NVPTXTargetTransformInfo.cpp

Go to the documentation of this file.

//===-- NVPTXTargetTransformInfo.cpp - NVPTX specific TTI -----------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
 
#include "NVPTXTargetTransformInfo.h"
#include "NVPTXUtilities.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/Analysis/LoopInfo.h"
#include "llvm/Analysis/TargetTransformInfo.h"
#include "llvm/Analysis/ValueTracking.h"
#include "llvm/CodeGen/BasicTTIImpl.h"
#include "llvm/CodeGen/TargetLowering.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/IntrinsicInst.h"
#include "llvm/IR/Intrinsics.h"
#include "llvm/IR/IntrinsicsNVPTX.h"
#include "llvm/IR/Value.h"
#include "llvm/Support/Casting.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/NVPTXAddrSpace.h"
#include "llvm/Transforms/InstCombine/InstCombiner.h"
#include <optional>
using namespace llvm;
 
#define DEBUG_TYPE "NVPTXtti"
 
// Whether the given intrinsic reads threadIdx.x/y/z.
static bool readsThreadIndex(const IntrinsicInst *II) {
  switch (II->getIntrinsicID()) {
    default: return false;
    case Intrinsic::nvvm_read_ptx_sreg_tid_x:
    case Intrinsic::nvvm_read_ptx_sreg_tid_y:
    case Intrinsic::nvvm_read_ptx_sreg_tid_z:
      return true;
  }
}
 
static bool readsLaneId(const IntrinsicInst *II) {
  return II->getIntrinsicID() == Intrinsic::nvvm_read_ptx_sreg_laneid;
}
 
// Whether the given intrinsic is an atomic instruction in PTX.
static bool isNVVMAtomic(const IntrinsicInst *II) {
  switch (II->getIntrinsicID()) {
  default:
    return false;
  case Intrinsic::nvvm_atomic_add_gen_f_cta:
  case Intrinsic::nvvm_atomic_add_gen_f_sys:
  case Intrinsic::nvvm_atomic_add_gen_i_cta:
  case Intrinsic::nvvm_atomic_add_gen_i_sys:
  case Intrinsic::nvvm_atomic_and_gen_i_cta:
  case Intrinsic::nvvm_atomic_and_gen_i_sys:
  case Intrinsic::nvvm_atomic_cas_gen_i_cta:
  case Intrinsic::nvvm_atomic_cas_gen_i_sys:
  case Intrinsic::nvvm_atomic_dec_gen_i_cta:
  case Intrinsic::nvvm_atomic_dec_gen_i_sys:
  case Intrinsic::nvvm_atomic_inc_gen_i_cta:
  case Intrinsic::nvvm_atomic_inc_gen_i_sys:
  case Intrinsic::nvvm_atomic_max_gen_i_cta:
  case Intrinsic::nvvm_atomic_max_gen_i_sys:
  case Intrinsic::nvvm_atomic_min_gen_i_cta:
  case Intrinsic::nvvm_atomic_min_gen_i_sys:
  case Intrinsic::nvvm_atomic_or_gen_i_cta:
  case Intrinsic::nvvm_atomic_or_gen_i_sys:
  case Intrinsic::nvvm_atomic_exch_gen_i_cta:
  case Intrinsic::nvvm_atomic_exch_gen_i_sys:
  case Intrinsic::nvvm_atomic_xor_gen_i_cta:
  case Intrinsic::nvvm_atomic_xor_gen_i_sys:
    return true;
  }
}
 
bool NVPTXTTIImpl::isSourceOfDivergence(const Value *V) const {
  // Without inter-procedural analysis, we conservatively assume that arguments
  // to __device__ functions are divergent.
  if (const Argument *Arg = dyn_cast<Argument>(V))
    return !isKernelFunction(*Arg->getParent());
 
