maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/lib/Target/AMDGPU/AMDGPULibCalls.cpp

1771 lines
54 KiB
C++
Raw Blame History

//===- AMDGPULibCalls.cpp -------------------------------------------------===//
//
// 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
//
//===----------------------------------------------------------------------===//
//
/// \file
/// This file does AMD library function optimizations.
//
//===----------------------------------------------------------------------===//
#include "AMDGPU.h"
#include "AMDGPULibFunc.h"
#include "AMDGPUSubtarget.h"
#include "llvm/ADT/StringRef.h"
#include "llvm/ADT/StringSet.h"
#include "llvm/Analysis/AliasAnalysis.h"
#include "llvm/Analysis/Loads.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/DerivedTypes.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/IRBuilder.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/Intrinsics.h"
#include "llvm/IR/LLVMContext.h"
#include "llvm/IR/Module.h"
#include "llvm/IR/ValueSymbolTable.h"
#include "llvm/InitializePasses.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/MathExtras.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/Target/TargetMachine.h"
#include <cmath>
#include <vector>
#define DEBUG_TYPE "amdgpu-simplifylib"
using namespace llvm;
static cl::opt<bool> EnablePreLink("amdgpu-prelink",
cl::desc("Enable pre-link mode optimizations"),
cl::init(false),
cl::Hidden);
static cl::list<std::string> UseNative("amdgpu-use-native",
cl::desc("Comma separated list of functions to replace with native, or all"),
cl::CommaSeparated, cl::ValueOptional,
cl::Hidden);
#define MATH_PI numbers::pi
#define MATH_E numbers::e
#define MATH_SQRT2 numbers::sqrt2
#define MATH_SQRT1_2 numbers::inv_sqrt2
namespace llvm {
class AMDGPULibCalls {
private:
typedef llvm::AMDGPULibFunc FuncInfo;
const TargetMachine *TM;
// -fuse-native.
bool AllNative = false;
bool useNativeFunc(const StringRef F) const;
// Return a pointer (pointer expr) to the function if function defintion with
// "FuncName" exists. It may create a new function prototype in pre-link mode.
FunctionCallee getFunction(Module *M, const FuncInfo &fInfo);
// Replace a normal function with its native version.
bool replaceWithNative(CallInst *CI, const FuncInfo &FInfo);
bool parseFunctionName(const StringRef& FMangledName,
FuncInfo *FInfo=nullptr /*out*/);
bool TDOFold(CallInst *CI, const FuncInfo &FInfo);
/* Specialized optimizations */
// recip (half or native)
bool fold_recip(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// divide (half or native)
bool fold_divide(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// pow/powr/pown
bool fold_pow(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// rootn
bool fold_rootn(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// fma/mad
bool fold_fma_mad(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// -fuse-native for sincos
bool sincosUseNative(CallInst *aCI, const FuncInfo &FInfo);
// evaluate calls if calls' arguments are constants.
bool evaluateScalarMathFunc(FuncInfo &FInfo, double& Res0,
double& Res1, Constant *copr0, Constant *copr1, Constant *copr2);
bool evaluateCall(CallInst *aCI, FuncInfo &FInfo);
// exp
bool fold_exp(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// exp2
bool fold_exp2(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// exp10
bool fold_exp10(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// log
bool fold_log(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// log2
bool fold_log2(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// log10
bool fold_log10(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// sqrt
bool fold_sqrt(CallInst *CI, IRBuilder<> &B, const FuncInfo &FInfo);
// sin/cos
bool fold_sincos(CallInst * CI, IRBuilder<> &B, AliasAnalysis * AA);
// __read_pipe/__write_pipe
bool fold_read_write_pipe(CallInst *CI, IRBuilder<> &B, FuncInfo &FInfo);
// llvm.amdgcn.wavefrontsize
bool fold_wavefrontsize(CallInst *CI, IRBuilder<> &B);
// Get insertion point at entry.
BasicBlock::iterator getEntryIns(CallInst * UI);
// Insert an Alloc instruction.
AllocaInst* insertAlloca(CallInst * UI, IRBuilder<> &B, const char *prefix);
// Get a scalar native builtin signle argument FP function
FunctionCallee getNativeFunction(Module *M, const FuncInfo &FInfo);
protected:
CallInst *CI;
bool isUnsafeMath(const CallInst *CI) const;
void replaceCall(Value *With) {
CI->replaceAllUsesWith(With);
CI->eraseFromParent();
}
public:
AMDGPULibCalls(const TargetMachine *TM_ = nullptr) : TM(TM_) {}
bool fold(CallInst *CI, AliasAnalysis *AA = nullptr);
void initNativeFuncs();
// Replace a normal math function call with that native version
bool useNative(CallInst *CI);
};
} // end llvm namespace
namespace {
class AMDGPUSimplifyLibCalls : public FunctionPass {
AMDGPULibCalls Simplifier;
public:
static char ID; // Pass identification
AMDGPUSimplifyLibCalls(const TargetMachine *TM = nullptr)
: FunctionPass(ID), Simplifier(TM) {
initializeAMDGPUSimplifyLibCallsPass(*PassRegistry::getPassRegistry());
}
void getAnalysisUsage(AnalysisUsage &AU) const override {
AU.addRequired<AAResultsWrapperPass>();
}
bool runOnFunction(Function &M) override;
};
class AMDGPUUseNativeCalls : public FunctionPass {
AMDGPULibCalls Simplifier;
public:
static char ID; // Pass identification
AMDGPUUseNativeCalls() : FunctionPass(ID) {
initializeAMDGPUUseNativeCallsPass(*PassRegistry::getPassRegistry());
Simplifier.initNativeFuncs();
}
bool runOnFunction(Function &F) override;
};
} // end anonymous namespace.
char AMDGPUSimplifyLibCalls::ID = 0;
char AMDGPUUseNativeCalls::ID = 0;
INITIALIZE_PASS_BEGIN(AMDGPUSimplifyLibCalls, "amdgpu-simplifylib",
"Simplify well-known AMD library calls", false, false)
INITIALIZE_PASS_DEPENDENCY(AAResultsWrapperPass)
INITIALIZE_PASS_END(AMDGPUSimplifyLibCalls, "amdgpu-simplifylib",
"Simplify well-known AMD library calls", false, false)
INITIALIZE_PASS(AMDGPUUseNativeCalls, "amdgpu-usenative",
"Replace builtin math calls with that native versions.",
false, false)
template <typename IRB>
static CallInst *CreateCallEx(IRB &B, FunctionCallee Callee, Value *Arg,
const Twine &Name = "") {
CallInst *R = B.CreateCall(Callee, Arg, Name);
if (Function *F = dyn_cast<Function>(Callee.getCallee()))
R->setCallingConv(F->getCallingConv());
return R;
}
template <typename IRB>
static CallInst *CreateCallEx2(IRB &B, FunctionCallee Callee, Value *Arg1,
Value *Arg2, const Twine &Name = "") {
CallInst *R = B.CreateCall(Callee, {Arg1, Arg2}, Name);
if (Function *F = dyn_cast<Function>(Callee.getCallee()))
R->setCallingConv(F->getCallingConv());
return R;
}
// Data structures for table-driven optimizations.
