//===-- X86TargetTransformInfo.cpp - X86 specific TTI pass ----------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
/// \file
/// This file implements a TargetTransformInfo analysis pass specific to the
/// X86 target machine. It uses the target's detailed information to provide
/// more precise answers to certain TTI queries, while letting the target
/// independent and default TTI implementations handle the rest.
///
//===----------------------------------------------------------------------===//
#include "X86TargetTransformInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/CodeGen/BasicTTIImpl.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/Support/Debug.h"
#include "llvm/Target/CostTable.h"
#include "llvm/Target/TargetLowering.h"
using namespace llvm;
#define DEBUG_TYPE "x86tti"
//===----------------------------------------------------------------------===//
//
// X86 cost model.
//
//===----------------------------------------------------------------------===//
TargetTransformInfo::PopcntSupportKind
X86TTIImpl::getPopcntSupport(unsigned TyWidth) {
assert(isPowerOf2_32(TyWidth) && "Ty width must be power of 2");
// TODO: Currently the __builtin_popcount() implementation using SSE3
// instructions is inefficient. Once the problem is fixed, we should
// call ST->hasSSE3() instead of ST->hasPOPCNT().
return ST->hasPOPCNT() ? TTI::PSK_FastHardware : TTI::PSK_Software;
}
unsigned X86TTIImpl::getNumberOfRegisters(bool Vector) {
if (Vector && !ST->hasSSE1())
return 0;
if (ST->is64Bit()) {
if (Vector && ST->hasAVX512())
return 32;
return 16;
}
return 8;
}
unsigned X86TTIImpl::getRegisterBitWidth(bool Vector) {
if (Vector) {
if (ST->hasAVX512()) return 512;
if (ST->hasAVX()) return 256;
if (ST->hasSSE1()) return 128;
return 0;
}
if (ST->is64Bit())
return 64;
return 32;
}
unsigned X86TTIImpl::getMaxInterleaveFactor(unsigned VF) {
// If the loop will not be vectorized, don't interleave the loop.
// Let regular unroll to unroll the loop, which saves the overflow
// check and memory check cost.
if (VF == 1)
return 1;
if (ST->isAtom())
return 1;
// Sandybridge and Haswell have multiple execution ports and pipelined
// vector units.
if (ST->hasAVX())
return 4;
return 2;
}
int X86TTIImpl::getArithmeticInstrCost(
unsigned Opcode, Type *Ty, TTI::OperandValueKind Op1Info,
TTI::OperandValueKind Op2Info, TTI::OperandValueProperties Opd1PropInfo,
TTI::OperandValueProperties Opd2PropInfo) {
// Legalize the type.
std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, Ty);
int ISD = TLI->InstructionOpcodeToISD(Opcode);
assert(ISD && "Invalid opcode");
if (ISD == ISD::SDIV &&
Op2Info == TargetTransformInfo::OK_UniformConstantValue &&
Opd2PropInfo == TargetTransformInfo::OP_PowerOf2) {
// On X86, vector signed division by constants power-of-two are
// normally expanded to the sequence SRA + SRL + ADD + SRA.
// The OperandValue properties many not be same as that of previous
// operation;conservatively assume OP_None.
int Cost = 2 * getArithmeticInstrCost(Instruction::AShr, Ty, Op1Info,
Op2Info, TargetTransformInfo::OP_None,
TargetTransformInfo::OP_None);
Cost += getArithmeticInstrCost(Instruction::LShr, Ty, Op1Info, Op2Info,
TargetTransformInfo::OP_None,
TargetTransformInfo::OP_None);
Cost += getArithmeticInstrCost(Instruction::Add, Ty, Op1Info, Op2Info,
TargetTransformInfo::OP_None,
TargetTransformInfo::OP_None);
return Cost;
}
static const CostTblEntry AVX2UniformConstCostTable[] = {
{ ISD::SRA, MVT::v4i64, 4 }, // 2 x psrad + shuffle.
{ ISD::SDIV, MVT::v16i16, 6 }, // vpmulhw sequence
{ ISD::UDIV, MVT::v16i16, 6 }, // vpmulhuw sequence
{ ISD::SDIV, MVT::v8i32, 15 }, // vpmuldq sequence
{ ISD::UDIV, MVT::v8i32, 15 }, // vpmuludq sequence
};
if (Op2Info == TargetTransformInfo::OK_UniformConstantValue &&
ST->hasAVX2()) {
if (const auto *Entry = CostTableLookup(AVX2UniformConstCostTable, ISD,
LT.second))
return LT.first * Entry->Cost;
}
static const CostTblEntry AVX512CostTable[] = {
{ ISD::SHL, MVT::v16i32, 1 },
{ ISD::SRL, MVT::v16i32, 1 },
{ ISD::SRA, MVT::v16i32, 1 },
{ ISD::SHL, MVT::v8i64, 1 },
{ ISD::SRL, MVT::v8i64, 1 },
{ ISD::SRA, MVT::v8i64, 1 },
};
if (ST->hasAVX512()) {
if (const auto *Entry = CostTableLookup(AVX512CostTable, ISD, LT.second))
return LT.first * Entry->Cost;
}
static const CostTblEntry AVX2CostTable[] = {
// Shifts on v4i64/v8i32 on AVX2 is legal even though we declare to
// customize them to detect the cases where shift amount is a scalar one.
{ ISD::SHL, MVT::v4i32, 1 },
{ ISD::SRL, MVT::v4i32, 1 },
{ ISD::SRA, MVT::v4i32, 1 },
{ ISD::SHL, MVT::v8i32, 1 },
{ ISD::SRL, MVT::v8i32, 1 },
{ ISD::SRA, MVT::v8i32, 1 },
{ ISD::SHL, MVT::v2i64, 1 },
{ ISD::SRL, MVT::v2i64, 1 },
{ ISD::SHL, MVT::v4i64, 1 },
{ ISD::SRL, MVT::v4i64, 1 },
};
// Look for AVX2 lowering tricks.
if (ST->hasAVX2()) {
if (ISD == ISD::SHL && LT.second == MVT::v16i16 &&
(Op2Info == TargetTransformInfo::OK_UniformConstantValue ||
Op2Info == TargetTransformInfo::OK_NonUniformConstantValue))
// On AVX2, a packed v16i16 shift left by a constant build_vector
// is lowered into a vector multiply (vpmullw).
return LT.first;
if (const auto *Entry = CostTableLookup(AVX2CostTable, ISD, LT.second))
return LT.first * Entry->Cost;
}
static const CostTblEntry XOPCostTable[] = {
// 128bit shifts take 1cy, but right shifts require negation beforehand.
{ ISD::SHL, MVT::v16i8, 1 },
{ ISD::SRL, MVT::v16i8, 2 },
{ ISD::SRA, MVT::v16i8, 2 },
{ ISD::SHL, MVT::v8i16, 1 },
{ ISD::SRL, MVT::v8i16, 2 },
{ ISD::SRA, MVT::v8i16, 2 },
{ ISD::SHL, MVT::v4i32, 1 },
{ ISD::SRL, MVT::v4i32, 2 },
{ ISD::SRA, MVT::v4i32, 2 },
{ ISD::SHL, MVT::v2i64, 1 },
{ ISD::SRL, MVT::v2i64, 2 },
{ ISD::SRA, MVT::v2i64, 2 },
// 256bit shifts require splitting if AVX2 didn't catch them above.
{ ISD::SHL, MVT::v32i8, 2 },
{ ISD::SRL, MVT::v32i8, 4 },
{ ISD::SRA, MVT::v32i8, 4 },
{ ISD::SHL, MVT::v16i16, 2 },
{ ISD::SRL, MVT::v16i16, 4 },
{ ISD::SRA, MVT::v16i16, 4 },
{ ISD::SHL, MVT::v8i32, 2 },
{ ISD::SRL, MVT::v8i32, 4 },
{ ISD::SRA, MVT::v8i32, 4 },
{ ISD::SHL, MVT::v4i64, 2 },
{ ISD::SRL, MVT::v4i64, 4 },
{ ISD::SRA, MVT::v4i64, 4 },
};
// Look for XOP lowering tricks.
if (ST->hasXOP()) {
if (const auto *Entry = CostTableLookup(XOPCostTable, ISD, LT.second))
return LT.first * Entry->Cost;
}
static const CostTblEntry AVX2CustomCostTable[] = {
{ ISD::SHL, MVT::v32i8, 11 }, // vpblendvb sequence.
{ ISD::SHL, MVT::v16i16, 10 }, // extend/vpsrlvd/pack sequence.
{ ISD::SRL, MVT::v32i8, 11 }, // vpblendvb sequence.
{ ISD::SRL, MVT::v16i16, 10 }, // extend/vpsrlvd/pack sequence.
{ ISD::SRA, MVT::v32i8, 24 }, // vpblendvb sequence.
{ ISD::SRA, MVT::v16i16, 10 }, // extend/vpsravd/pack sequence.
{ ISD::SRA, MVT::v2i64, 4 }, // srl/xor/sub sequence.
{ ISD::SRA, MVT::v4i64, 4 }, // srl/xor/sub sequence.
// Vectorizing division is a bad idea. See the SSE2 table for more comments.