  if (const Instruction *I = dyn_cast<Instruction>(V)) {
    // Without pointer analysis, we conservatively assume values loaded from
    // generic or local address space are divergent.
    if (const LoadInst *LI = dyn_cast<LoadInst>(I)) {
      unsigned AS = LI->getPointerAddressSpace();
      return AS == ADDRESS_SPACE_GENERIC || AS == ADDRESS_SPACE_LOCAL;
    }
    // Atomic instructions may cause divergence. Atomic instructions are
    // executed sequentially across all threads in a warp. Therefore, an earlier
    // executed thread may see different memory inputs than a later executed
    // thread. For example, suppose *a = 0 initially.
    //
    //   atom.global.add.s32 d, [a], 1
    //
    // returns 0 for the first thread that enters the critical region, and 1 for
    // the second thread.
    if (I->isAtomic())
      return true;
    if (const IntrinsicInst *II = dyn_cast<IntrinsicInst>(I)) {
      // Instructions that read threadIdx are obviously divergent.
      if (readsThreadIndex(II) || readsLaneId(II))
        return true;
      // Handle the NVPTX atomic intrinsics that cannot be represented as an
      // atomic IR instruction.
      if (isNVVMAtomic(II))
        return true;
    }
    // Conservatively consider the return value of function calls as divergent.
    // We could analyze callees with bodies more precisely using
    // inter-procedural analysis.
    if (isa<CallInst>(I))
      return true;
  }
 
  return false;
}
 
// Convert NVVM intrinsics to target-generic LLVM code where possible.
static Instruction *convertNvvmIntrinsicToLlvm(InstCombiner &IC,
                                               IntrinsicInst *II) {
  // Each NVVM intrinsic we can simplify can be replaced with one of:
  //
  //  * an LLVM intrinsic,
  //  * an LLVM cast operation,
  //  * an LLVM binary operation, or
  //  * ad-hoc LLVM IR for the particular operation.
 
  // Some transformations are only valid when the module's
  // flush-denormals-to-zero (ftz) setting is true/false, whereas other
  // transformations are valid regardless of the module's ftz setting.
  enum FtzRequirementTy {
    FTZ_Any,       // Any ftz setting is ok.
    FTZ_MustBeOn,  // Transformation is valid only if ftz is on.
    FTZ_MustBeOff, // Transformation is valid only if ftz is off.
  };
  // Classes of NVVM intrinsics that can't be replaced one-to-one with a
  // target-generic intrinsic, cast op, or binary op but that we can nonetheless
  // simplify.
  enum SpecialCase {
    SPC_Reciprocal,
    SCP_FunnelShiftClamp,
  };
 
  // SimplifyAction is a poor-man's variant (plus an additional flag) that
  // represents how to replace an NVVM intrinsic with target-generic LLVM IR.
  struct SimplifyAction {
    // Invariant: At most one of these Optionals has a value.
    std::optional<Intrinsic::ID> IID;
    std::optional<Instruction::CastOps> CastOp;
    std::optional<Instruction::BinaryOps> BinaryOp;
    std::optional<SpecialCase> Special;
 
    FtzRequirementTy FtzRequirement = FTZ_Any;
    // Denormal handling is guarded by different attributes depending on the
    // type (denormal-fp-math vs denormal-fp-math-f32), take note of halfs.
    bool IsHalfTy = false;
 
    SimplifyAction() = default;
 
    SimplifyAction(Intrinsic::ID IID, FtzRequirementTy FtzReq,
                   bool IsHalfTy = false)
        : IID(IID), FtzRequirement(FtzReq), IsHalfTy(IsHalfTy) {}
 
    // Cast operations don't have anything to do with FTZ, so we skip that
    // argument.
    SimplifyAction(Instruction::CastOps CastOp) : CastOp(CastOp) {}
 
    SimplifyAction(Instruction::BinaryOps BinaryOp, FtzRequirementTy FtzReq)
        : BinaryOp(BinaryOp), FtzRequirement(FtzReq) {}
 
    SimplifyAction(SpecialCase Special, FtzRequirementTy FtzReq)
        : Special(Special), FtzRequirement(FtzReq) {}
  };
 
