64 [ [[VEC_EPILOG_RESUME_VAL]], [[VEC_EPILOG_PH]] ], [ [[INDEX_NEXT33:%.*]], [[VEC_EPILOG_VECTOR_BODY]] ]
; CHECK-NEXT: [[NEXT_GEP28:%.*]] = getelementptr i8, ptr [[PSRC]], i64 [[INDEX27]]
; CHECK-NEXT: [[NEXT_GEP29:%.*]] = getelementptr i8, ptr [[PDST]], i64 [[INDEX27]]
; CHECK-NEXT: [[WIDE_LOAD30:%.*]] = load <16 x i8>, ptr [[NEXT_GEP28]], align 2
; CHECK-NEXT: [[TMP12:%.*]] = call <16 x i8> @llvm.fshl.v16i8(<16 x i8> [[WIDE_LOAD30]], <16 x i8> [[WIDE_LOAD30]], <16 x i8> [[BROADCAST_SPLAT32]])
; CHECK-NEXT: store <16 x i8> [[TMP12]], ptr [[NEXT_GEP29]], align 2
; CHECK-NEXT: [[INDEX_NEXT33]] = add nuw i64 [[INDEX27]], 16
; CHECK-NEXT: [[TMP13:%.*]] = icmp eq i64 [[INDEX_NEXT33]], [[N_VEC15]]
; CHECK-NEXT: br i1 [[TMP13]], label [[VEC_EPILOG_MIDDLE_BLOCK:%.*]], label [[VEC_EPILOG_VECTOR_BODY]], !llvm.loop [[LOOP6:![0-9]+]]
; CHECK: vec.epilog.middle.block:
; CHECK-NEXT: [[CMP_N26:%.*]] = icmp eq i64 [[N_VEC15]], [[TMP0]]
; CHECK-NEXT: br i1 [[CMP_N26]], label [[WHILE_END]], label [[VEC_EPILOG_SCALAR_PH]]
; CHECK: vec.epilog.scalar.ph:
; CHECK-NEXT: [[BC_RESUME_VAL19:%.*]] = phi i32 [ [[IND_END16]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END18]], [[VEC_EPILOG_ITER_CHECK]] ], [ [[BLOCKSIZE]], [[ITER_CHECK]] ]
; CHECK-NEXT: [[BC_RESUME_VAL22:%.*]] = phi ptr [ [[IND_END20]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END21]], [[VEC_EPILOG_ITER_CHECK]] ], [ [[PSRC]], [[ITER_CHECK]] ]
; CHECK-NEXT: [[BC_RESUME_VAL25:%.*]] = phi ptr [ [[IND_END23]], [[VEC_EPILOG_MIDDLE_BLOCK]] ], [ [[IND_END24]], [[VEC_EPILOG_ITER_CHECK]] ], [ [[PDST]], [[ITER_CHECK]] ]
; CHECK-NEXT: br label [[WHILE_BODY:%.*]]
; CHECK: while.body:
; CHECK-NEXT: [[BLKCNT_09:%.*]] = phi i32 [ [[DEC:%.*]], [[WHILE_BODY]] ], [ [[BC_RESUME_VAL19]], [[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT: [[PSRC_ADDR_08:%.*]] = phi ptr [ [[INCDEC_PTR:%.*]], [[WHILE_BODY]] ], [ [[BC_RESUME_VAL22]], [[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT: [[PDST_ADDR_07:%.*]] = phi ptr [ [[INCDEC_PTR3:%.*]], [[WHILE_BODY]] ], [ [[BC_RESUME_VAL25]], [[VEC_EPILOG_SCALAR_PH]] ]
; CHECK-NEXT: [[INCDEC_PTR]] = getelementptr inbounds i8, ptr [[PSRC_ADDR_08]], i64 1
; CHECK-NEXT: [[TMP14:%.*]] = load i8, ptr [[PSRC_ADDR_08]], align 2
; CHECK-NEXT: [[TMP15:%.*]] = tail call i8 @llvm.fshl.i8(i8 [[TMP14]], i8 [[TMP14]], i8 [[OFFSET]])
; CHECK-NEXT: [[INCDEC_PTR3]] = getelementptr inbounds i8, ptr [[PDST_ADDR_07]], i64 1
; CHECK-NEXT: store i8 [[TMP15]], ptr [[PDST_ADDR_07]], align 2
; CHECK-NEXT: [[DEC]] = add i32 [[BLKCNT_09]], -1
; CHECK-NEXT: [[CMP_NOT:%.*]] = icmp eq i32 [[DEC]], 0
; CHECK-NEXT: br i1 [[CMP_NOT]], label [[WHILE_END]], label [[WHILE_BODY]], !llvm.loop [[LOOP7:![0-9]+]]
; CHECK: while.end:
; CHECK-NEXT: ret void
;
entry:
%cmp.not6 = icmp eq i32 %blockSize, 0
br i1 %cmp.not6, label %while.end, label %while.body
while.body: ; preds = %entry, %while.body
%blkCnt.09 = phi i32 [ %dec, %while.body ], [ %blockSize, %entry ]
%pSrc.addr.08 = phi ptr [ %incdec.ptr, %while.body ], [ %pSrc, %entry ]
%pDst.addr.07 = phi ptr [ %incdec.ptr3, %while.body ], [ %pDst, %entry ]
%incdec.ptr = getelementptr inbounds i8, ptr %pSrc.addr.08, i32 1
%0 = load i8, ptr %pSrc.addr.08, align 2
%1 = tail call i8 @llvm.fshl.i8(i8 %0, i8 %0, i8 %offset)
%incdec.ptr3 = getelementptr inbounds i8, ptr %pDst.addr.07, i32 1
store i8 %1, ptr %pDst.addr.07, align 2
%dec = add i32 %blkCnt.09, -1
%cmp.not = icmp eq i32 %dec, 0
br i1 %cmp.not, label %while.end, label %while.body
while.end: ; preds = %while.body, %entry
ret void
}
declare i16 @llvm.uadd.sat.i16(i16, i16)
declare i8 @llvm.fshl.i8(i8, i8, i8)
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; CHECK-COST: {{.*}}