{ ISD::SDIV, MVT::v32i8, 32*20 },
{ ISD::SDIV, MVT::v16i16, 16*20 },
{ ISD::SDIV, MVT::v8i32, 8*20 },
{ ISD::SDIV, MVT::v4i64, 4*20 },
{ ISD::UDIV, MVT::v32i8, 32*20 },
{ ISD::UDIV, MVT::v16i16, 16*20 },
{ ISD::UDIV, MVT::v8i32, 8*20 },
{ ISD::UDIV, MVT::v4i64, 4*20 },
};
// Look for AVX2 lowering tricks for custom cases.
if (ST->hasAVX2()) {
if (const auto *Entry = CostTableLookup(AVX2CustomCostTable, ISD,
LT.second))
return LT.first * Entry->Cost;
}
static const CostTblEntry
SSE2UniformConstCostTable[] = {
// We don't correctly identify costs of casts because they are marked as
// custom.
// Constant splats are cheaper for the following instructions.
{ ISD::SHL, MVT::v16i8, 1 }, // psllw.
{ ISD::SHL, MVT::v32i8, 2 }, // psllw.
{ ISD::SHL, MVT::v8i16, 1 }, // psllw.
{ ISD::SHL, MVT::v16i16, 2 }, // psllw.
{ ISD::SHL, MVT::v4i32, 1 }, // pslld
{ ISD::SHL, MVT::v8i32, 2 }, // pslld
{ ISD::SHL, MVT::v2i64, 1 }, // psllq.
{ ISD::SHL, MVT::v4i64, 2 }, // psllq.
{ ISD::SRL, MVT::v16i8, 1 }, // psrlw.
{ ISD::SRL, MVT::v32i8, 2 }, // psrlw.
{ ISD::SRL, MVT::v8i16, 1 }, // psrlw.
{ ISD::SRL, MVT::v16i16, 2 }, // psrlw.
{ ISD::SRL, MVT::v4i32, 1 }, // psrld.
{ ISD::SRL, MVT::v8i32, 2 }, // psrld.
{ ISD::SRL, MVT::v2i64, 1 }, // psrlq.
{ ISD::SRL, MVT::v4i64, 2 }, // psrlq.
{ ISD::SRA, MVT::v16i8, 4 }, // psrlw, pand, pxor, psubb.
{ ISD::SRA, MVT::v32i8, 8 }, // psrlw, pand, pxor, psubb.
{ ISD::SRA, MVT::v8i16, 1 }, // psraw.
{ ISD::SRA, MVT::v16i16, 2 }, // psraw.
{ ISD::SRA, MVT::v4i32, 1 }, // psrad.
{ ISD::SRA, MVT::v8i32, 2 }, // psrad.
{ ISD::SRA, MVT::v2i64, 4 }, // 2 x psrad + shuffle.
{ ISD::SRA, MVT::v4i64, 8 }, // 2 x psrad + shuffle.
{ ISD::SDIV, MVT::v8i16, 6 }, // pmulhw sequence
{ ISD::UDIV, MVT::v8i16, 6 }, // pmulhuw sequence
{ ISD::SDIV, MVT::v4i32, 19 }, // pmuludq sequence
{ ISD::UDIV, MVT::v4i32, 15 }, // pmuludq sequence
};
if (Op2Info == TargetTransformInfo::OK_UniformConstantValue &&
ST->hasSSE2()) {
// pmuldq sequence.
if (ISD == ISD::SDIV && LT.second == MVT::v4i32 && ST->hasSSE41())
return LT.first * 15;
if (const auto *Entry = CostTableLookup(SSE2UniformConstCostTable, ISD,
LT.second))
return LT.first * Entry->Cost;
}
if (ISD == ISD::SHL &&
Op2Info == TargetTransformInfo::OK_NonUniformConstantValue) {
MVT VT = LT.second;
// Vector shift left by non uniform constant can be lowered
// into vector multiply (pmullw/pmulld).
if ((VT == MVT::v8i16 && ST->hasSSE2()) ||
(VT == MVT::v4i32 && ST->hasSSE41()))
return LT.first;
// v16i16 and v8i32 shifts by non-uniform constants are lowered into a
// sequence of extract + two vector multiply + insert.
if ((VT == MVT::v8i32 || VT == MVT::v16i16) &&
(ST->hasAVX() && !ST->hasAVX2()))
ISD = ISD::MUL;
// A vector shift left by non uniform constant is converted
// into a vector multiply; the new multiply is eventually
// lowered into a sequence of shuffles and 2 x pmuludq.
if (VT == MVT::v4i32 && ST->hasSSE2())
ISD = ISD::MUL;
}
static const CostTblEntry SSE2CostTable[] = {
// We don't correctly identify costs of casts because they are marked as
// custom.
// For some cases, where the shift amount is a scalar we would be able
// to generate better code. Unfortunately, when this is the case the value
// (the splat) will get hoisted out of the loop, thereby making it invisible
// to ISel. The cost model must return worst case assumptions because it is
// used for vectorization and we don't want to make vectorized code worse
// than scalar code.
{ ISD::SHL, MVT::v16i8, 26 }, // cmpgtb sequence.
{ ISD::SHL, MVT::v32i8, 2*26 }, // cmpgtb sequence.
{ ISD::SHL, MVT::v8i16, 32 }, // cmpgtb sequence.
{ ISD::SHL, MVT::v16i16, 2*32 }, // cmpgtb sequence.
{ ISD::SHL, MVT::v4i32, 2*5 }, // We optimized this using mul.
{ ISD::SHL, MVT::v8i32, 2*2*5 }, // We optimized this using mul.
{ ISD::SHL, MVT::v2i64, 4 }, // splat+shuffle sequence.
{ ISD::SHL, MVT::v4i64, 2*4 }, // splat+shuffle sequence.
{ ISD::SRL, MVT::v16i8, 26 }, // cmpgtb sequence.
{ ISD::SRL, MVT::v32i8, 2*26 }, // cmpgtb sequence.
{ ISD::SRL, MVT::v8i16, 32 }, // cmpgtb sequence.
{ ISD::SRL, MVT::v16i16, 2*32 }, // cmpgtb sequence.
{ ISD::SRL, MVT::v4i32, 16 }, // Shift each lane + blend.
{ ISD::SRL, MVT::v8i32, 2*16 }, // Shift each lane + blend.
{ ISD::SRL, MVT::v2i64, 4 }, // splat+shuffle sequence.
{ ISD::SRL, MVT::v4i64, 2*4 }, // splat+shuffle sequence.
{ ISD::SRA, MVT::v16i8, 54 }, // unpacked cmpgtb sequence.
{ ISD::SRA, MVT::v32i8, 2*54 }, // unpacked cmpgtb sequence.
{ ISD::SRA, MVT::v8i16, 32 }, // cmpgtb sequence.
{ ISD::SRA, MVT::v16i16, 2*32 }, // cmpgtb sequence.
{ ISD::SRA, MVT::v4i32, 16 }, // Shift each lane + blend.
{ ISD::SRA, MVT::v8i32, 2*16 }, // Shift each lane + blend.
{ ISD::SRA, MVT::v2i64, 12 }, // srl/xor/sub sequence.
{ ISD::SRA, MVT::v4i64, 2*12 }, // srl/xor/sub sequence.
// It is not a good idea to vectorize division. We have to scalarize it and
// in the process we will often end up having to spilling regular
// registers. The overhead of division is going to dominate most kernels
// anyways so try hard to prevent vectorization of division - it is
// generally a bad idea. Assume somewhat arbitrarily that we have to be able
// to hide "20 cycles" for each lane.
{ ISD::SDIV, MVT::v16i8, 16*20 },
{ ISD::SDIV, MVT::v8i16, 8*20 },
{ ISD::SDIV, MVT::v4i32, 4*20 },
{ ISD::SDIV, MVT::v2i64, 2*20 },
{ ISD::UDIV, MVT::v16i8, 16*20 },
{ ISD::UDIV, MVT::v8i16, 8*20 },
{ ISD::UDIV, MVT::v4i32, 4*20 },
{ ISD::UDIV, MVT::v2i64, 2*20 },
};
if (ST->hasSSE2()) {
if (const auto *Entry = CostTableLookup(SSE2CostTable, ISD, LT.second))
return LT.first * Entry->Cost;
}
static const CostTblEntry AVX1CostTable[] = {
// We don't have to scalarize unsupported ops. We can issue two half-sized
// operations and we only need to extract the upper YMM half.
// Two ops + 1 extract + 1 insert = 4.
{ ISD::MUL, MVT::v16i16, 4 },
{ ISD::MUL, MVT::v8i32, 4 },
{ ISD::SUB, MVT::v8i32, 4 },
{ ISD::ADD, MVT::v8i32, 4 },
{ ISD::SUB, MVT::v4i64, 4 },
{ ISD::ADD, MVT::v4i64, 4 },
// A v4i64 multiply is custom lowered as two split v2i64 vectors that then
// are lowered as a series of long multiplies(3), shifts(4) and adds(2)
// Because we believe v4i64 to be a legal type, we must also include the
// split factor of two in the cost table. Therefore, the cost here is 18
// instead of 9.
{ ISD::MUL, MVT::v4i64, 18 },
};
// Look for AVX1 lowering tricks.
if (ST->hasAVX() && !ST->hasAVX2()) {
MVT VT = LT.second;
if (const auto *Entry = CostTableLookup(AVX1CostTable, ISD, VT))
return LT.first * Entry->Cost;
}
// Custom lowering of vectors.
static const CostTblEntry CustomLowered[] = {
// A v2i64/v4i64 and multiply is custom lowered as a series of long
// multiplies(3), shifts(4) and adds(2).