// FuncTbl works for both f32 and f64 functions with 1 input argument
struct TableEntry {
double result;
double input;
};
/* a list of {result, input} */
static const TableEntry tbl_acos[] = {
{MATH_PI / 2.0, 0.0},
{MATH_PI / 2.0, -0.0},
{0.0, 1.0},
{MATH_PI, -1.0}
};
static const TableEntry tbl_acosh[] = {
{0.0, 1.0}
};
static const TableEntry tbl_acospi[] = {
{0.5, 0.0},
{0.5, -0.0},
{0.0, 1.0},
{1.0, -1.0}
};
static const TableEntry tbl_asin[] = {
{0.0, 0.0},
{-0.0, -0.0},
{MATH_PI / 2.0, 1.0},
{-MATH_PI / 2.0, -1.0}
};
static const TableEntry tbl_asinh[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_asinpi[] = {
{0.0, 0.0},
{-0.0, -0.0},
{0.5, 1.0},
{-0.5, -1.0}
};
static const TableEntry tbl_atan[] = {
{0.0, 0.0},
{-0.0, -0.0},
{MATH_PI / 4.0, 1.0},
{-MATH_PI / 4.0, -1.0}
};
static const TableEntry tbl_atanh[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_atanpi[] = {
{0.0, 0.0},
{-0.0, -0.0},
{0.25, 1.0},
{-0.25, -1.0}
};
static const TableEntry tbl_cbrt[] = {
{0.0, 0.0},
{-0.0, -0.0},
{1.0, 1.0},
{-1.0, -1.0},
};
static const TableEntry tbl_cos[] = {
{1.0, 0.0},
{1.0, -0.0}
};
static const TableEntry tbl_cosh[] = {
{1.0, 0.0},
{1.0, -0.0}
};
static const TableEntry tbl_cospi[] = {
{1.0, 0.0},
{1.0, -0.0}
};
static const TableEntry tbl_erfc[] = {
{1.0, 0.0},
{1.0, -0.0}
};
static const TableEntry tbl_erf[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_exp[] = {
{1.0, 0.0},
{1.0, -0.0},
{MATH_E, 1.0}
};
static const TableEntry tbl_exp2[] = {
{1.0, 0.0},
{1.0, -0.0},
{2.0, 1.0}
};
static const TableEntry tbl_exp10[] = {
{1.0, 0.0},
{1.0, -0.0},
{10.0, 1.0}
};
static const TableEntry tbl_expm1[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_log[] = {
{0.0, 1.0},
{1.0, MATH_E}
};
static const TableEntry tbl_log2[] = {
{0.0, 1.0},
{1.0, 2.0}
};
static const TableEntry tbl_log10[] = {
{0.0, 1.0},
{1.0, 10.0}
};
static const TableEntry tbl_rsqrt[] = {
{1.0, 1.0},
{MATH_SQRT1_2, 2.0}
};
static const TableEntry tbl_sin[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_sinh[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_sinpi[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_sqrt[] = {
{0.0, 0.0},
{1.0, 1.0},
{MATH_SQRT2, 2.0}
};
static const TableEntry tbl_tan[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_tanh[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_tanpi[] = {
{0.0, 0.0},
{-0.0, -0.0}
};
static const TableEntry tbl_tgamma[] = {
{1.0, 1.0},
{1.0, 2.0},
{2.0, 3.0},
{6.0, 4.0}
};
static bool HasNative(AMDGPULibFunc::EFuncId id) {
switch(id) {
case AMDGPULibFunc::EI_DIVIDE:
case AMDGPULibFunc::EI_COS:
case AMDGPULibFunc::EI_EXP:
case AMDGPULibFunc::EI_EXP2:
case AMDGPULibFunc::EI_EXP10:
case AMDGPULibFunc::EI_LOG:
case AMDGPULibFunc::EI_LOG2:
case AMDGPULibFunc::EI_LOG10:
case AMDGPULibFunc::EI_POWR:
case AMDGPULibFunc::EI_RECIP:
case AMDGPULibFunc::EI_RSQRT:
case AMDGPULibFunc::EI_SIN:
case AMDGPULibFunc::EI_SINCOS:
case AMDGPULibFunc::EI_SQRT:
case AMDGPULibFunc::EI_TAN:
return true;
default:;
}
return false;
}
struct TableRef {
size_t size;
const TableEntry *table; // variable size: from 0 to (size - 1)
TableRef() : size(0), table(nullptr) {}
template <size_t N>
TableRef(const TableEntry (&tbl)[N]) : size(N), table(&tbl[0]) {}
};
static TableRef getOptTable(AMDGPULibFunc::EFuncId id) {
switch(id) {
case AMDGPULibFunc::EI_ACOS: return TableRef(tbl_acos);
case AMDGPULibFunc::EI_ACOSH: return TableRef(tbl_acosh);
case AMDGPULibFunc::EI_ACOSPI: return TableRef(tbl_acospi);
case AMDGPULibFunc::EI_ASIN: return TableRef(tbl_asin);
case AMDGPULibFunc::EI_ASINH: return TableRef(tbl_asinh);
case AMDGPULibFunc::EI_ASINPI: return TableRef(tbl_asinpi);
case AMDGPULibFunc::EI_ATAN: return TableRef(tbl_atan);
case AMDGPULibFunc::EI_ATANH: return TableRef(tbl_atanh);
case AMDGPULibFunc::EI_ATANPI: return TableRef(tbl_atanpi);
case AMDGPULibFunc::EI_CBRT: return TableRef(tbl_cbrt);
case AMDGPULibFunc::EI_NCOS:
case AMDGPULibFunc::EI_COS: return TableRef(tbl_cos);
case AMDGPULibFunc::EI_COSH: return TableRef(tbl_cosh);
case AMDGPULibFunc::EI_COSPI: return TableRef(tbl_cospi);
case AMDGPULibFunc::EI_ERFC: return TableRef(tbl_erfc);
case AMDGPULibFunc::EI_ERF: return TableRef(tbl_erf);
case AMDGPULibFunc::EI_EXP: return TableRef(tbl_exp);
case AMDGPULibFunc::EI_NEXP2:
case AMDGPULibFunc::EI_EXP2: return TableRef(tbl_exp2);
case AMDGPULibFunc::EI_EXP10: return TableRef(tbl_exp10);
case AMDGPULibFunc::EI_EXPM1: return TableRef(tbl_expm1);
case AMDGPULibFunc::EI_LOG: return TableRef(tbl_log);
case AMDGPULibFunc::EI_NLOG2:
case AMDGPULibFunc::EI_LOG2: return TableRef(tbl_log2);
case AMDGPULibFunc::EI_LOG10: return TableRef(tbl_log10);
case AMDGPULibFunc::EI_NRSQRT:
case AMDGPULibFunc::EI_RSQRT: return TableRef(tbl_rsqrt);
case AMDGPULibFunc::EI_NSIN:
case AMDGPULibFunc::EI_SIN: return TableRef(tbl_sin);
case AMDGPULibFunc::EI_SINH: return TableRef(tbl_sinh);
case AMDGPULibFunc::EI_SINPI: return TableRef(tbl_sinpi);
case AMDGPULibFunc::EI_NSQRT:
case AMDGPULibFunc::EI_SQRT: return TableRef(tbl_sqrt);
case AMDGPULibFunc::EI_TAN: return TableRef(tbl_tan);
case AMDGPULibFunc::EI_TANH: return TableRef(tbl_tanh);
case AMDGPULibFunc::EI_TANPI: return TableRef(tbl_tanpi);
case AMDGPULibFunc::EI_TGAMMA: return TableRef(tbl_tgamma);
default:;
}
return TableRef();
}
static inline int getVecSize(const AMDGPULibFunc& FInfo) {
return FInfo.getLeads()[0].VectorSize;
}
static inline AMDGPULibFunc::EType getArgType(const AMDGPULibFunc& FInfo) {
return (AMDGPULibFunc::EType)FInfo.getLeads()[0].ArgType;
}
FunctionCallee AMDGPULibCalls::getFunction(Module *M, const FuncInfo &fInfo) {
// If we are doing PreLinkOpt, the function is external. So it is safe to
// use getOrInsertFunction() at this stage.