  // Try to generate a SimplifyAction describing how to replace our
  // IntrinsicInstr with target-generic LLVM IR.
  const SimplifyAction Action = [II]() -> SimplifyAction {
    switch (II->getIntrinsicID()) {
    // NVVM intrinsics that map directly to LLVM intrinsics.
    case Intrinsic::nvvm_ceil_d:
      return {Intrinsic::ceil, FTZ_Any};
    case Intrinsic::nvvm_ceil_f:
      return {Intrinsic::ceil, FTZ_MustBeOff};
    case Intrinsic::nvvm_ceil_ftz_f:
      return {Intrinsic::ceil, FTZ_MustBeOn};
    case Intrinsic::nvvm_floor_d:
      return {Intrinsic::floor, FTZ_Any};
    case Intrinsic::nvvm_floor_f:
      return {Intrinsic::floor, FTZ_MustBeOff};
    case Intrinsic::nvvm_floor_ftz_f:
      return {Intrinsic::floor, FTZ_MustBeOn};
    case Intrinsic::nvvm_fma_rn_d:
      return {Intrinsic::fma, FTZ_Any};
    case Intrinsic::nvvm_fma_rn_f:
      return {Intrinsic::fma, FTZ_MustBeOff};
    case Intrinsic::nvvm_fma_rn_ftz_f:
      return {Intrinsic::fma, FTZ_MustBeOn};
    case Intrinsic::nvvm_fma_rn_f16:
      return {Intrinsic::fma, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fma_rn_ftz_f16:
      return {Intrinsic::fma, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fma_rn_f16x2:
      return {Intrinsic::fma, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fma_rn_ftz_f16x2:
      return {Intrinsic::fma, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fma_rn_bf16:
      return {Intrinsic::fma, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fma_rn_ftz_bf16:
      return {Intrinsic::fma, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fma_rn_bf16x2:
      return {Intrinsic::fma, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fma_rn_ftz_bf16x2:
      return {Intrinsic::fma, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fmax_d:
      return {Intrinsic::maxnum, FTZ_Any};
    case Intrinsic::nvvm_fmax_f:
      return {Intrinsic::maxnum, FTZ_MustBeOff};
    case Intrinsic::nvvm_fmax_ftz_f:
      return {Intrinsic::maxnum, FTZ_MustBeOn};
    case Intrinsic::nvvm_fmax_nan_f:
      return {Intrinsic::maximum, FTZ_MustBeOff};
    case Intrinsic::nvvm_fmax_ftz_nan_f:
      return {Intrinsic::maximum, FTZ_MustBeOn};
    case Intrinsic::nvvm_fmax_f16:
      return {Intrinsic::maxnum, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fmax_ftz_f16:
      return {Intrinsic::maxnum, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fmax_f16x2:
      return {Intrinsic::maxnum, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fmax_ftz_f16x2:
      return {Intrinsic::maxnum, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fmax_nan_f16:
      return {Intrinsic::maximum, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fmax_ftz_nan_f16:
      return {Intrinsic::maximum, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fmax_nan_f16x2:
      return {Intrinsic::maximum, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fmax_ftz_nan_f16x2:
      return {Intrinsic::maximum, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fmin_d:
      return {Intrinsic::minnum, FTZ_Any};
    case Intrinsic::nvvm_fmin_f:
      return {Intrinsic::minnum, FTZ_MustBeOff};
    case Intrinsic::nvvm_fmin_ftz_f:
      return {Intrinsic::minnum, FTZ_MustBeOn};
    case Intrinsic::nvvm_fmin_nan_f:
      return {Intrinsic::minimum, FTZ_MustBeOff};
    case Intrinsic::nvvm_fmin_ftz_nan_f:
      return {Intrinsic::minimum, FTZ_MustBeOn};
    case Intrinsic::nvvm_fmin_f16:
      return {Intrinsic::minnum, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fmin_ftz_f16:
      return {Intrinsic::minnum, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fmin_f16x2:
      return {Intrinsic::minnum, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fmin_ftz_f16x2:
      return {Intrinsic::minnum, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fmin_nan_f16:
      return {Intrinsic::minimum, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fmin_ftz_nan_f16:
      return {Intrinsic::minimum, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_fmin_nan_f16x2:
      return {Intrinsic::minimum, FTZ_MustBeOff, true};
    case Intrinsic::nvvm_fmin_ftz_nan_f16x2:
      return {Intrinsic::minimum, FTZ_MustBeOn, true};
    case Intrinsic::nvvm_sqrt_rn_d:
      return {Intrinsic::sqrt, FTZ_Any};
    case Intrinsic::nvvm_sqrt_f:
      // nvvm_sqrt_f is a special case.  For  most intrinsics, foo_ftz_f is the
      // ftz version, and foo_f is the non-ftz version.  But nvvm_sqrt_f adopts
      // the ftz-ness of the surrounding code.  sqrt_rn_f and sqrt_rn_ftz_f are
      // the versions with explicit ftz-ness.
      return {Intrinsic::sqrt, FTZ_Any};
    case Intrinsic::nvvm_trunc_d:
      return {Intrinsic::trunc, FTZ_Any};
    case Intrinsic::nvvm_trunc_f:
      return {Intrinsic::trunc, FTZ_MustBeOff};
    case Intrinsic::nvvm_trunc_ftz_f:
      return {Intrinsic::trunc, FTZ_MustBeOn};
 