{ ISD::MUL, MVT::v2i64, 9 },
{ ISD::MUL, MVT::v4i64, 9 },
};
if (const auto *Entry = CostTableLookup(CustomLowered, ISD, LT.second))
return LT.first * Entry->Cost;
// Special lowering of v4i32 mul on sse2, sse3: Lower v4i32 mul as 2x shuffle,
// 2x pmuludq, 2x shuffle.
if (ISD == ISD::MUL && LT.second == MVT::v4i32 && ST->hasSSE2() &&
!ST->hasSSE41())
return LT.first * 6;
// Fallback to the default implementation.
return BaseT::getArithmeticInstrCost(Opcode, Ty, Op1Info, Op2Info);
}
int X86TTIImpl::getShuffleCost(TTI::ShuffleKind Kind, Type *Tp, int Index,
Type *SubTp) {
// We only estimate the cost of reverse and alternate shuffles.
if (Kind != TTI::SK_Reverse && Kind != TTI::SK_Alternate)
return BaseT::getShuffleCost(Kind, Tp, Index, SubTp);
if (Kind == TTI::SK_Reverse) {
std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, Tp);
int Cost = 1;
if (LT.second.getSizeInBits() > 128)
Cost = 3; // Extract + insert + copy.
// Multiple by the number of parts.
return Cost * LT.first;
}
if (Kind == TTI::SK_Alternate) {
// 64-bit packed float vectors (v2f32) are widened to type v4f32.
// 64-bit packed integer vectors (v2i32) are promoted to type v2i64.
std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, Tp);
// The backend knows how to generate a single VEX.256 version of
// instruction VPBLENDW if the target supports AVX2.
if (ST->hasAVX2() && LT.second == MVT::v16i16)
return LT.first;
static const CostTblEntry AVXAltShuffleTbl[] = {
{ISD::VECTOR_SHUFFLE, MVT::v4i64, 1}, // vblendpd
{ISD::VECTOR_SHUFFLE, MVT::v4f64, 1}, // vblendpd
{ISD::VECTOR_SHUFFLE, MVT::v8i32, 1}, // vblendps
{ISD::VECTOR_SHUFFLE, MVT::v8f32, 1}, // vblendps
// This shuffle is custom lowered into a sequence of:
// 2x vextractf128 , 2x vpblendw , 1x vinsertf128
{ISD::VECTOR_SHUFFLE, MVT::v16i16, 5},
// This shuffle is custom lowered into a long sequence of:
// 2x vextractf128 , 4x vpshufb , 2x vpor , 1x vinsertf128
{ISD::VECTOR_SHUFFLE, MVT::v32i8, 9}
};
if (ST->hasAVX())
if (const auto *Entry = CostTableLookup(AVXAltShuffleTbl,
ISD::VECTOR_SHUFFLE, LT.second))
return LT.first * Entry->Cost;
static const CostTblEntry SSE41AltShuffleTbl[] = {
// These are lowered into movsd.
{ISD::VECTOR_SHUFFLE, MVT::v2i64, 1},
{ISD::VECTOR_SHUFFLE, MVT::v2f64, 1},
// packed float vectors with four elements are lowered into BLENDI dag
// nodes. A v4i32/v4f32 BLENDI generates a single 'blendps'/'blendpd'.
{ISD::VECTOR_SHUFFLE, MVT::v4i32, 1},
{ISD::VECTOR_SHUFFLE, MVT::v4f32, 1},
// This shuffle generates a single pshufw.
{ISD::VECTOR_SHUFFLE, MVT::v8i16, 1},
// There is no instruction that matches a v16i8 alternate shuffle.
// The backend will expand it into the sequence 'pshufb + pshufb + or'.
{ISD::VECTOR_SHUFFLE, MVT::v16i8, 3}
};
if (ST->hasSSE41())
if (const auto *Entry = CostTableLookup(SSE41AltShuffleTbl, ISD::VECTOR_SHUFFLE,
LT.second))
return LT.first * Entry->Cost;
static const CostTblEntry SSSE3AltShuffleTbl[] = {
{ISD::VECTOR_SHUFFLE, MVT::v2i64, 1}, // movsd
{ISD::VECTOR_SHUFFLE, MVT::v2f64, 1}, // movsd
// SSE3 doesn't have 'blendps'. The following shuffles are expanded into
// the sequence 'shufps + pshufd'
{ISD::VECTOR_SHUFFLE, MVT::v4i32, 2},
{ISD::VECTOR_SHUFFLE, MVT::v4f32, 2},
{ISD::VECTOR_SHUFFLE, MVT::v8i16, 3}, // pshufb + pshufb + or
{ISD::VECTOR_SHUFFLE, MVT::v16i8, 3} // pshufb + pshufb + or
};
if (ST->hasSSSE3())
if (const auto *Entry = CostTableLookup(SSSE3AltShuffleTbl,
ISD::VECTOR_SHUFFLE, LT.second))
return LT.first * Entry->Cost;
static const CostTblEntry SSEAltShuffleTbl[] = {
{ISD::VECTOR_SHUFFLE, MVT::v2i64, 1}, // movsd
{ISD::VECTOR_SHUFFLE, MVT::v2f64, 1}, // movsd
{ISD::VECTOR_SHUFFLE, MVT::v4i32, 2}, // shufps + pshufd
{ISD::VECTOR_SHUFFLE, MVT::v4f32, 2}, // shufps + pshufd
// This is expanded into a long sequence of four extract + four insert.
{ISD::VECTOR_SHUFFLE, MVT::v8i16, 8}, // 4 x pextrw + 4 pinsrw.
// 8 x (pinsrw + pextrw + and + movb + movzb + or)
{ISD::VECTOR_SHUFFLE, MVT::v16i8, 48}
};
// Fall-back (SSE3 and SSE2).
if (const auto *Entry = CostTableLookup(SSEAltShuffleTbl,
ISD::VECTOR_SHUFFLE, LT.second))
return LT.first * Entry->Cost;
return BaseT::getShuffleCost(Kind, Tp, Index, SubTp);
}
return BaseT::getShuffleCost(Kind, Tp, Index, SubTp);
}
int X86TTIImpl::getCastInstrCost(unsigned Opcode, Type *Dst, Type *Src) {
int ISD = TLI->InstructionOpcodeToISD(Opcode);
assert(ISD && "Invalid opcode");
// FIXME: Need a better design of the cost table to handle non-simple types of
// potential massive combinations (elem_num x src_type x dst_type).
static const TypeConversionCostTblEntry AVX512DQConversionTbl[] = {
{ ISD::UINT_TO_FP, MVT::v2f32, MVT::v2i64, 1 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v2i64, 1 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i64, 1 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i64, 1 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i64, 1 },
{ ISD::UINT_TO_FP, MVT::v8f64, MVT::v8i64, 1 },
{ ISD::FP_TO_UINT, MVT::v2i64, MVT::v2f32, 1 },
{ ISD::FP_TO_UINT, MVT::v4i64, MVT::v4f32, 1 },
{ ISD::FP_TO_UINT, MVT::v8i64, MVT::v8f32, 1 },
{ ISD::FP_TO_UINT, MVT::v2i64, MVT::v2f64, 1 },
{ ISD::FP_TO_UINT, MVT::v4i64, MVT::v4f64, 1 },
{ ISD::FP_TO_UINT, MVT::v8i64, MVT::v8f64, 1 },
};
// TODO: For AVX512DQ + AVX512VL, we also have cheap casts for 128-bit and
// 256-bit wide vectors.