return EnablePreLink ? AMDGPULibFunc::getOrInsertFunction(M, fInfo)
: AMDGPULibFunc::getFunction(M, fInfo);
}
bool AMDGPULibCalls::parseFunctionName(const StringRef& FMangledName,
FuncInfo *FInfo) {
return AMDGPULibFunc::parse(FMangledName, *FInfo);
}
bool AMDGPULibCalls::isUnsafeMath(const CallInst *CI) const {
if (auto Op = dyn_cast<FPMathOperator>(CI))
if (Op->isFast())
return true;
const Function *F = CI->getParent()->getParent();
Attribute Attr = F->getFnAttribute("unsafe-fp-math");
return Attr.getValueAsString() == "true";
}
bool AMDGPULibCalls::useNativeFunc(const StringRef F) const {
return AllNative || llvm::is_contained(UseNative, F);
}
void AMDGPULibCalls::initNativeFuncs() {
AllNative = useNativeFunc("all") ||
(UseNative.getNumOccurrences() && UseNative.size() == 1 &&
UseNative.begin()->empty());
}
bool AMDGPULibCalls::sincosUseNative(CallInst *aCI, const FuncInfo &FInfo) {
bool native_sin = useNativeFunc("sin");
bool native_cos = useNativeFunc("cos");
if (native_sin && native_cos) {
Module *M = aCI->getModule();
Value *opr0 = aCI->getArgOperand(0);
AMDGPULibFunc nf;
nf.getLeads()[0].ArgType = FInfo.getLeads()[0].ArgType;
nf.getLeads()[0].VectorSize = FInfo.getLeads()[0].VectorSize;
nf.setPrefix(AMDGPULibFunc::NATIVE);
nf.setId(AMDGPULibFunc::EI_SIN);
FunctionCallee sinExpr = getFunction(M, nf);
nf.setPrefix(AMDGPULibFunc::NATIVE);
nf.setId(AMDGPULibFunc::EI_COS);
FunctionCallee cosExpr = getFunction(M, nf);
if (sinExpr && cosExpr) {
Value *sinval = CallInst::Create(sinExpr, opr0, "splitsin", aCI);
Value *cosval = CallInst::Create(cosExpr, opr0, "splitcos", aCI);
new StoreInst(cosval, aCI->getArgOperand(1), aCI);
DEBUG_WITH_TYPE("usenative", dbgs() << "<useNative> replace " << *aCI
<< " with native version of sin/cos");
replaceCall(sinval);
return true;
}
}
return false;
}
bool AMDGPULibCalls::useNative(CallInst *aCI) {
CI = aCI;
Function *Callee = aCI->getCalledFunction();
FuncInfo FInfo;
if (!parseFunctionName(Callee->getName(), &FInfo) || !FInfo.isMangled() ||
FInfo.getPrefix() != AMDGPULibFunc::NOPFX ||
getArgType(FInfo) == AMDGPULibFunc::F64 || !HasNative(FInfo.getId()) ||
!(AllNative || useNativeFunc(FInfo.getName()))) {
return false;
}
if (FInfo.getId() == AMDGPULibFunc::EI_SINCOS)
return sincosUseNative(aCI, FInfo);
FInfo.setPrefix(AMDGPULibFunc::NATIVE);
FunctionCallee F = getFunction(aCI->getModule(), FInfo);
if (!F)
return false;
aCI->setCalledFunction(F);
DEBUG_WITH_TYPE("usenative", dbgs() << "<useNative> replace " << *aCI
<< " with native version");
return true;
}
// Clang emits call of __read_pipe_2 or __read_pipe_4 for OpenCL read_pipe
// builtin, with appended type size and alignment arguments, where 2 or 4
// indicates the original number of arguments. The library has optimized version
// of __read_pipe_2/__read_pipe_4 when the type size and alignment has the same
// power of 2 value. This function transforms __read_pipe_2 to __read_pipe_2_N
// for such cases where N is the size in bytes of the type (N = 1, 2, 4, 8, ...,
// 128). The same for __read_pipe_4, write_pipe_2, and write_pipe_4.
bool AMDGPULibCalls::fold_read_write_pipe(CallInst *CI, IRBuilder<> &B,
FuncInfo &FInfo) {
auto *Callee = CI->getCalledFunction();
if (!Callee->isDeclaration())
return false;
assert(Callee->hasName() && "Invalid read_pipe/write_pipe function");
auto *M = Callee->getParent();
auto &Ctx = M->getContext();
std::string Name = std::string(Callee->getName());
auto NumArg = CI->getNumArgOperands();
if (NumArg != 4 && NumArg != 6)
return false;
auto *PacketSize = CI->getArgOperand(NumArg - 2);
auto *PacketAlign = CI->getArgOperand(NumArg - 1);
if (!isa<ConstantInt>(PacketSize) || !isa<ConstantInt>(PacketAlign))
return false;
unsigned Size = cast<ConstantInt>(PacketSize)->getZExtValue();
Align Alignment = cast<ConstantInt>(PacketAlign)->getAlignValue();
if (Alignment != Size)
return false;
Type *PtrElemTy;
if (Size <= 8)
PtrElemTy = Type::getIntNTy(Ctx, Size * 8);
else
PtrElemTy = FixedVectorType::get(Type::getInt64Ty(Ctx), Size / 8);
unsigned PtrArgLoc = CI->getNumArgOperands() - 3;
auto PtrArg = CI->getArgOperand(PtrArgLoc);
unsigned PtrArgAS = PtrArg->getType()->getPointerAddressSpace();
auto *PtrTy = llvm::PointerType::get(PtrElemTy, PtrArgAS);
SmallVector<llvm::Type *, 6> ArgTys;
for (unsigned I = 0; I != PtrArgLoc; ++I)
ArgTys.push_back(CI->getArgOperand(I)->getType());
ArgTys.push_back(PtrTy);
Name = Name + "_" + std::to_string(Size);
auto *FTy = FunctionType::get(Callee->getReturnType(),
ArrayRef<Type *>(ArgTys), false);
AMDGPULibFunc NewLibFunc(Name, FTy);
FunctionCallee F = AMDGPULibFunc::getOrInsertFunction(M, NewLibFunc);
if (!F)
return false;
auto *BCast = B.CreatePointerCast(PtrArg, PtrTy);
SmallVector<Value *, 6> Args;
for (unsigned I = 0; I != PtrArgLoc; ++I)
Args.push_back(CI->getArgOperand(I));
Args.push_back(BCast);
auto *NCI = B.CreateCall(F, Args);
NCI->setAttributes(CI->getAttributes());
CI->replaceAllUsesWith(NCI);
CI->dropAllReferences();
CI->eraseFromParent();
return true;
}
// This function returns false if no change; return true otherwise.
bool AMDGPULibCalls::fold(CallInst *CI, AliasAnalysis *AA) {
this->CI = CI;
Function *Callee = CI->getCalledFunction();
// Ignore indirect calls.
if (Callee == 0) return false;
BasicBlock *BB = CI->getParent();
LLVMContext &Context = CI->getParent()->getContext();
IRBuilder<> B(Context);
// Set the builder to the instruction after the call.
B.SetInsertPoint(BB, CI->getIterator());
// Copy fast flags from the original call.
if (const FPMathOperator *FPOp = dyn_cast<const FPMathOperator>(CI))
B.setFastMathFlags(FPOp->getFastMathFlags());
switch (Callee->getIntrinsicID()) {
default:
break;
case Intrinsic::amdgcn_wavefrontsize:
return !EnablePreLink && fold_wavefrontsize(CI, B);
}
FuncInfo FInfo;
if (!parseFunctionName(Callee->getName(), &FInfo))
return false;
// Further check the number of arguments to see if they match.
if (CI->getNumArgOperands() != FInfo.getNumArgs())
return false;
if (TDOFold(CI, FInfo))
return true;
// Under unsafe-math, evaluate calls if possible.