    // NVVM intrinsics that map to LLVM cast operations.
    //
    // Note that llvm's target-generic conversion operators correspond to the rz
    // (round to zero) versions of the nvvm conversion intrinsics, even though
    // most everything else here uses the rn (round to nearest even) nvvm ops.
    case Intrinsic::nvvm_d2i_rz:
    case Intrinsic::nvvm_f2i_rz:
    case Intrinsic::nvvm_d2ll_rz:
    case Intrinsic::nvvm_f2ll_rz:
      return {Instruction::FPToSI};
    case Intrinsic::nvvm_d2ui_rz:
    case Intrinsic::nvvm_f2ui_rz:
    case Intrinsic::nvvm_d2ull_rz:
    case Intrinsic::nvvm_f2ull_rz:
      return {Instruction::FPToUI};
    // Integer to floating-point uses RN rounding, not RZ
    case Intrinsic::nvvm_i2d_rn:
    case Intrinsic::nvvm_i2f_rn:
    case Intrinsic::nvvm_ll2d_rn:
    case Intrinsic::nvvm_ll2f_rn:
      return {Instruction::SIToFP};
    case Intrinsic::nvvm_ui2d_rn:
    case Intrinsic::nvvm_ui2f_rn:
    case Intrinsic::nvvm_ull2d_rn:
    case Intrinsic::nvvm_ull2f_rn:
      return {Instruction::UIToFP};
 
    // NVVM intrinsics that map to LLVM binary ops.
    case Intrinsic::nvvm_div_rn_d:
      return {Instruction::FDiv, FTZ_Any};
 
    // The remainder of cases are NVVM intrinsics that map to LLVM idioms, but
    // need special handling.
    //
    // We seem to be missing intrinsics for rcp.approx.{ftz.}f32, which is just
    // as well.
    case Intrinsic::nvvm_rcp_rn_d:
      return {SPC_Reciprocal, FTZ_Any};
 
    case Intrinsic::nvvm_fshl_clamp:
    case Intrinsic::nvvm_fshr_clamp:
      return {SCP_FunnelShiftClamp, FTZ_Any};
 
      // We do not currently simplify intrinsics that give an approximate
      // answer. These include:
      //
      //   - nvvm_cos_approx_{f,ftz_f}
      //   - nvvm_ex2_approx_{d,f,ftz_f}
      //   - nvvm_lg2_approx_{d,f,ftz_f}
      //   - nvvm_sin_approx_{f,ftz_f}
      //   - nvvm_sqrt_approx_{f,ftz_f}
      //   - nvvm_rsqrt_approx_{d,f,ftz_f}
      //   - nvvm_div_approx_{ftz_d,ftz_f,f}
      //   - nvvm_rcp_approx_ftz_d
      //
      // Ideally we'd encode them as e.g. "fast call @llvm.cos", where "fast"
      // means that fastmath is enabled in the intrinsic.  Unfortunately only
      // binary operators (currently) have a fastmath bit in SelectionDAG, so
      // this information gets lost and we can't select on it.
      //
      // TODO: div and rcp are lowered to a binary op, so these we could in
      // theory lower them to "fast fdiv".
 
    default:
      return {};
    }
  }();
 
  // If Action.FtzRequirementTy is not satisfied by the module's ftz state, we
  // can bail out now.  (Notice that in the case that IID is not an NVVM
  // intrinsic, we don't have to look up any module metadata, as
  // FtzRequirementTy will be FTZ_Any.)
  if (Action.FtzRequirement != FTZ_Any) {
    // FIXME: Broken for f64
    DenormalMode Mode = II->getFunction()->getDenormalMode(
        Action.IsHalfTy ? APFloat::IEEEhalf() : APFloat::IEEEsingle());
    bool FtzEnabled = Mode.Output == DenormalMode::PreserveSign;
 
    if (FtzEnabled != (Action.FtzRequirement == FTZ_MustBeOn))
      return nullptr;
  }
 