static const TypeConversionCostTblEntry AVX512FConversionTbl[] = {
{ ISD::FP_EXTEND, MVT::v8f64, MVT::v8f32, 1 },
{ ISD::FP_EXTEND, MVT::v8f64, MVT::v16f32, 3 },
{ ISD::FP_ROUND, MVT::v8f32, MVT::v8f64, 1 },
{ ISD::TRUNCATE, MVT::v16i8, MVT::v16i32, 1 },
{ ISD::TRUNCATE, MVT::v16i16, MVT::v16i32, 1 },
{ ISD::TRUNCATE, MVT::v8i16, MVT::v8i64, 1 },
{ ISD::TRUNCATE, MVT::v8i32, MVT::v8i64, 1 },
// v16i1 -> v16i32 - load + broadcast
{ ISD::SIGN_EXTEND, MVT::v16i32, MVT::v16i1, 2 },
{ ISD::ZERO_EXTEND, MVT::v16i32, MVT::v16i1, 2 },
{ ISD::SIGN_EXTEND, MVT::v16i32, MVT::v16i8, 1 },
{ ISD::ZERO_EXTEND, MVT::v16i32, MVT::v16i8, 1 },
{ ISD::SIGN_EXTEND, MVT::v16i32, MVT::v16i16, 1 },
{ ISD::ZERO_EXTEND, MVT::v16i32, MVT::v16i16, 1 },
{ ISD::ZERO_EXTEND, MVT::v8i64, MVT::v8i16, 1 },
{ ISD::SIGN_EXTEND, MVT::v8i64, MVT::v8i16, 1 },
{ ISD::SIGN_EXTEND, MVT::v8i64, MVT::v8i32, 1 },
{ ISD::ZERO_EXTEND, MVT::v8i64, MVT::v8i32, 1 },
{ ISD::SINT_TO_FP, MVT::v8f64, MVT::v8i1, 4 },
{ ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i1, 3 },
{ ISD::SINT_TO_FP, MVT::v8f64, MVT::v8i8, 2 },
{ ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i8, 2 },
{ ISD::SINT_TO_FP, MVT::v8f64, MVT::v8i16, 2 },
{ ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i16, 2 },
{ ISD::SINT_TO_FP, MVT::v16f32, MVT::v16i32, 1 },
{ ISD::SINT_TO_FP, MVT::v8f64, MVT::v8i32, 1 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i64, 26 },
{ ISD::UINT_TO_FP, MVT::v8f64, MVT::v8i64, 26 },
{ ISD::UINT_TO_FP, MVT::v8f64, MVT::v8i1, 4 },
{ ISD::UINT_TO_FP, MVT::v16f32, MVT::v16i1, 3 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v2i8, 2 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i8, 2 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i8, 2 },
{ ISD::UINT_TO_FP, MVT::v8f64, MVT::v8i8, 2 },
{ ISD::UINT_TO_FP, MVT::v16f32, MVT::v16i8, 2 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v2i16, 5 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i16, 2 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i16, 2 },
{ ISD::UINT_TO_FP, MVT::v8f64, MVT::v8i16, 2 },
{ ISD::UINT_TO_FP, MVT::v16f32, MVT::v16i16, 2 },
{ ISD::UINT_TO_FP, MVT::v2f32, MVT::v2i32, 2 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v2i32, 1 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i32, 1 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i32, 1 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i32, 1 },
{ ISD::UINT_TO_FP, MVT::v8f64, MVT::v8i32, 1 },
{ ISD::UINT_TO_FP, MVT::v16f32, MVT::v16i32, 1 },
{ ISD::UINT_TO_FP, MVT::v2f32, MVT::v2i64, 5 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v2i64, 5 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i64, 12 },
{ ISD::UINT_TO_FP, MVT::v8f64, MVT::v8i64, 26 },
{ ISD::FP_TO_UINT, MVT::v2i32, MVT::v2f32, 1 },
{ ISD::FP_TO_UINT, MVT::v4i32, MVT::v4f32, 1 },
{ ISD::FP_TO_UINT, MVT::v8i32, MVT::v8f32, 1 },
{ ISD::FP_TO_UINT, MVT::v16i32, MVT::v16f32, 1 },
};
static const TypeConversionCostTblEntry AVX2ConversionTbl[] = {
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i1, 3 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i1, 3 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i1, 3 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i1, 3 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i8, 3 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i8, 3 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i8, 3 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i8, 3 },
{ ISD::SIGN_EXTEND, MVT::v16i16, MVT::v16i8, 1 },
{ ISD::ZERO_EXTEND, MVT::v16i16, MVT::v16i8, 1 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i16, 3 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i16, 3 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i16, 1 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i16, 1 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i32, 1 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i32, 1 },
{ ISD::TRUNCATE, MVT::v4i8, MVT::v4i64, 2 },
{ ISD::TRUNCATE, MVT::v4i16, MVT::v4i64, 2 },
{ ISD::TRUNCATE, MVT::v4i32, MVT::v4i64, 2 },
{ ISD::TRUNCATE, MVT::v8i8, MVT::v8i32, 2 },
{ ISD::TRUNCATE, MVT::v8i16, MVT::v8i32, 2 },
{ ISD::TRUNCATE, MVT::v8i32, MVT::v8i64, 4 },
{ ISD::FP_EXTEND, MVT::v8f64, MVT::v8f32, 3 },
{ ISD::FP_ROUND, MVT::v8f32, MVT::v8f64, 3 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i32, 8 },
};
static const TypeConversionCostTblEntry AVXConversionTbl[] = {
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i1, 6 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i1, 4 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i1, 7 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i1, 4 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i8, 6 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i8, 4 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i8, 7 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i8, 4 },
{ ISD::SIGN_EXTEND, MVT::v16i16, MVT::v16i8, 4 },
{ ISD::ZERO_EXTEND, MVT::v16i16, MVT::v16i8, 4 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i16, 6 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i16, 3 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i16, 4 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i16, 4 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i32, 4 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i32, 4 },
{ ISD::TRUNCATE, MVT::v16i8, MVT::v16i16, 4 },
{ ISD::TRUNCATE, MVT::v8i8, MVT::v8i32, 4 },
{ ISD::TRUNCATE, MVT::v8i16, MVT::v8i32, 5 },
{ ISD::TRUNCATE, MVT::v4i8, MVT::v4i64, 4 },
{ ISD::TRUNCATE, MVT::v4i16, MVT::v4i64, 4 },
{ ISD::TRUNCATE, MVT::v4i32, MVT::v4i64, 4 },
{ ISD::TRUNCATE, MVT::v8i32, MVT::v8i64, 9 },
{ ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i1, 3 },
{ ISD::SINT_TO_FP, MVT::v4f64, MVT::v4i1, 3 },
{ ISD::SINT_TO_FP, MVT::v8f32, MVT::v8i1, 8 },
{ ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i8, 3 },
{ ISD::SINT_TO_FP, MVT::v4f64, MVT::v4i8, 3 },
{ ISD::SINT_TO_FP, MVT::v8f32, MVT::v8i8, 8 },
{ ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i16, 3 },
{ ISD::SINT_TO_FP, MVT::v4f64, MVT::v4i16, 3 },
{ ISD::SINT_TO_FP, MVT::v8f32, MVT::v8i16, 5 },
{ ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i32, 1 },
{ ISD::SINT_TO_FP, MVT::v4f64, MVT::v4i32, 1 },
{ ISD::SINT_TO_FP, MVT::v8f32, MVT::v8i32, 1 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i1, 7 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i1, 7 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i1, 6 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i8, 2 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i8, 2 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i8, 5 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i16, 2 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i16, 2 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i16, 5 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v2i32, 6 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i32, 6 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i32, 6 },
{ ISD::UINT_TO_FP, MVT::v8f32, MVT::v8i32, 9 },
// The generic code to compute the scalar overhead is currently broken.
// Workaround this limitation by estimating the scalarization overhead
// here. We have roughly 10 instructions per scalar element.
// Multiply that by the vector width.
// FIXME: remove that when PR19268 is fixed.
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v2i64, 10 },
{ ISD::UINT_TO_FP, MVT::v4f64, MVT::v4i64, 20 },
{ ISD::SINT_TO_FP, MVT::v4f64, MVT::v4i64, 13 },
{ ISD::SINT_TO_FP, MVT::v4f64, MVT::v4i64, 13 },
{ ISD::FP_TO_SINT, MVT::v4i8, MVT::v4f32, 1 },
{ ISD::FP_TO_SINT, MVT::v8i8, MVT::v8f32, 7 },
// This node is expanded into scalarized operations but BasicTTI is overly
// optimistic estimating its cost. It computes 3 per element (one
// vector-extract, one scalar conversion and one vector-insert). The
// problem is that the inserts form a read-modify-write chain so latency
// should be factored in too. Inflating the cost per element by 1.
{ ISD::FP_TO_UINT, MVT::v8i32, MVT::v8f32, 8*4 },
{ ISD::FP_TO_UINT, MVT::v4i32, MVT::v4f64, 4*4 },
{ ISD::FP_EXTEND, MVT::v4f64, MVT::v4f32, 1 },
{ ISD::FP_ROUND, MVT::v4f32, MVT::v4f64, 1 },
};
static const TypeConversionCostTblEntry SSE41ConversionTbl[] = {
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i8, 2 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i8, 2 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i16, 2 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i16, 2 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i32, 2 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i32, 2 },
{ ISD::ZERO_EXTEND, MVT::v4i16, MVT::v4i8, 1 },
{ ISD::SIGN_EXTEND, MVT::v4i16, MVT::v4i8, 2 },
{ ISD::ZERO_EXTEND, MVT::v4i32, MVT::v4i8, 1 },
{ ISD::SIGN_EXTEND, MVT::v4i32, MVT::v4i8, 1 },
{ ISD::ZERO_EXTEND, MVT::v8i16, MVT::v8i8, 1 },
{ ISD::SIGN_EXTEND, MVT::v8i16, MVT::v8i8, 1 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i8, 2 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i8, 2 },
{ ISD::ZERO_EXTEND, MVT::v16i16, MVT::v16i8, 2 },
{ ISD::SIGN_EXTEND, MVT::v16i16, MVT::v16i8, 2 },
{ ISD::ZERO_EXTEND, MVT::v16i32, MVT::v16i8, 4 },
{ ISD::SIGN_EXTEND, MVT::v16i32, MVT::v16i8, 4 },
{ ISD::ZERO_EXTEND, MVT::v4i32, MVT::v4i16, 1 },
{ ISD::SIGN_EXTEND, MVT::v4i32, MVT::v4i16, 1 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i16, 2 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i16, 2 },
{ ISD::ZERO_EXTEND, MVT::v16i32, MVT::v16i16, 4 },
{ ISD::SIGN_EXTEND, MVT::v16i32, MVT::v16i16, 4 },
{ ISD::TRUNCATE, MVT::v4i8, MVT::v4i16, 2 },
{ ISD::TRUNCATE, MVT::v8i8, MVT::v8i16, 1 },
{ ISD::TRUNCATE, MVT::v4i8, MVT::v4i32, 1 },
{ ISD::TRUNCATE, MVT::v4i16, MVT::v4i32, 1 },
{ ISD::TRUNCATE, MVT::v8i8, MVT::v8i32, 3 },
{ ISD::TRUNCATE, MVT::v8i16, MVT::v8i32, 3 },
{ ISD::TRUNCATE, MVT::v16i16, MVT::v16i32, 6 },
};
static const TypeConversionCostTblEntry SSE2ConversionTbl[] = {
// These are somewhat magic numbers justified by looking at the output of
// Intel's IACA, running some kernels and making sure when we take
// legalization into account the throughput will be overestimated.