// According to Brian Sumner, we can do this for all f32 function calls
// using host's double function calls.
if (isUnsafeMath(CI) && evaluateCall(CI, FInfo))
return true;
// Specilized optimizations for each function call
switch (FInfo.getId()) {
case AMDGPULibFunc::EI_RECIP:
// skip vector function
assert ((FInfo.getPrefix() == AMDGPULibFunc::NATIVE ||
FInfo.getPrefix() == AMDGPULibFunc::HALF) &&
"recip must be an either native or half function");
return (getVecSize(FInfo) != 1) ? false : fold_recip(CI, B, FInfo);
case AMDGPULibFunc::EI_DIVIDE:
// skip vector function
assert ((FInfo.getPrefix() == AMDGPULibFunc::NATIVE ||
FInfo.getPrefix() == AMDGPULibFunc::HALF) &&
"divide must be an either native or half function");
return (getVecSize(FInfo) != 1) ? false : fold_divide(CI, B, FInfo);
case AMDGPULibFunc::EI_POW:
case AMDGPULibFunc::EI_POWR:
case AMDGPULibFunc::EI_POWN:
return fold_pow(CI, B, FInfo);
case AMDGPULibFunc::EI_ROOTN:
// skip vector function
return (getVecSize(FInfo) != 1) ? false : fold_rootn(CI, B, FInfo);
case AMDGPULibFunc::EI_FMA:
case AMDGPULibFunc::EI_MAD:
case AMDGPULibFunc::EI_NFMA:
// skip vector function
return (getVecSize(FInfo) != 1) ? false : fold_fma_mad(CI, B, FInfo);
case AMDGPULibFunc::EI_SQRT:
return isUnsafeMath(CI) && fold_sqrt(CI, B, FInfo);
case AMDGPULibFunc::EI_COS:
case AMDGPULibFunc::EI_SIN:
if ((getArgType(FInfo) == AMDGPULibFunc::F32 ||
getArgType(FInfo) == AMDGPULibFunc::F64)
&& (FInfo.getPrefix() == AMDGPULibFunc::NOPFX))
return fold_sincos(CI, B, AA);
break;
case AMDGPULibFunc::EI_READ_PIPE_2:
case AMDGPULibFunc::EI_READ_PIPE_4:
case AMDGPULibFunc::EI_WRITE_PIPE_2:
case AMDGPULibFunc::EI_WRITE_PIPE_4:
return fold_read_write_pipe(CI, B, FInfo);
default:
break;
}
return false;
}
bool AMDGPULibCalls::TDOFold(CallInst *CI, const FuncInfo &FInfo) {
// Table-Driven optimization
const TableRef tr = getOptTable(FInfo.getId());
if (tr.size==0)
return false;
int const sz = (int)tr.size;
const TableEntry * const ftbl = tr.table;
Value *opr0 = CI->getArgOperand(0);
if (getVecSize(FInfo) > 1) {
if (ConstantDataVector *CV = dyn_cast<ConstantDataVector>(opr0)) {
SmallVector<double, 0> DVal;
for (int eltNo = 0; eltNo < getVecSize(FInfo); ++eltNo) {
ConstantFP *eltval = dyn_cast<ConstantFP>(
CV->getElementAsConstant((unsigned)eltNo));
assert(eltval && "Non-FP arguments in math function!");
bool found = false;
for (int i=0; i < sz; ++i) {
if (eltval->isExactlyValue(ftbl[i].input)) {
DVal.push_back(ftbl[i].result);
found = true;
break;
}
}
if (!found) {
// This vector constants not handled yet.
return false;
}
}
LLVMContext &context = CI->getParent()->getParent()->getContext();
Constant *nval;
if (getArgType(FInfo) == AMDGPULibFunc::F32) {
SmallVector<float, 0> FVal;
for (unsigned i = 0; i < DVal.size(); ++i) {
FVal.push_back((float)DVal[i]);
}
ArrayRef<float> tmp(FVal);
nval = ConstantDataVector::get(context, tmp);
} else { // F64
ArrayRef<double> tmp(DVal);
nval = ConstantDataVector::get(context, tmp);
}
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *nval << "\n");
replaceCall(nval);
return true;
}
} else {
// Scalar version
if (ConstantFP *CF = dyn_cast<ConstantFP>(opr0)) {
for (int i = 0; i < sz; ++i) {
if (CF->isExactlyValue(ftbl[i].input)) {
Value *nval = ConstantFP::get(CF->getType(), ftbl[i].result);
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *nval << "\n");
replaceCall(nval);
return true;
}
}
}
}
return false;
}
bool AMDGPULibCalls::replaceWithNative(CallInst *CI, const FuncInfo &FInfo) {
Module *M = CI->getModule();
if (getArgType(FInfo) != AMDGPULibFunc::F32 ||
FInfo.getPrefix() != AMDGPULibFunc::NOPFX ||
!HasNative(FInfo.getId()))
return false;
AMDGPULibFunc nf = FInfo;
nf.setPrefix(AMDGPULibFunc::NATIVE);
if (FunctionCallee FPExpr = getFunction(M, nf)) {
LLVM_DEBUG(dbgs() << "AMDIC: " << *CI << " ---> ");
CI->setCalledFunction(FPExpr);
LLVM_DEBUG(dbgs() << *CI << '\n');
return true;
}
return false;
}
// [native_]half_recip(c) ==> 1.0/c
bool AMDGPULibCalls::fold_recip(CallInst *CI, IRBuilder<> &B,
const FuncInfo &FInfo) {
Value *opr0 = CI->getArgOperand(0);
if (ConstantFP *CF = dyn_cast<ConstantFP>(opr0)) {
// Just create a normal div. Later, InstCombine will be able
// to compute the divide into a constant (avoid check float infinity
// or subnormal at this point).
Value *nval = B.CreateFDiv(ConstantFP::get(CF->getType(), 1.0),
opr0,
"recip2div");
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *nval << "\n");
replaceCall(nval);
return true;
}
return false;
}
// [native_]half_divide(x, c) ==> x/c
bool AMDGPULibCalls::fold_divide(CallInst *CI, IRBuilder<> &B,
const FuncInfo &FInfo) {
Value *opr0 = CI->getArgOperand(0);
Value *opr1 = CI->getArgOperand(1);
ConstantFP *CF0 = dyn_cast<ConstantFP>(opr0);
ConstantFP *CF1 = dyn_cast<ConstantFP>(opr1);
if ((CF0 && CF1) || // both are constants
(CF1 && (getArgType(FInfo) == AMDGPULibFunc::F32)))
// CF1 is constant && f32 divide
{
Value *nval1 = B.CreateFDiv(ConstantFP::get(opr1->getType(), 1.0),
opr1, "__div2recip");
Value *nval = B.CreateFMul(opr0, nval1, "__div2mul");
replaceCall(nval);
return true;
}
return false;
}
namespace llvm {
static double log2(double V) {
#if _XOPEN_SOURCE >= 600 || defined(_ISOC99_SOURCE) || _POSIX_C_SOURCE >= 200112L
return ::log2(V);
#else
return log(V) / numbers::ln2;
#endif
}
}
bool AMDGPULibCalls::fold_pow(CallInst *CI, IRBuilder<> &B,
const FuncInfo &FInfo) {
assert((FInfo.getId() == AMDGPULibFunc::EI_POW ||
FInfo.getId() == AMDGPULibFunc::EI_POWR ||
FInfo.getId() == AMDGPULibFunc::EI_POWN) &&
"fold_pow: encounter a wrong function call");
Value *opr0, *opr1;
ConstantFP *CF;
ConstantInt *CINT;
ConstantAggregateZero *CZero;
Type *eltType;
opr0 = CI->getArgOperand(0);
opr1 = CI->getArgOperand(1);
CZero = dyn_cast<ConstantAggregateZero>(opr1);
if (getVecSize(FInfo) == 1) {
eltType = opr0->getType();
CF = dyn_cast<ConstantFP>(opr1);
CINT = dyn_cast<ConstantInt>(opr1);
} else {
VectorType *VTy = dyn_cast<VectorType>(opr0->getType());
assert(VTy && "Oprand of vector function should be of vectortype");
eltType = VTy->getElementType();
ConstantDataVector *CDV = dyn_cast<ConstantDataVector>(opr1);
// Now, only Handle vector const whose elements have the same value.