  // Simplify to target-generic intrinsic.
  if (Action.IID) {
    SmallVector<Value *, 4> Args(II->args());
    // All the target-generic intrinsics currently of interest to us have one
    // type argument, equal to that of the nvvm intrinsic's argument.
    Type *Tys[] = {II->getArgOperand(0)->getType()};
    return CallInst::Create(
        Intrinsic::getOrInsertDeclaration(II->getModule(), *Action.IID, Tys),
        Args);
  }
 
  // Simplify to target-generic binary op.
  if (Action.BinaryOp)
    return BinaryOperator::Create(*Action.BinaryOp, II->getArgOperand(0),
                                  II->getArgOperand(1), II->getName());
 
  // Simplify to target-generic cast op.
  if (Action.CastOp)
    return CastInst::Create(*Action.CastOp, II->getArgOperand(0), II->getType(),
                            II->getName());
 
  // All that's left are the special cases.
  if (!Action.Special)
    return nullptr;
 
  switch (*Action.Special) {
  case SPC_Reciprocal:
    // Simplify reciprocal.
    return BinaryOperator::Create(
        Instruction::FDiv, ConstantFP::get(II->getArgOperand(0)->getType(), 1),
        II->getArgOperand(0), II->getName());
 
  case SCP_FunnelShiftClamp: {
    // Canonicalize a clamping funnel shift to the generic llvm funnel shift
    // when possible, as this is easier for llvm to optimize further.
    if (const auto *ShiftConst = dyn_cast<ConstantInt>(II->getArgOperand(2))) {
      const bool IsLeft = II->getIntrinsicID() == Intrinsic::nvvm_fshl_clamp;
      if (ShiftConst->getZExtValue() >= II->getType()->getIntegerBitWidth())
        return IC.replaceInstUsesWith(*II, II->getArgOperand(IsLeft ? 1 : 0));
 
      const unsigned FshIID = IsLeft ? Intrinsic::fshl : Intrinsic::fshr;
      return CallInst::Create(Intrinsic::getOrInsertDeclaration(
                                  II->getModule(), FshIID, II->getType()),
                              SmallVector<Value *, 3>(II->args()));
    }
    return nullptr;
  }
  }
  llvm_unreachable("All SpecialCase enumerators should be handled in switch.");
}
 
// Returns true/false when we know the answer, nullopt otherwise.
static std::optional<bool> evaluateIsSpace(Intrinsic::ID IID, unsigned AS) {
  if (AS == NVPTXAS::ADDRESS_SPACE_GENERIC ||
      AS == NVPTXAS::ADDRESS_SPACE_PARAM)
    return std::nullopt; // Got to check at run-time.
  switch (IID) {
  case Intrinsic::nvvm_isspacep_global:
    return AS == NVPTXAS::ADDRESS_SPACE_GLOBAL;
  case Intrinsic::nvvm_isspacep_local:
    return AS == NVPTXAS::ADDRESS_SPACE_LOCAL;
  case Intrinsic::nvvm_isspacep_shared:
    // If shared cluster this can't be evaluated at compile time.
    if (AS == NVPTXAS::ADDRESS_SPACE_SHARED_CLUSTER)
      return std::nullopt;
    return AS == NVPTXAS::ADDRESS_SPACE_SHARED;
  case Intrinsic::nvvm_isspacep_shared_cluster:
    return AS == NVPTXAS::ADDRESS_SPACE_SHARED_CLUSTER ||
           AS == NVPTXAS::ADDRESS_SPACE_SHARED;
  case Intrinsic::nvvm_isspacep_const:
    return AS == NVPTXAS::ADDRESS_SPACE_CONST;
  default:
    llvm_unreachable("Unexpected intrinsic");
  }
}
 