{ ISD::SINT_TO_FP, MVT::v4f32, MVT::v16i8, 8 },
{ ISD::SINT_TO_FP, MVT::v2f64, MVT::v16i8, 16*10 },
{ ISD::SINT_TO_FP, MVT::v4f32, MVT::v8i16, 15 },
{ ISD::SINT_TO_FP, MVT::v2f64, MVT::v8i16, 8*10 },
{ ISD::SINT_TO_FP, MVT::v4f32, MVT::v4i32, 5 },
{ ISD::SINT_TO_FP, MVT::v2f64, MVT::v4i32, 4*10 },
{ ISD::SINT_TO_FP, MVT::v4f32, MVT::v2i64, 15 },
{ ISD::SINT_TO_FP, MVT::v2f64, MVT::v2i64, 2*10 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v16i8, 16*10 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v16i8, 8 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v8i16, 15 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v8i16, 8*10 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v4i32, 4*10 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v4i32, 8 },
{ ISD::UINT_TO_FP, MVT::v2f64, MVT::v2i64, 2*10 },
{ ISD::UINT_TO_FP, MVT::v4f32, MVT::v2i64, 15 },
{ ISD::ZERO_EXTEND, MVT::v4i16, MVT::v4i8, 1 },
{ ISD::SIGN_EXTEND, MVT::v4i16, MVT::v4i8, 6 },
{ ISD::ZERO_EXTEND, MVT::v4i32, MVT::v4i8, 2 },
{ ISD::SIGN_EXTEND, MVT::v4i32, MVT::v4i8, 3 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i8, 4 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i8, 8 },
{ ISD::ZERO_EXTEND, MVT::v8i16, MVT::v8i8, 1 },
{ ISD::SIGN_EXTEND, MVT::v8i16, MVT::v8i8, 2 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i8, 6 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i8, 6 },
{ ISD::ZERO_EXTEND, MVT::v16i16, MVT::v16i8, 3 },
{ ISD::SIGN_EXTEND, MVT::v16i16, MVT::v16i8, 4 },
{ ISD::ZERO_EXTEND, MVT::v16i32, MVT::v16i8, 9 },
{ ISD::SIGN_EXTEND, MVT::v16i32, MVT::v16i8, 12 },
{ ISD::ZERO_EXTEND, MVT::v4i32, MVT::v4i16, 1 },
{ ISD::SIGN_EXTEND, MVT::v4i32, MVT::v4i16, 2 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i16, 3 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i16, 10 },
{ ISD::ZERO_EXTEND, MVT::v8i32, MVT::v8i16, 3 },
{ ISD::SIGN_EXTEND, MVT::v8i32, MVT::v8i16, 4 },
{ ISD::ZERO_EXTEND, MVT::v16i32, MVT::v16i16, 6 },
{ ISD::SIGN_EXTEND, MVT::v16i32, MVT::v16i16, 8 },
{ ISD::ZERO_EXTEND, MVT::v4i64, MVT::v4i32, 3 },
{ ISD::SIGN_EXTEND, MVT::v4i64, MVT::v4i32, 5 },
{ ISD::TRUNCATE, MVT::v4i8, MVT::v4i16, 4 },
{ ISD::TRUNCATE, MVT::v8i8, MVT::v8i16, 2 },
{ ISD::TRUNCATE, MVT::v16i8, MVT::v16i16, 3 },
{ ISD::TRUNCATE, MVT::v4i8, MVT::v4i32, 3 },
{ ISD::TRUNCATE, MVT::v4i16, MVT::v4i32, 3 },
{ ISD::TRUNCATE, MVT::v8i8, MVT::v8i32, 4 },
{ ISD::TRUNCATE, MVT::v16i8, MVT::v16i32, 7 },
{ ISD::TRUNCATE, MVT::v8i16, MVT::v8i32, 5 },
{ ISD::TRUNCATE, MVT::v16i16, MVT::v16i32, 10 },
};
std::pair<int, MVT> LTSrc = TLI->getTypeLegalizationCost(DL, Src);
std::pair<int, MVT> LTDest = TLI->getTypeLegalizationCost(DL, Dst);
if (ST->hasSSE2() && !ST->hasAVX()) {
if (const auto *Entry = ConvertCostTableLookup(SSE2ConversionTbl, ISD,
LTDest.second, LTSrc.second))
return LTSrc.first * Entry->Cost;
}
EVT SrcTy = TLI->getValueType(DL, Src);
EVT DstTy = TLI->getValueType(DL, Dst);
// The function getSimpleVT only handles simple value types.
if (!SrcTy.isSimple() || !DstTy.isSimple())
return BaseT::getCastInstrCost(Opcode, Dst, Src);
if (ST->hasDQI())
if (const auto *Entry = ConvertCostTableLookup(AVX512DQConversionTbl, ISD,
DstTy.getSimpleVT(),
SrcTy.getSimpleVT()))
return Entry->Cost;
if (ST->hasAVX512())
if (const auto *Entry = ConvertCostTableLookup(AVX512FConversionTbl, ISD,
DstTy.getSimpleVT(),
SrcTy.getSimpleVT()))
return Entry->Cost;
if (ST->hasAVX2()) {
if (const auto *Entry = ConvertCostTableLookup(AVX2ConversionTbl, ISD,
DstTy.getSimpleVT(),
SrcTy.getSimpleVT()))
return Entry->Cost;
}
if (ST->hasAVX()) {
if (const auto *Entry = ConvertCostTableLookup(AVXConversionTbl, ISD,
DstTy.getSimpleVT(),
SrcTy.getSimpleVT()))
return Entry->Cost;
}
if (ST->hasSSE41()) {
if (const auto *Entry = ConvertCostTableLookup(SSE41ConversionTbl, ISD,
DstTy.getSimpleVT(),
SrcTy.getSimpleVT()))
return Entry->Cost;
}
if (ST->hasSSE2()) {
if (const auto *Entry = ConvertCostTableLookup(SSE2ConversionTbl, ISD,
DstTy.getSimpleVT(),
SrcTy.getSimpleVT()))
return Entry->Cost;
}
return BaseT::getCastInstrCost(Opcode, Dst, Src);
}
int X86TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy, Type *CondTy) {
// Legalize the type.
std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, ValTy);
MVT MTy = LT.second;
int ISD = TLI->InstructionOpcodeToISD(Opcode);
assert(ISD && "Invalid opcode");
static const CostTblEntry SSE2CostTbl[] = {
{ ISD::SETCC, MVT::v2i64, 8 },
{ ISD::SETCC, MVT::v4i32, 1 },
{ ISD::SETCC, MVT::v8i16, 1 },
{ ISD::SETCC, MVT::v16i8, 1 },
};
static const CostTblEntry SSE42CostTbl[] = {
{ ISD::SETCC, MVT::v2f64, 1 },
{ ISD::SETCC, MVT::v4f32, 1 },
{ ISD::SETCC, MVT::v2i64, 1 },
};
static const CostTblEntry AVX1CostTbl[] = {
{ ISD::SETCC, MVT::v4f64, 1 },
{ ISD::SETCC, MVT::v8f32, 1 },
// AVX1 does not support 8-wide integer compare.