CF = CDV ? dyn_cast_or_null<ConstantFP>(CDV->getSplatValue()) : nullptr;
CINT = CDV ? dyn_cast_or_null<ConstantInt>(CDV->getSplatValue()) : nullptr;
}
// No unsafe math , no constant argument, do nothing
if (!isUnsafeMath(CI) && !CF && !CINT && !CZero)
return false;
// 0x1111111 means that we don't do anything for this call.
int ci_opr1 = (CINT ? (int)CINT->getSExtValue() : 0x1111111);
if ((CF && CF->isZero()) || (CINT && ci_opr1 == 0) || CZero) {
// pow/powr/pown(x, 0) == 1
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> 1\n");
Constant *cnval = ConstantFP::get(eltType, 1.0);
if (getVecSize(FInfo) > 1) {
cnval = ConstantDataVector::getSplat(getVecSize(FInfo), cnval);
}
replaceCall(cnval);
return true;
}
if ((CF && CF->isExactlyValue(1.0)) || (CINT && ci_opr1 == 1)) {
// pow/powr/pown(x, 1.0) = x
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *opr0 << "\n");
replaceCall(opr0);
return true;
}
if ((CF && CF->isExactlyValue(2.0)) || (CINT && ci_opr1 == 2)) {
// pow/powr/pown(x, 2.0) = x*x
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *opr0 << " * " << *opr0
<< "\n");
Value *nval = B.CreateFMul(opr0, opr0, "__pow2");
replaceCall(nval);
return true;
}
if ((CF && CF->isExactlyValue(-1.0)) || (CINT && ci_opr1 == -1)) {
// pow/powr/pown(x, -1.0) = 1.0/x
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> 1 / " << *opr0 << "\n");
Constant *cnval = ConstantFP::get(eltType, 1.0);
if (getVecSize(FInfo) > 1) {
cnval = ConstantDataVector::getSplat(getVecSize(FInfo), cnval);
}
Value *nval = B.CreateFDiv(cnval, opr0, "__powrecip");
replaceCall(nval);
return true;
}
Module *M = CI->getModule();
if (CF && (CF->isExactlyValue(0.5) || CF->isExactlyValue(-0.5))) {
// pow[r](x, [-]0.5) = sqrt(x)
bool issqrt = CF->isExactlyValue(0.5);
if (FunctionCallee FPExpr =
getFunction(M, AMDGPULibFunc(issqrt ? AMDGPULibFunc::EI_SQRT
: AMDGPULibFunc::EI_RSQRT,
FInfo))) {
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> "
<< FInfo.getName().c_str() << "(" << *opr0 << ")\n");
Value *nval = CreateCallEx(B,FPExpr, opr0, issqrt ? "__pow2sqrt"
: "__pow2rsqrt");
replaceCall(nval);
return true;
}
}
if (!isUnsafeMath(CI))
return false;
// Unsafe Math optimization
// Remember that ci_opr1 is set if opr1 is integral
if (CF) {
double dval = (getArgType(FInfo) == AMDGPULibFunc::F32)
? (double)CF->getValueAPF().convertToFloat()
: CF->getValueAPF().convertToDouble();
int ival = (int)dval;
if ((double)ival == dval) {
ci_opr1 = ival;
} else
ci_opr1 = 0x11111111;
}
// pow/powr/pown(x, c) = [1/](x*x*..x); where
// trunc(c) == c && the number of x == c && |c| <= 12
unsigned abs_opr1 = (ci_opr1 < 0) ? -ci_opr1 : ci_opr1;
if (abs_opr1 <= 12) {
Constant *cnval;
Value *nval;
if (abs_opr1 == 0) {
cnval = ConstantFP::get(eltType, 1.0);
if (getVecSize(FInfo) > 1) {
cnval = ConstantDataVector::getSplat(getVecSize(FInfo), cnval);
}
nval = cnval;
} else {
Value *valx2 = nullptr;
nval = nullptr;
while (abs_opr1 > 0) {
valx2 = valx2 ? B.CreateFMul(valx2, valx2, "__powx2") : opr0;
if (abs_opr1 & 1) {
nval = nval ? B.CreateFMul(nval, valx2, "__powprod") : valx2;
}
abs_opr1 >>= 1;
}
}
if (ci_opr1 < 0) {
cnval = ConstantFP::get(eltType, 1.0);
if (getVecSize(FInfo) > 1) {
cnval = ConstantDataVector::getSplat(getVecSize(FInfo), cnval);
}
nval = B.CreateFDiv(cnval, nval, "__1powprod");
}
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> "
<< ((ci_opr1 < 0) ? "1/prod(" : "prod(") << *opr0
<< ")\n");
replaceCall(nval);
return true;
}
// powr ---> exp2(y * log2(x))
// pown/pow ---> powr(fabs(x), y) | (x & ((int)y << 31))
FunctionCallee ExpExpr =
getFunction(M, AMDGPULibFunc(AMDGPULibFunc::EI_EXP2, FInfo));
if (!ExpExpr)
return false;
bool needlog = false;
bool needabs = false;
bool needcopysign = false;
Constant *cnval = nullptr;
if (getVecSize(FInfo) == 1) {
CF = dyn_cast<ConstantFP>(opr0);
if (CF) {
double V = (getArgType(FInfo) == AMDGPULibFunc::F32)
? (double)CF->getValueAPF().convertToFloat()
: CF->getValueAPF().convertToDouble();
V = log2(std::abs(V));
cnval = ConstantFP::get(eltType, V);
needcopysign = (FInfo.getId() != AMDGPULibFunc::EI_POWR) &&
CF->isNegative();
} else {
needlog = true;
needcopysign = needabs = FInfo.getId() != AMDGPULibFunc::EI_POWR &&
(!CF || CF->isNegative());
}
} else {
ConstantDataVector *CDV = dyn_cast<ConstantDataVector>(opr0);
if (!CDV) {
needlog = true;
needcopysign = needabs = FInfo.getId() != AMDGPULibFunc::EI_POWR;
} else {
assert ((int)CDV->getNumElements() == getVecSize(FInfo) &&
"Wrong vector size detected");
SmallVector<double, 0> DVal;
for (int i=0; i < getVecSize(FInfo); ++i) {
double V = (getArgType(FInfo) == AMDGPULibFunc::F32)
? (double)CDV->getElementAsFloat(i)
: CDV->getElementAsDouble(i);
if (V < 0.0) needcopysign = true;
V = log2(std::abs(V));
DVal.push_back(V);
}
if (getArgType(FInfo) == AMDGPULibFunc::F32) {
SmallVector<float, 0> FVal;
for (unsigned i=0; i < DVal.size(); ++i) {
FVal.push_back((float)DVal[i]);
}
ArrayRef<float> tmp(FVal);
cnval = ConstantDataVector::get(M->getContext(), tmp);
} else {
ArrayRef<double> tmp(DVal);
cnval = ConstantDataVector::get(M->getContext(), tmp);
}
}
}
if (needcopysign && (FInfo.getId() == AMDGPULibFunc::EI_POW)) {
// We cannot handle corner cases for a general pow() function, give up
// unless y is a constant integral value. Then proceed as if it were pown.