// Returns an instruction pointer (may be nullptr if we do not know the answer).
// Returns nullopt if `II` is not one of the `isspacep` intrinsics.
//
// TODO: If InferAddressSpaces were run early enough in the pipeline this could
// be removed in favor of the constant folding that occurs there through
// rewriteIntrinsicWithAddressSpace
static std::optional<Instruction *>
handleSpaceCheckIntrinsics(InstCombiner &IC, IntrinsicInst &II) {
 
  switch (auto IID = II.getIntrinsicID()) {
  case Intrinsic::nvvm_isspacep_global:
  case Intrinsic::nvvm_isspacep_local:
  case Intrinsic::nvvm_isspacep_shared:
  case Intrinsic::nvvm_isspacep_shared_cluster:
  case Intrinsic::nvvm_isspacep_const: {
    Value *Op0 = II.getArgOperand(0);
    unsigned AS = Op0->getType()->getPointerAddressSpace();
    // Peek through ASC to generic AS.
    // TODO: we could dig deeper through both ASCs and GEPs.
    if (AS == NVPTXAS::ADDRESS_SPACE_GENERIC)
      if (auto *ASCO = dyn_cast<AddrSpaceCastOperator>(Op0))
        AS = ASCO->getOperand(0)->getType()->getPointerAddressSpace();
 
    if (std::optional<bool> Answer = evaluateIsSpace(IID, AS))
      return IC.replaceInstUsesWith(II,
                                    ConstantInt::get(II.getType(), *Answer));
    return nullptr; // Don't know the answer, got to check at run time.
  }
  default:
    return std::nullopt;
  }
}
 
std::optional<Instruction *>
NVPTXTTIImpl::instCombineIntrinsic(InstCombiner &IC, IntrinsicInst &II) const {
  if (std::optional<Instruction *> I = handleSpaceCheckIntrinsics(IC, II))
    return *I;
  if (Instruction *I = convertNvvmIntrinsicToLlvm(IC, &II))
    return I;
 
  return std::nullopt;
}
 
InstructionCost
NVPTXTTIImpl::getInstructionCost(const User *U,
                                 ArrayRef<const Value *> Operands,
                                 TTI::TargetCostKind CostKind) const {
  if (const auto *CI = dyn_cast<CallInst>(U))
    if (const auto *IA = dyn_cast<InlineAsm>(CI->getCalledOperand())) {
      // Without this implementation getCallCost() would return the number
      // of arguments+1 as the cost. Because the cost-model assumes it is a call
      // since it is classified as a call in the IR. A better cost model would
      // be to return the number of asm instructions embedded in the asm
      // string.
      StringRef AsmStr = IA->getAsmString();
      const unsigned InstCount =
          count_if(split(AsmStr, ';'), [](StringRef AsmInst) {
            // Trim off scopes denoted by '{' and '}' as these can be ignored
            AsmInst = AsmInst.trim().ltrim("{} \t\n\v\f\r");
            // This is pretty coarse but does a reasonably good job of
            // identifying things that look like instructions, possibly with a
            // predicate ("@").
            return !AsmInst.empty() &&
                   (AsmInst[0] == '@' || isAlpha(AsmInst[0]) ||
                    AsmInst.find(".pragma") != StringRef::npos);
          });
      return InstCount * TargetTransformInfo::TCC_Basic;
    }
 
  return BaseT::getInstructionCost(U, Operands, CostKind);
}
 
InstructionCost NVPTXTTIImpl::getArithmeticInstrCost(
    unsigned Opcode, Type *Ty, TTI::TargetCostKind CostKind,
    TTI::OperandValueInfo Op1Info, TTI::OperandValueInfo Op2Info,
    ArrayRef<const Value *> Args, const Instruction *CxtI) const {
  // Legalize the type.
  std::pair<InstructionCost, MVT> LT = getTypeLegalizationCost(Ty);
 
  int ISD = TLI->InstructionOpcodeToISD(Opcode);
 
  switch (ISD) {
  default:
    return BaseT::getArithmeticInstrCost(Opcode, Ty, CostKind, Op1Info,
                                         Op2Info);
  case ISD::ADD:
  case ISD::MUL:
  case ISD::XOR:
  case ISD::OR:
  case ISD::AND:
    // The machine code (SASS) simulates an i64 with two i32. Therefore, we
    // estimate that arithmetic operations on i64 are twice as expensive as
    // those on types that can fit into one machine register.
    if (LT.second.SimpleTy == MVT::i64)
      return 2 * LT.first;
    // Delegate other cases to the basic TTI.
    return BaseT::getArithmeticInstrCost(Opcode, Ty, CostKind, Op1Info,
                                         Op2Info);
  }
}
 
void NVPTXTTIImpl::getUnrollingPreferences(
    Loop *L, ScalarEvolution &SE, TTI::UnrollingPreferences &UP,
    OptimizationRemarkEmitter *ORE) const {
  BaseT::getUnrollingPreferences(L, SE, UP, ORE);
 