{ ISD::SETCC, MVT::v4i64, 4 },
{ ISD::SETCC, MVT::v8i32, 4 },
{ ISD::SETCC, MVT::v16i16, 4 },
{ ISD::SETCC, MVT::v32i8, 4 },
};
static const CostTblEntry AVX2CostTbl[] = {
{ ISD::SETCC, MVT::v4i64, 1 },
{ ISD::SETCC, MVT::v8i32, 1 },
{ ISD::SETCC, MVT::v16i16, 1 },
{ ISD::SETCC, MVT::v32i8, 1 },
};
static const CostTblEntry AVX512CostTbl[] = {
{ ISD::SETCC, MVT::v8i64, 1 },
{ ISD::SETCC, MVT::v16i32, 1 },
{ ISD::SETCC, MVT::v8f64, 1 },
{ ISD::SETCC, MVT::v16f32, 1 },
};
if (ST->hasAVX512())
if (const auto *Entry = CostTableLookup(AVX512CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasAVX2())
if (const auto *Entry = CostTableLookup(AVX2CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasAVX())
if (const auto *Entry = CostTableLookup(AVX1CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasSSE42())
if (const auto *Entry = CostTableLookup(SSE42CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasSSE2())
if (const auto *Entry = CostTableLookup(SSE2CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
return BaseT::getCmpSelInstrCost(Opcode, ValTy, CondTy);
}
int X86TTIImpl::getIntrinsicInstrCost(Intrinsic::ID IID, Type *RetTy,
ArrayRef<Type *> Tys, FastMathFlags FMF) {
static const CostTblEntry XOPCostTbl[] = {
{ ISD::BITREVERSE, MVT::v4i64, 4 },
{ ISD::BITREVERSE, MVT::v8i32, 4 },
{ ISD::BITREVERSE, MVT::v16i16, 4 },
{ ISD::BITREVERSE, MVT::v32i8, 4 },
{ ISD::BITREVERSE, MVT::v2i64, 1 },
{ ISD::BITREVERSE, MVT::v4i32, 1 },
{ ISD::BITREVERSE, MVT::v8i16, 1 },
{ ISD::BITREVERSE, MVT::v16i8, 1 },
{ ISD::BITREVERSE, MVT::i64, 3 },
{ ISD::BITREVERSE, MVT::i32, 3 },
{ ISD::BITREVERSE, MVT::i16, 3 },
{ ISD::BITREVERSE, MVT::i8, 3 }
};
static const CostTblEntry AVX2CostTbl[] = {
{ ISD::BITREVERSE, MVT::v4i64, 5 },
{ ISD::BITREVERSE, MVT::v8i32, 5 },
{ ISD::BITREVERSE, MVT::v16i16, 5 },
{ ISD::BITREVERSE, MVT::v32i8, 5 },
{ ISD::BSWAP, MVT::v4i64, 1 },
{ ISD::BSWAP, MVT::v8i32, 1 },
{ ISD::BSWAP, MVT::v16i16, 1 }
};
static const CostTblEntry AVX1CostTbl[] = {
{ ISD::BITREVERSE, MVT::v4i64, 10 },
{ ISD::BITREVERSE, MVT::v8i32, 10 },
{ ISD::BITREVERSE, MVT::v16i16, 10 },
{ ISD::BITREVERSE, MVT::v32i8, 10 },
{ ISD::BSWAP, MVT::v4i64, 4 },
{ ISD::BSWAP, MVT::v8i32, 4 },
{ ISD::BSWAP, MVT::v16i16, 4 }
};
static const CostTblEntry SSSE3CostTbl[] = {
{ ISD::BITREVERSE, MVT::v2i64, 5 },
{ ISD::BITREVERSE, MVT::v4i32, 5 },
{ ISD::BITREVERSE, MVT::v8i16, 5 },
{ ISD::BITREVERSE, MVT::v16i8, 5 },
{ ISD::BSWAP, MVT::v2i64, 1 },
{ ISD::BSWAP, MVT::v4i32, 1 },
{ ISD::BSWAP, MVT::v8i16, 1 }
};
static const CostTblEntry SSE2CostTbl[] = {
{ ISD::BSWAP, MVT::v2i64, 7 },
{ ISD::BSWAP, MVT::v4i32, 7 },
{ ISD::BSWAP, MVT::v8i16, 7 }
};
unsigned ISD = ISD::DELETED_NODE;
switch (IID) {
default:
break;
case Intrinsic::bitreverse:
ISD = ISD::BITREVERSE;
break;
case Intrinsic::bswap:
ISD = ISD::BSWAP;
break;
}
// Legalize the type.
std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, RetTy);
MVT MTy = LT.second;
// Attempt to lookup cost.
if (ST->hasXOP())
if (const auto *Entry = CostTableLookup(XOPCostTbl, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasAVX2())
if (const auto *Entry = CostTableLookup(AVX2CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasAVX())
if (const auto *Entry = CostTableLookup(AVX1CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasSSSE3())
if (const auto *Entry = CostTableLookup(SSSE3CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasSSE2())
if (const auto *Entry = CostTableLookup(SSE2CostTbl, ISD, MTy))
return LT.first * Entry->Cost;
return BaseT::getIntrinsicInstrCost(IID, RetTy, Tys, FMF);
}
int X86TTIImpl::getIntrinsicInstrCost(Intrinsic::ID IID, Type *RetTy,
ArrayRef<Value *> Args, FastMathFlags FMF) {
return BaseT::getIntrinsicInstrCost(IID, RetTy, Args, FMF);
}
int X86TTIImpl::getVectorInstrCost(unsigned Opcode, Type *Val, unsigned Index) {
assert(Val->isVectorTy() && "This must be a vector type");
Type *ScalarType = Val->getScalarType();
if (Index != -1U) {
// Legalize the type.
std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, Val);
// This type is legalized to a scalar type.
if (!LT.second.isVector())
return 0;
// The type may be split. Normalize the index to the new type.
unsigned Width = LT.second.getVectorNumElements();
Index = Index % Width;
// Floating point scalars are already located in index #0.
if (ScalarType->isFloatingPointTy() && Index == 0)
return 0;
}
// Add to the base cost if we know that the extracted element of a vector is
// destined to be moved to and used in the integer register file.
int RegisterFileMoveCost = 0;
if (Opcode == Instruction::ExtractElement && ScalarType->isPointerTy())
RegisterFileMoveCost = 1;
return BaseT::getVectorInstrCost(Opcode, Val, Index) + RegisterFileMoveCost;
}
int X86TTIImpl::getScalarizationOverhead(Type *Ty, bool Insert, bool Extract) {
assert (Ty->isVectorTy() && "Can only scalarize vectors");
int Cost = 0;
for (int i = 0, e = Ty->getVectorNumElements(); i < e; ++i) {
if (Insert)
Cost += getVectorInstrCost(Instruction::InsertElement, Ty, i);
if (Extract)
Cost += getVectorInstrCost(Instruction::ExtractElement, Ty, i);
}
return Cost;
}
int X86TTIImpl::getMemoryOpCost(unsigned Opcode, Type *Src, unsigned Alignment,
unsigned AddressSpace) {
// Handle non-power-of-two vectors such as <3 x float>
if (VectorType *VTy = dyn_cast<VectorType>(Src)) {
unsigned NumElem = VTy->getVectorNumElements();
// Handle a few common cases:
// <3 x float>
if (NumElem == 3 && VTy->getScalarSizeInBits() == 32)
// Cost = 64 bit store + extract + 32 bit store.
return 3;
// <3 x double>
if (NumElem == 3 && VTy->getScalarSizeInBits() == 64)
// Cost = 128 bit store + unpack + 64 bit store.
return 3;
// Assume that all other non-power-of-two numbers are scalarized.
if (!isPowerOf2_32(NumElem)) {
int Cost = BaseT::getMemoryOpCost(Opcode, VTy->getScalarType(), Alignment,
AddressSpace);
int SplitCost = getScalarizationOverhead(Src, Opcode == Instruction::Load,
Opcode == Instruction::Store);
return NumElem * Cost + SplitCost;
}
}
// Legalize the type.
std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, Src);
assert((Opcode == Instruction::Load || Opcode == Instruction::Store) &&
"Invalid Opcode");
// Each load/store unit costs 1.
int Cost = LT.first * 1;
// This isn't exactly right. We're using slow unaligned 32-byte accesses as a
// proxy for a double-pumped AVX memory interface such as on Sandybridge.
if (LT.second.getStoreSize() == 32 && ST->isUnalignedMem32Slow())
Cost *= 2;
return Cost;
}
int X86TTIImpl::getMaskedMemoryOpCost(unsigned Opcode, Type *SrcTy,
unsigned Alignment,
unsigned AddressSpace) {
VectorType *SrcVTy = dyn_cast<VectorType>(SrcTy);
if (!SrcVTy)
// To calculate scalar take the regular cost, without mask
return getMemoryOpCost(Opcode, SrcTy, Alignment, AddressSpace);
unsigned NumElem = SrcVTy->getVectorNumElements();
VectorType *MaskTy =
VectorType::get(Type::getInt8Ty(SrcVTy->getContext()), NumElem);
if ((Opcode == Instruction::Load && !isLegalMaskedLoad(SrcVTy)) ||
(Opcode == Instruction::Store && !isLegalMaskedStore(SrcVTy)) ||
!isPowerOf2_32(NumElem)) {
// Scalarization
int MaskSplitCost = getScalarizationOverhead(MaskTy, false, true);
int ScalarCompareCost = getCmpSelInstrCost(
Instruction::ICmp, Type::getInt8Ty(SrcVTy->getContext()), nullptr);
int BranchCost = getCFInstrCost(Instruction::Br);
int MaskCmpCost = NumElem * (BranchCost + ScalarCompareCost);
int ValueSplitCost = getScalarizationOverhead(
SrcVTy, Opcode == Instruction::Load, Opcode == Instruction::Store);
int MemopCost =
NumElem * BaseT::getMemoryOpCost(Opcode, SrcVTy->getScalarType(),
Alignment, AddressSpace);
return MemopCost + ValueSplitCost + MaskSplitCost + MaskCmpCost;
}
// Legalize the type.
std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, SrcVTy);
auto VT = TLI->getValueType(DL, SrcVTy);
int Cost = 0;
if (VT.isSimple() && LT.second != VT.getSimpleVT() &&
LT.second.getVectorNumElements() == NumElem)
// Promotion requires expand/truncate for data and a shuffle for mask.