if (getVecSize(FInfo) == 1) {
if (const ConstantFP *CF = dyn_cast<ConstantFP>(opr1)) {
double y = (getArgType(FInfo) == AMDGPULibFunc::F32)
? (double)CF->getValueAPF().convertToFloat()
: CF->getValueAPF().convertToDouble();
if (y != (double)(int64_t)y)
return false;
} else
return false;
} else {
if (const ConstantDataVector *CDV = dyn_cast<ConstantDataVector>(opr1)) {
for (int i=0; i < getVecSize(FInfo); ++i) {
double y = (getArgType(FInfo) == AMDGPULibFunc::F32)
? (double)CDV->getElementAsFloat(i)
: CDV->getElementAsDouble(i);
if (y != (double)(int64_t)y)
return false;
}
} else
return false;
}
}
Value *nval;
if (needabs) {
FunctionCallee AbsExpr =
getFunction(M, AMDGPULibFunc(AMDGPULibFunc::EI_FABS, FInfo));
if (!AbsExpr)
return false;
nval = CreateCallEx(B, AbsExpr, opr0, "__fabs");
} else {
nval = cnval ? cnval : opr0;
}
if (needlog) {
FunctionCallee LogExpr =
getFunction(M, AMDGPULibFunc(AMDGPULibFunc::EI_LOG2, FInfo));
if (!LogExpr)
return false;
nval = CreateCallEx(B,LogExpr, nval, "__log2");
}
if (FInfo.getId() == AMDGPULibFunc::EI_POWN) {
// convert int(32) to fp(f32 or f64)
opr1 = B.CreateSIToFP(opr1, nval->getType(), "pownI2F");
}
nval = B.CreateFMul(opr1, nval, "__ylogx");
nval = CreateCallEx(B,ExpExpr, nval, "__exp2");
if (needcopysign) {
Value *opr_n;
Type* rTy = opr0->getType();
Type* nTyS = eltType->isDoubleTy() ? B.getInt64Ty() : B.getInt32Ty();
Type *nTy = nTyS;
if (const auto *vTy = dyn_cast<FixedVectorType>(rTy))
nTy = FixedVectorType::get(nTyS, vTy);
unsigned size = nTy->getScalarSizeInBits();
opr_n = CI->getArgOperand(1);
if (opr_n->getType()->isIntegerTy())
opr_n = B.CreateZExtOrBitCast(opr_n, nTy, "__ytou");
else
opr_n = B.CreateFPToSI(opr1, nTy, "__ytou");
Value *sign = B.CreateShl(opr_n, size-1, "__yeven");
sign = B.CreateAnd(B.CreateBitCast(opr0, nTy), sign, "__pow_sign");
nval = B.CreateOr(B.CreateBitCast(nval, nTy), sign);
nval = B.CreateBitCast(nval, opr0->getType());
}
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> "
<< "exp2(" << *opr1 << " * log2(" << *opr0 << "))\n");
replaceCall(nval);
return true;
}
bool AMDGPULibCalls::fold_rootn(CallInst *CI, IRBuilder<> &B,
const FuncInfo &FInfo) {
Value *opr0 = CI->getArgOperand(0);
Value *opr1 = CI->getArgOperand(1);
ConstantInt *CINT = dyn_cast<ConstantInt>(opr1);
if (!CINT) {
return false;
}
int ci_opr1 = (int)CINT->getSExtValue();
if (ci_opr1 == 1) { // rootn(x, 1) = x
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *opr0 << "\n");
replaceCall(opr0);
return true;
}
if (ci_opr1 == 2) { // rootn(x, 2) = sqrt(x)
Module *M = CI->getModule();
if (FunctionCallee FPExpr =
getFunction(M, AMDGPULibFunc(AMDGPULibFunc::EI_SQRT, FInfo))) {
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> sqrt(" << *opr0 << ")\n");
Value *nval = CreateCallEx(B,FPExpr, opr0, "__rootn2sqrt");
replaceCall(nval);
return true;
}
} else if (ci_opr1 == 3) { // rootn(x, 3) = cbrt(x)
Module *M = CI->getModule();
if (FunctionCallee FPExpr =
getFunction(M, AMDGPULibFunc(AMDGPULibFunc::EI_CBRT, FInfo))) {
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> cbrt(" << *opr0 << ")\n");
Value *nval = CreateCallEx(B,FPExpr, opr0, "__rootn2cbrt");
replaceCall(nval);
return true;
}
} else if (ci_opr1 == -1) { // rootn(x, -1) = 1.0/x
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> 1.0 / " << *opr0 << "\n");
Value *nval = B.CreateFDiv(ConstantFP::get(opr0->getType(), 1.0),
opr0,
"__rootn2div");
replaceCall(nval);
return true;
} else if (ci_opr1 == -2) { // rootn(x, -2) = rsqrt(x)
Module *M = CI->getModule();
if (FunctionCallee FPExpr =
getFunction(M, AMDGPULibFunc(AMDGPULibFunc::EI_RSQRT, FInfo))) {
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> rsqrt(" << *opr0
<< ")\n");
Value *nval = CreateCallEx(B,FPExpr, opr0, "__rootn2rsqrt");
replaceCall(nval);
return true;
}
}
return false;
}
bool AMDGPULibCalls::fold_fma_mad(CallInst *CI, IRBuilder<> &B,
const FuncInfo &FInfo) {
Value *opr0 = CI->getArgOperand(0);
Value *opr1 = CI->getArgOperand(1);
Value *opr2 = CI->getArgOperand(2);
ConstantFP *CF0 = dyn_cast<ConstantFP>(opr0);
ConstantFP *CF1 = dyn_cast<ConstantFP>(opr1);
if ((CF0 && CF0->isZero()) || (CF1 && CF1->isZero())) {
// fma/mad(a, b, c) = c if a=0 || b=0
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *opr2 << "\n");
replaceCall(opr2);
return true;
}
if (CF0 && CF0->isExactlyValue(1.0f)) {
// fma/mad(a, b, c) = b+c if a=1
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *opr1 << " + " << *opr2
<< "\n");
Value *nval = B.CreateFAdd(opr1, opr2, "fmaadd");
replaceCall(nval);
return true;
}
if (CF1 && CF1->isExactlyValue(1.0f)) {
// fma/mad(a, b, c) = a+c if b=1
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *opr0 << " + " << *opr2
<< "\n");
Value *nval = B.CreateFAdd(opr0, opr2, "fmaadd");
replaceCall(nval);
return true;
}
if (ConstantFP *CF = dyn_cast<ConstantFP>(opr2)) {
if (CF->isZero()) {
// fma/mad(a, b, c) = a*b if c=0
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> " << *opr0 << " * "
<< *opr1 << "\n");
Value *nval = B.CreateFMul(opr0, opr1, "fmamul");
replaceCall(nval);
return true;
}
}
return false;
}
// Get a scalar native builtin signle argument FP function
FunctionCallee AMDGPULibCalls::getNativeFunction(Module *M,
const FuncInfo &FInfo) {
if (getArgType(FInfo) == AMDGPULibFunc::F64 || !HasNative(FInfo.getId()))
return nullptr;
FuncInfo nf = FInfo;
nf.setPrefix(AMDGPULibFunc::NATIVE);
return getFunction(M, nf);
}
// fold sqrt -> native_sqrt (x)
bool AMDGPULibCalls::fold_sqrt(CallInst *CI, IRBuilder<> &B,
const FuncInfo &FInfo) {
if (getArgType(FInfo) == AMDGPULibFunc::F32 && (getVecSize(FInfo) == 1) &&
(FInfo.getPrefix() != AMDGPULibFunc::NATIVE)) {
if (FunctionCallee FPExpr = getNativeFunction(
CI->getModule(), AMDGPULibFunc(AMDGPULibFunc::EI_SQRT, FInfo))) {
Value *opr0 = CI->getArgOperand(0);
LLVM_DEBUG(errs() << "AMDIC: " << *CI << " ---> "
<< "sqrt(" << *opr0 << ")\n");
Value *nval = CreateCallEx(B,FPExpr, opr0, "__sqrt");
replaceCall(nval);
return true;
}
}
return false;
}
// fold sin, cos -> sincos.