  // Enable partial unrolling and runtime unrolling, but reduce the
  // threshold.  This partially unrolls small loops which are often
  // unrolled by the PTX to SASS compiler and unrolling earlier can be
  // beneficial.
  UP.Partial = UP.Runtime = true;
  UP.PartialThreshold = UP.Threshold / 4;
}
 
void NVPTXTTIImpl::getPeelingPreferences(Loop *L, ScalarEvolution &SE,
                                         TTI::PeelingPreferences &PP) const {
  BaseT::getPeelingPreferences(L, SE, PP);
}
 
bool NVPTXTTIImpl::collectFlatAddressOperands(SmallVectorImpl<int> &OpIndexes,
                                              Intrinsic::ID IID) const {
  switch (IID) {
  case Intrinsic::nvvm_isspacep_const:
  case Intrinsic::nvvm_isspacep_global:
  case Intrinsic::nvvm_isspacep_local:
  case Intrinsic::nvvm_isspacep_shared:
  case Intrinsic::nvvm_isspacep_shared_cluster:
  case Intrinsic::nvvm_prefetch_tensormap: {
    OpIndexes.push_back(0);
    return true;
  }
  }
  return false;
}
 
Value *NVPTXTTIImpl::rewriteIntrinsicWithAddressSpace(IntrinsicInst *II,
                                                      Value *OldV,
                                                      Value *NewV) const {
  const Intrinsic::ID IID = II->getIntrinsicID();
  switch (IID) {
  case Intrinsic::nvvm_isspacep_const:
  case Intrinsic::nvvm_isspacep_global:
  case Intrinsic::nvvm_isspacep_local:
  case Intrinsic::nvvm_isspacep_shared:
  case Intrinsic::nvvm_isspacep_shared_cluster: {
    const unsigned NewAS = NewV->getType()->getPointerAddressSpace();
    if (const auto R = evaluateIsSpace(IID, NewAS))
      return ConstantInt::get(II->getType(), *R);
    return nullptr;
  }
  case Intrinsic::nvvm_prefetch_tensormap: {
    IRBuilder<> Builder(II);
    return Builder.CreateUnaryIntrinsic(Intrinsic::nvvm_prefetch_tensormap,
                                        NewV);
  }
  }
  return nullptr;
}
 
unsigned NVPTXTTIImpl::getLoadStoreVecRegBitWidth(unsigned AddrSpace) const {
  // 256 bit loads/stores are currently only supported for global address space
  if (ST->has256BitVectorLoadStore(AddrSpace))
    return 256;
  return 128;
}
 
unsigned NVPTXTTIImpl::getAssumedAddrSpace(const Value *V) const {
  if (isa<AllocaInst>(V))
    return ADDRESS_SPACE_LOCAL;
 
  if (const Argument *Arg = dyn_cast<Argument>(V)) {
    if (isKernelFunction(*Arg->getParent())) {
      const NVPTXTargetMachine &TM =
          static_cast<const NVPTXTargetMachine &>(getTLI()->getTargetMachine());
      if (TM.getDrvInterface() == NVPTX::CUDA && !Arg->hasByValAttr())
        return ADDRESS_SPACE_GLOBAL;
    } else {
      // We assume that all device parameters that are passed byval will be
      // placed in the local AS. Very simple cases will be updated after ISel to
      // use the device param space where possible.
      if (Arg->hasByValAttr())
        return ADDRESS_SPACE_LOCAL;
    }
  }
 
  return -1;
}
 
void NVPTXTTIImpl::collectKernelLaunchBounds(
    const Function &F,
    SmallVectorImpl<std::pair<StringRef, int64_t>> &LB) const {
  if (const auto Val = getMaxClusterRank(F))
    LB.push_back({"maxclusterrank", *Val});
 
  const auto MaxNTID = getMaxNTID(F);
  if (MaxNTID.size() > 0)
    LB.push_back({"maxntidx", MaxNTID[0]});
  if (MaxNTID.size() > 1)
    LB.push_back({"maxntidy", MaxNTID[1]});
  if (MaxNTID.size() > 2)
    LB.push_back({"maxntidz", MaxNTID[2]});
}