Cost += getShuffleCost(TTI::SK_Alternate, SrcVTy, 0, nullptr) +
getShuffleCost(TTI::SK_Alternate, MaskTy, 0, nullptr);
else if (LT.second.getVectorNumElements() > NumElem) {
VectorType *NewMaskTy = VectorType::get(MaskTy->getVectorElementType(),
LT.second.getVectorNumElements());
// Expanding requires fill mask with zeroes
Cost += getShuffleCost(TTI::SK_InsertSubvector, NewMaskTy, 0, MaskTy);
}
if (!ST->hasAVX512())
return Cost + LT.first*4; // Each maskmov costs 4
// AVX-512 masked load/store is cheapper
return Cost+LT.first;
}
int X86TTIImpl::getAddressComputationCost(Type *Ty, bool IsComplex) {
// Address computations in vectorized code with non-consecutive addresses will
// likely result in more instructions compared to scalar code where the
// computation can more often be merged into the index mode. The resulting
// extra micro-ops can significantly decrease throughput.
unsigned NumVectorInstToHideOverhead = 10;
if (Ty->isVectorTy() && IsComplex)
return NumVectorInstToHideOverhead;
return BaseT::getAddressComputationCost(Ty, IsComplex);
}
int X86TTIImpl::getReductionCost(unsigned Opcode, Type *ValTy,
bool IsPairwise) {
std::pair<int, MVT> LT = TLI->getTypeLegalizationCost(DL, ValTy);
MVT MTy = LT.second;
int ISD = TLI->InstructionOpcodeToISD(Opcode);
assert(ISD && "Invalid opcode");
// We use the Intel Architecture Code Analyzer(IACA) to measure the throughput
// and make it as the cost.
static const CostTblEntry SSE42CostTblPairWise[] = {
{ ISD::FADD, MVT::v2f64, 2 },
{ ISD::FADD, MVT::v4f32, 4 },
{ ISD::ADD, MVT::v2i64, 2 }, // The data reported by the IACA tool is "1.6".
{ ISD::ADD, MVT::v4i32, 3 }, // The data reported by the IACA tool is "3.5".
{ ISD::ADD, MVT::v8i16, 5 },
};
static const CostTblEntry AVX1CostTblPairWise[] = {
{ ISD::FADD, MVT::v4f32, 4 },
{ ISD::FADD, MVT::v4f64, 5 },
{ ISD::FADD, MVT::v8f32, 7 },
{ ISD::ADD, MVT::v2i64, 1 }, // The data reported by the IACA tool is "1.5".
{ ISD::ADD, MVT::v4i32, 3 }, // The data reported by the IACA tool is "3.5".
{ ISD::ADD, MVT::v4i64, 5 }, // The data reported by the IACA tool is "4.8".
{ ISD::ADD, MVT::v8i16, 5 },
{ ISD::ADD, MVT::v8i32, 5 },
};
static const CostTblEntry SSE42CostTblNoPairWise[] = {
{ ISD::FADD, MVT::v2f64, 2 },
{ ISD::FADD, MVT::v4f32, 4 },
{ ISD::ADD, MVT::v2i64, 2 }, // The data reported by the IACA tool is "1.6".
{ ISD::ADD, MVT::v4i32, 3 }, // The data reported by the IACA tool is "3.3".
{ ISD::ADD, MVT::v8i16, 4 }, // The data reported by the IACA tool is "4.3".
};
static const CostTblEntry AVX1CostTblNoPairWise[] = {
{ ISD::FADD, MVT::v4f32, 3 },
{ ISD::FADD, MVT::v4f64, 3 },
{ ISD::FADD, MVT::v8f32, 4 },
{ ISD::ADD, MVT::v2i64, 1 }, // The data reported by the IACA tool is "1.5".
{ ISD::ADD, MVT::v4i32, 3 }, // The data reported by the IACA tool is "2.8".
{ ISD::ADD, MVT::v4i64, 3 },
{ ISD::ADD, MVT::v8i16, 4 },
{ ISD::ADD, MVT::v8i32, 5 },
};
if (IsPairwise) {
if (ST->hasAVX())
if (const auto *Entry = CostTableLookup(AVX1CostTblPairWise, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasSSE42())
if (const auto *Entry = CostTableLookup(SSE42CostTblPairWise, ISD, MTy))
return LT.first * Entry->Cost;
} else {
if (ST->hasAVX())
if (const auto *Entry = CostTableLookup(AVX1CostTblNoPairWise, ISD, MTy))
return LT.first * Entry->Cost;
if (ST->hasSSE42())
if (const auto *Entry = CostTableLookup(SSE42CostTblNoPairWise, ISD, MTy))
return LT.first * Entry->Cost;
}
return BaseT::getReductionCost(Opcode, ValTy, IsPairwise);
}
/// \brief Calculate the cost of materializing a 64-bit value. This helper
/// method might only calculate a fraction of a larger immediate. Therefore it
/// is valid to return a cost of ZERO.
int X86TTIImpl::getIntImmCost(int64_t Val) {
if (Val == 0)
return TTI::TCC_Free;
if (isInt<32>(Val))
return TTI::TCC_Basic;
return 2 * TTI::TCC_Basic;
}
int X86TTIImpl::getIntImmCost(const APInt &Imm, Type *Ty) {
assert(Ty->isIntegerTy());
unsigned BitSize = Ty->getPrimitiveSizeInBits();
if (BitSize == 0)
return ~0U;
// Never hoist constants larger than 128bit, because this might lead to
// incorrect code generation or assertions in codegen.
// Fixme: Create a cost model for types larger than i128 once the codegen
// issues have been fixed.
if (BitSize > 128)
return TTI::TCC_Free;
if (Imm == 0)
return TTI::TCC_Free;
// Sign-extend all constants to a multiple of 64-bit.
APInt ImmVal = Imm;
if (BitSize & 0x3f)
ImmVal = Imm.sext((BitSize + 63) & ~0x3fU);
// Split the constant into 64-bit chunks and calculate the cost for each
// chunk.
int Cost = 0;
for (unsigned ShiftVal = 0; ShiftVal < BitSize; ShiftVal += 64) {
APInt Tmp = ImmVal.ashr(ShiftVal).sextOrTrunc(64);
int64_t Val = Tmp.getSExtValue();
Cost += getIntImmCost(Val);
}
// We need at least one instruction to materialize the constant.
return std::max(1, Cost);
}
int X86TTIImpl::getIntImmCost(unsigned Opcode, unsigned Idx, const APInt &Imm,
Type *Ty) {
assert(Ty->isIntegerTy());
unsigned BitSize = Ty->getPrimitiveSizeInBits();
// There is no cost model for constants with a bit size of 0. Return TCC_Free
// here, so that constant hoisting will ignore this constant.
if (BitSize == 0)
return TTI::TCC_Free;
unsigned ImmIdx = ~0U;
switch (Opcode) {
default:
return TTI::TCC_Free;
case Instruction::GetElementPtr:
// Always hoist the base address of a GetElementPtr. This prevents the
// creation of new constants for every base constant that gets constant
// folded with the offset.
if (Idx == 0)
return 2 * TTI::TCC_Basic;
return TTI::TCC_Free;
case Instruction::Store:
ImmIdx = 0;
break;
case Instruction::ICmp:
// This is an imperfect hack to prevent constant hoisting of
// compares that might be trying to check if a 64-bit value fits in
// 32-bits. The backend can optimize these cases using a right shift by 32.
// Ideally we would check the compare predicate here. There also other
// similar immediates the backend can use shifts for.
if (Idx == 1 && Imm.getBitWidth() == 64) {
uint64_t ImmVal = Imm.getZExtValue();
if (ImmVal == 0x100000000ULL || ImmVal == 0xffffffff)
return TTI::TCC_Free;
}
ImmIdx = 1;
break;
case Instruction::And:
// We support 64-bit ANDs with immediates with 32-bits of leading zeroes
// by using a 32-bit operation with implicit zero extension. Detect such
// immediates here as the normal path expects bit 31 to be sign extended.
if (Idx == 1 && Imm.getBitWidth() == 64 && isUInt<32>(Imm.getZExtValue()))
return TTI::TCC_Free;
// Fallthrough
case Instruction::Add:
case Instruction::Sub:
case Instruction::Mul:
case Instruction::UDiv:
case Instruction::SDiv:
case Instruction::URem:
case Instruction::SRem:
case Instruction::Or:
case Instruction::Xor:
ImmIdx = 1;
break;
// Always return TCC_Free for the shift value of a shift instruction.
case Instruction::Shl:
case Instruction::LShr:
case Instruction::AShr:
if (Idx == 1)
return TTI::TCC_Free;
break;
case Instruction::Trunc:
case Instruction::ZExt:
case Instruction::SExt:
case Instruction::IntToPtr:
case Instruction::PtrToInt:
case Instruction::BitCast:
case Instruction::PHI:
case Instruction::Call:
case Instruction::Select:
case Instruction::Ret:
case Instruction::Load:
break;
}
if (Idx == ImmIdx) {
int NumConstants = (BitSize + 63) / 64;
int Cost = X86TTIImpl::getIntImmCost(Imm, Ty);
return (Cost <= NumConstants * TTI::TCC_Basic)
? static_cast<int>(TTI::TCC_Free)
: Cost;
}
return X86TTIImpl::getIntImmCost(Imm, Ty);
}
int X86TTIImpl::getIntImmCost(Intrinsic::ID IID, unsigned Idx, const APInt &Imm,
Type *Ty) {
assert(Ty->isIntegerTy());
unsigned BitSize = Ty->getPrimitiveSizeInBits();
// There is no cost model for constants with a bit size of 0. Return TCC_Free
// here, so that constant hoisting will ignore this constant.