bool AMDGPULibCalls::fold_sincos(CallInst *CI, IRBuilder<> &B,
AliasAnalysis *AA) {
AMDGPULibFunc fInfo;
if (!AMDGPULibFunc::parse(CI->getCalledFunction()->getName(), fInfo))
return false;
assert(fInfo.getId() == AMDGPULibFunc::EI_SIN ||
fInfo.getId() == AMDGPULibFunc::EI_COS);
bool const isSin = fInfo.getId() == AMDGPULibFunc::EI_SIN;
Value *CArgVal = CI->getArgOperand(0);
BasicBlock * const CBB = CI->getParent();
int const MaxScan = 30;
{ // fold in load value.
LoadInst *LI = dyn_cast<LoadInst>(CArgVal);
if (LI && LI->getParent() == CBB) {
BasicBlock::iterator BBI = LI->getIterator();
Value *AvailableVal = FindAvailableLoadedValue(LI, CBB, BBI, MaxScan, AA);
if (AvailableVal) {
CArgVal->replaceAllUsesWith(AvailableVal);
if (CArgVal->getNumUses() == 0)
LI->eraseFromParent();
CArgVal = CI->getArgOperand(0);
}
}
}
Module *M = CI->getModule();
fInfo.setId(isSin ? AMDGPULibFunc::EI_COS : AMDGPULibFunc::EI_SIN);
std::string const PairName = fInfo.mangle();
CallInst *UI = nullptr;
for (User* U : CArgVal->users()) {
CallInst *XI = dyn_cast_or_null<CallInst>(U);
if (!XI || XI == CI || XI->getParent() != CBB)
continue;
Function *UCallee = XI->getCalledFunction();
if (!UCallee || !UCallee->getName().equals(PairName))
continue;
BasicBlock::iterator BBI = CI->getIterator();
if (BBI == CI->getParent()->begin())
break;
--BBI;
for (int I = MaxScan; I > 0 && BBI != CBB->begin(); --BBI, --I) {
if (cast<Instruction>(BBI) == XI) {
UI = XI;
break;
}
}
if (UI) break;
}
if (!UI) return false;
// Merge the sin and cos.
// for OpenCL 2.0 we have only generic implementation of sincos
// function.
AMDGPULibFunc nf(AMDGPULibFunc::EI_SINCOS, fInfo);
nf.getLeads()[0].PtrKind = AMDGPULibFunc::getEPtrKindFromAddrSpace(AMDGPUAS::FLAT_ADDRESS);
FunctionCallee Fsincos = getFunction(M, nf);
if (!Fsincos) return false;
BasicBlock::iterator ItOld = B.GetInsertPoint();
AllocaInst *Alloc = insertAlloca(UI, B, "__sincos_");
B.SetInsertPoint(UI);
Value *P = Alloc;
Type *PTy = Fsincos.getFunctionType()->getParamType(1);
// The allocaInst allocates the memory in private address space. This need
// to be bitcasted to point to the address space of cos pointer type.
// In OpenCL 2.0 this is generic, while in 1.2 that is private.
if (PTy->getPointerAddressSpace() != AMDGPUAS::PRIVATE_ADDRESS)
P = B.CreateAddrSpaceCast(Alloc, PTy);
CallInst *Call = CreateCallEx2(B, Fsincos, UI->getArgOperand(0), P);
LLVM_DEBUG(errs() << "AMDIC: fold_sincos (" << *CI << ", " << *UI << ") with "
<< *Call << "\n");
if (!isSin) { // CI->cos, UI->sin
B.SetInsertPoint(&*ItOld);
UI->replaceAllUsesWith(&*Call);
Instruction *Reload = B.CreateLoad(Alloc->getAllocatedType(), Alloc);
CI->replaceAllUsesWith(Reload);
UI->eraseFromParent();
CI->eraseFromParent();
} else { // CI->sin, UI->cos
Instruction *Reload = B.CreateLoad(Alloc->getAllocatedType(), Alloc);
UI->replaceAllUsesWith(Reload);
CI->replaceAllUsesWith(Call);
UI->eraseFromParent();
CI->eraseFromParent();
}
return true;
}
bool AMDGPULibCalls::fold_wavefrontsize(CallInst *CI, IRBuilder<> &B) {
if (!TM)
return false;
StringRef CPU = TM->getTargetCPU();
StringRef Features = TM->getTargetFeatureString();
if ((CPU.empty() || CPU.equals_lower("generic")) &&
(Features.empty() ||
Features.find_lower("wavefrontsize") == StringRef::npos))
return false;
Function *F = CI->getParent()->getParent();
const GCNSubtarget &ST = TM->getSubtarget<GCNSubtarget>(*F);
unsigned N = ST.getWavefrontSize();
LLVM_DEBUG(errs() << "AMDIC: fold_wavefrontsize (" << *CI << ") with "
<< N << "\n");
CI->replaceAllUsesWith(ConstantInt::get(B.getInt32Ty(), N));
CI->eraseFromParent();
return true;
}
// Get insertion point at entry.
BasicBlock::iterator AMDGPULibCalls::getEntryIns(CallInst * UI) {
Function * Func = UI->getParent()->getParent();
BasicBlock * BB = &Func->getEntryBlock();
assert(BB && "Entry block not found!");
BasicBlock::iterator ItNew = BB->begin();
return ItNew;
}
// Insert a AllocsInst at the beginning of function entry block.
AllocaInst* AMDGPULibCalls::insertAlloca(CallInst *UI, IRBuilder<> &B,
const char *prefix) {
BasicBlock::iterator ItNew = getEntryIns(UI);
Function *UCallee = UI->getCalledFunction();
Type *RetType = UCallee->getReturnType();
B.SetInsertPoint(&*ItNew);
AllocaInst *Alloc = B.CreateAlloca(RetType, 0,
std::string(prefix) + UI->getName());
Alloc->setAlignment(
Align(UCallee->getParent()->getDataLayout().getTypeAllocSize(RetType)));
return Alloc;
}
bool AMDGPULibCalls::evaluateScalarMathFunc(FuncInfo &FInfo,
double& Res0, double& Res1,
Constant *copr0, Constant *copr1,
Constant *copr2) {
// By default, opr0/opr1/opr3 holds values of float/double type.