if (BitSize == 0)
return TTI::TCC_Free;
switch (IID) {
default:
return TTI::TCC_Free;
case Intrinsic::sadd_with_overflow:
case Intrinsic::uadd_with_overflow:
case Intrinsic::ssub_with_overflow:
case Intrinsic::usub_with_overflow:
case Intrinsic::smul_with_overflow:
case Intrinsic::umul_with_overflow:
if ((Idx == 1) && Imm.getBitWidth() <= 64 && isInt<32>(Imm.getSExtValue()))
return TTI::TCC_Free;
break;
case Intrinsic::experimental_stackmap:
if ((Idx < 2) || (Imm.getBitWidth() <= 64 && isInt<64>(Imm.getSExtValue())))
return TTI::TCC_Free;
break;
case Intrinsic::experimental_patchpoint_void:
case Intrinsic::experimental_patchpoint_i64:
if ((Idx < 4) || (Imm.getBitWidth() <= 64 && isInt<64>(Imm.getSExtValue())))
return TTI::TCC_Free;
break;
}
return X86TTIImpl::getIntImmCost(Imm, Ty);
}
// Return an average cost of Gather / Scatter instruction, maybe improved later
int X86TTIImpl::getGSVectorCost(unsigned Opcode, Type *SrcVTy, Value *Ptr,
unsigned Alignment, unsigned AddressSpace) {
assert(isa<VectorType>(SrcVTy) && "Unexpected type in getGSVectorCost");
unsigned VF = SrcVTy->getVectorNumElements();
// Try to reduce index size from 64 bit (default for GEP)
// to 32. It is essential for VF 16. If the index can't be reduced to 32, the
// operation will use 16 x 64 indices which do not fit in a zmm and needs
// to split. Also check that the base pointer is the same for all lanes,
// and that there's at most one variable index.
auto getIndexSizeInBits = [](Value *Ptr, const DataLayout& DL) {
unsigned IndexSize = DL.getPointerSizeInBits();
GetElementPtrInst *GEP = dyn_cast<GetElementPtrInst>(Ptr);
if (IndexSize < 64 || !GEP)
return IndexSize;
unsigned NumOfVarIndices = 0;
Value *Ptrs = GEP->getPointerOperand();
if (Ptrs->getType()->isVectorTy() && !getSplatValue(Ptrs))
return IndexSize;
for (unsigned i = 1; i < GEP->getNumOperands(); ++i) {
if (isa<Constant>(GEP->getOperand(i)))
continue;
Type *IndxTy = GEP->getOperand(i)->getType();
if (IndxTy->isVectorTy())
IndxTy = IndxTy->getVectorElementType();
if ((IndxTy->getPrimitiveSizeInBits() == 64 &&
!isa<SExtInst>(GEP->getOperand(i))) ||
++NumOfVarIndices > 1)
return IndexSize; // 64
}
return (unsigned)32;
};
// Trying to reduce IndexSize to 32 bits for vector 16.
// By default the IndexSize is equal to pointer size.
unsigned IndexSize = (VF >= 16) ? getIndexSizeInBits(Ptr, DL) :
DL.getPointerSizeInBits();
Type *IndexVTy = VectorType::get(IntegerType::get(SrcVTy->getContext(),
IndexSize), VF);
std::pair<int, MVT> IdxsLT = TLI->getTypeLegalizationCost(DL, IndexVTy);
std::pair<int, MVT> SrcLT = TLI->getTypeLegalizationCost(DL, SrcVTy);
int SplitFactor = std::max(IdxsLT.first, SrcLT.first);
if (SplitFactor > 1) {
// Handle splitting of vector of pointers
Type *SplitSrcTy = VectorType::get(SrcVTy->getScalarType(), VF / SplitFactor);
return SplitFactor * getGSVectorCost(Opcode, SplitSrcTy, Ptr, Alignment,
AddressSpace);
}
// The gather / scatter cost is given by Intel architects. It is a rough
// number since we are looking at one instruction in a time.
const int GSOverhead = 2;
return GSOverhead + VF * getMemoryOpCost(Opcode, SrcVTy->getScalarType(),
Alignment, AddressSpace);
}
/// Return the cost of full scalarization of gather / scatter operation.
///
/// Opcode - Load or Store instruction.
/// SrcVTy - The type of the data vector that should be gathered or scattered.
/// VariableMask - The mask is non-constant at compile time.
/// Alignment - Alignment for one element.
/// AddressSpace - pointer[s] address space.
///
int X86TTIImpl::getGSScalarCost(unsigned Opcode, Type *SrcVTy,
bool VariableMask, unsigned Alignment,
unsigned AddressSpace) {
unsigned VF = SrcVTy->getVectorNumElements();
int MaskUnpackCost = 0;
if (VariableMask) {
VectorType *MaskTy =
VectorType::get(Type::getInt1Ty(SrcVTy->getContext()), VF);
MaskUnpackCost = getScalarizationOverhead(MaskTy, false, true);
int ScalarCompareCost =
getCmpSelInstrCost(Instruction::ICmp, Type::getInt1Ty(SrcVTy->getContext()),
nullptr);
int BranchCost = getCFInstrCost(Instruction::Br);
MaskUnpackCost += VF * (BranchCost + ScalarCompareCost);
}
// The cost of the scalar loads/stores.
int MemoryOpCost = VF * getMemoryOpCost(Opcode, SrcVTy->getScalarType(),
Alignment, AddressSpace);
int InsertExtractCost = 0;
if (Opcode == Instruction::Load)
for (unsigned i = 0; i < VF; ++i)
// Add the cost of inserting each scalar load into the vector
InsertExtractCost +=
getVectorInstrCost(Instruction::InsertElement, SrcVTy, i);
else
for (unsigned i = 0; i < VF; ++i)
// Add the cost of extracting each element out of the data vector
InsertExtractCost +=
getVectorInstrCost(Instruction::ExtractElement, SrcVTy, i);
return MemoryOpCost + MaskUnpackCost + InsertExtractCost;
}
/// Calculate the cost of Gather / Scatter operation
int X86TTIImpl::getGatherScatterOpCost(unsigned Opcode, Type *SrcVTy,
Value *Ptr, bool VariableMask,
unsigned Alignment) {
assert(SrcVTy->isVectorTy() && "Unexpected data type for Gather/Scatter");
unsigned VF = SrcVTy->getVectorNumElements();
PointerType *PtrTy = dyn_cast<PointerType>(Ptr->getType());
if (!PtrTy && Ptr->getType()->isVectorTy())
PtrTy = dyn_cast<PointerType>(Ptr->getType()->getVectorElementType());
assert(PtrTy && "Unexpected type for Ptr argument");
unsigned AddressSpace = PtrTy->getAddressSpace();
bool Scalarize = false;
if ((Opcode == Instruction::Load && !isLegalMaskedGather(SrcVTy)) ||
(Opcode == Instruction::Store && !isLegalMaskedScatter(SrcVTy)))
Scalarize = true;
// Gather / Scatter for vector 2 is not profitable on KNL / SKX
// Vector-4 of gather/scatter instruction does not exist on KNL.
// We can extend it to 8 elements, but zeroing upper bits of
// the mask vector will add more instructions. Right now we give the scalar
// cost of vector-4 for KNL. TODO: Check, maybe the gather/scatter instruction is
// better in the VariableMask case.
if (VF == 2 || (VF == 4 && !ST->hasVLX()))
Scalarize = true;
if (Scalarize)
return getGSScalarCost(Opcode, SrcVTy, VariableMask, Alignment, AddressSpace);
return getGSVectorCost(Opcode, SrcVTy, Ptr, Alignment, AddressSpace);
}
bool X86TTIImpl::isLegalMaskedLoad(Type *DataTy) {
Type *ScalarTy = DataTy->getScalarType();
int DataWidth = isa<PointerType>(ScalarTy) ?
DL.getPointerSizeInBits() : ScalarTy->getPrimitiveSizeInBits();
return (DataWidth >= 32 && ST->hasAVX()) ||
(DataWidth >= 8 && ST->hasBWI());
}
bool X86TTIImpl::isLegalMaskedStore(Type *DataType) {
return isLegalMaskedLoad(DataType);
}
bool X86TTIImpl::isLegalMaskedGather(Type *DataTy) {
// This function is called now in two cases: from the Loop Vectorizer
// and from the Scalarizer.
// When the Loop Vectorizer asks about legality of the feature,
// the vectorization factor is not calculated yet. The Loop Vectorizer
// sends a scalar type and the decision is based on the width of the
// scalar element.
// Later on, the cost model will estimate usage this intrinsic based on
// the vector type.
// The Scalarizer asks again about legality. It sends a vector type.
// In this case we can reject non-power-of-2 vectors.
if (isa<VectorType>(DataTy) && !isPowerOf2_32(DataTy->getVectorNumElements()))
return false;
Type *ScalarTy = DataTy->getScalarType();
int DataWidth = isa<PointerType>(ScalarTy) ?
DL.getPointerSizeInBits() : ScalarTy->getPrimitiveSizeInBits();
// AVX-512 allows gather and scatter
return DataWidth >= 32 && ST->hasAVX512();
}
bool X86TTIImpl::isLegalMaskedScatter(Type *DataType) {
return isLegalMaskedGather(DataType);
}
bool X86TTIImpl::areInlineCompatible(const Function *Caller,
const Function *Callee) const {
const TargetMachine &TM = getTLI()->getTargetMachine();
// Work this as a subsetting of subtarget features.
const FeatureBitset &CallerBits =
TM.getSubtargetImpl(*Caller)->getFeatureBits();
const FeatureBitset &CalleeBits =
TM.getSubtargetImpl(*Callee)->getFeatureBits();
// FIXME: This is likely too limiting as it will include subtarget features
// that we might not care about for inlining, but it is conservatively
// correct.
return (CallerBits & CalleeBits) == CalleeBits;
}