// If they are not float/double, each function has to its
// operand separately.
double opr0=0.0, opr1=0.0, opr2=0.0;
ConstantFP *fpopr0 = dyn_cast_or_null<ConstantFP>(copr0);
ConstantFP *fpopr1 = dyn_cast_or_null<ConstantFP>(copr1);
ConstantFP *fpopr2 = dyn_cast_or_null<ConstantFP>(copr2);
if (fpopr0) {
opr0 = (getArgType(FInfo) == AMDGPULibFunc::F64)
? fpopr0->getValueAPF().convertToDouble()
: (double)fpopr0->getValueAPF().convertToFloat();
}
if (fpopr1) {
opr1 = (getArgType(FInfo) == AMDGPULibFunc::F64)
? fpopr1->getValueAPF().convertToDouble()
: (double)fpopr1->getValueAPF().convertToFloat();
}
if (fpopr2) {
opr2 = (getArgType(FInfo) == AMDGPULibFunc::F64)
? fpopr2->getValueAPF().convertToDouble()
: (double)fpopr2->getValueAPF().convertToFloat();
}
switch (FInfo.getId()) {
default : return false;
case AMDGPULibFunc::EI_ACOS:
Res0 = acos(opr0);
return true;
case AMDGPULibFunc::EI_ACOSH:
// acosh(x) == log(x + sqrt(x*x - 1))
Res0 = log(opr0 + sqrt(opr0*opr0 - 1.0));
return true;
case AMDGPULibFunc::EI_ACOSPI:
Res0 = acos(opr0) / MATH_PI;
return true;
case AMDGPULibFunc::EI_ASIN:
Res0 = asin(opr0);
return true;
case AMDGPULibFunc::EI_ASINH:
// asinh(x) == log(x + sqrt(x*x + 1))
Res0 = log(opr0 + sqrt(opr0*opr0 + 1.0));
return true;
case AMDGPULibFunc::EI_ASINPI:
Res0 = asin(opr0) / MATH_PI;
return true;
case AMDGPULibFunc::EI_ATAN:
Res0 = atan(opr0);
return true;
case AMDGPULibFunc::EI_ATANH:
// atanh(x) == (log(x+1) - log(x-1))/2;
Res0 = (log(opr0 + 1.0) - log(opr0 - 1.0))/2.0;
return true;
case AMDGPULibFunc::EI_ATANPI:
Res0 = atan(opr0) / MATH_PI;
return true;
case AMDGPULibFunc::EI_CBRT:
Res0 = (opr0 < 0.0) ? -pow(-opr0, 1.0/3.0) : pow(opr0, 1.0/3.0);
return true;
case AMDGPULibFunc::EI_COS:
Res0 = cos(opr0);
return true;
case AMDGPULibFunc::EI_COSH:
Res0 = cosh(opr0);
return true;
case AMDGPULibFunc::EI_COSPI:
Res0 = cos(MATH_PI * opr0);
return true;
case AMDGPULibFunc::EI_EXP:
Res0 = exp(opr0);
return true;
case AMDGPULibFunc::EI_EXP2:
Res0 = pow(2.0, opr0);
return true;
case AMDGPULibFunc::EI_EXP10:
Res0 = pow(10.0, opr0);
return true;
case AMDGPULibFunc::EI_EXPM1:
Res0 = exp(opr0) - 1.0;
return true;
case AMDGPULibFunc::EI_LOG:
Res0 = log(opr0);
return true;
case AMDGPULibFunc::EI_LOG2:
Res0 = log(opr0) / log(2.0);
return true;
case AMDGPULibFunc::EI_LOG10:
Res0 = log(opr0) / log(10.0);
return true;
case AMDGPULibFunc::EI_RSQRT:
Res0 = 1.0 / sqrt(opr0);
return true;
case AMDGPULibFunc::EI_SIN:
Res0 = sin(opr0);
return true;
case AMDGPULibFunc::EI_SINH:
Res0 = sinh(opr0);
return true;
case AMDGPULibFunc::EI_SINPI:
Res0 = sin(MATH_PI * opr0);
return true;
case AMDGPULibFunc::EI_SQRT:
Res0 = sqrt(opr0);
return true;
case AMDGPULibFunc::EI_TAN:
Res0 = tan(opr0);
return true;
case AMDGPULibFunc::EI_TANH:
Res0 = tanh(opr0);
return true;
case AMDGPULibFunc::EI_TANPI:
Res0 = tan(MATH_PI * opr0);
return true;
case AMDGPULibFunc::EI_RECIP:
Res0 = 1.0 / opr0;
return true;
// two-arg functions
case AMDGPULibFunc::EI_DIVIDE:
Res0 = opr0 / opr1;
return true;
case AMDGPULibFunc::EI_POW:
case AMDGPULibFunc::EI_POWR:
Res0 = pow(opr0, opr1);
return true;
case AMDGPULibFunc::EI_POWN: {
if (ConstantInt *iopr1 = dyn_cast_or_null<ConstantInt>(copr1)) {
double val = (double)iopr1->getSExtValue();
Res0 = pow(opr0, val);
return true;
}
return false;
}
case AMDGPULibFunc::EI_ROOTN: {
if (ConstantInt *iopr1 = dyn_cast_or_null<ConstantInt>(copr1)) {
double val = (double)iopr1->getSExtValue();
Res0 = pow(opr0, 1.0 / val);
return true;
}
return false;
}
// with ptr arg
case AMDGPULibFunc::EI_SINCOS:
Res0 = sin(opr0);
Res1 = cos(opr0);
return true;
// three-arg functions
case AMDGPULibFunc::EI_FMA:
case AMDGPULibFunc::EI_MAD:
Res0 = opr0 * opr1 + opr2;
return true;
}
return false;
}
bool AMDGPULibCalls::evaluateCall(CallInst *aCI, FuncInfo &FInfo) {
int numArgs = (int)aCI->getNumArgOperands();
if (numArgs > 3)
return false;
Constant *copr0 = nullptr;
Constant *copr1 = nullptr;
Constant *copr2 = nullptr;
if (numArgs > 0) {
if ((copr0 = dyn_cast<Constant>(aCI->getArgOperand(0))) == nullptr)
return false;
}
if (numArgs > 1) {
if ((copr1 = dyn_cast<Constant>(aCI->getArgOperand(1))) == nullptr) {
if (FInfo.getId() != AMDGPULibFunc::EI_SINCOS)
return false;
}
}
if (numArgs > 2) {
if ((copr2 = dyn_cast<Constant>(aCI->getArgOperand(2))) == nullptr)
return false;
}
// At this point, all arguments to aCI are constants.
// max vector size is 16, and sincos will generate two results.
double DVal0[16], DVal1[16];
bool hasTwoResults = (FInfo.getId() == AMDGPULibFunc::EI_SINCOS);
if (getVecSize(FInfo) == 1) {
if (!evaluateScalarMathFunc(FInfo, DVal0[0],
DVal1[0], copr0, copr1, copr2)) {
return false;
}
} else {
ConstantDataVector *CDV0 = dyn_cast_or_null<ConstantDataVector>(copr0);
ConstantDataVector *CDV1 = dyn_cast_or_null<ConstantDataVector>(copr1);
ConstantDataVector *CDV2 = dyn_cast_or_null<ConstantDataVector>(copr2);
for (int i=0; i < getVecSize(FInfo); ++i) {
Constant *celt0 = CDV0 ? CDV0->getElementAsConstant(i) : nullptr;
Constant *celt1 = CDV1 ? CDV1->getElementAsConstant(i) : nullptr;
Constant *celt2 = CDV2 ? CDV2->getElementAsConstant(i) : nullptr;
if (!evaluateScalarMathFunc(FInfo, DVal0[i],
DVal1[i], celt0, celt1, celt2)) {
return false;
}
}
}
LLVMContext &context = CI->getParent()->getParent()->getContext();
Constant *nval0, *nval1;
if (getVecSize(FInfo) == 1) {
nval0 = ConstantFP::get(CI->getType(), DVal0[0]);
if (hasTwoResults)
nval1 = ConstantFP::get(CI->getType(), DVal1[0]);
} else {
if (getArgType(FInfo) == AMDGPULibFunc::F32) {
SmallVector <float, 0> FVal0, FVal1;
for (int i=0; i < getVecSize(FInfo); ++i)
FVal0.push_back((float)DVal0[i]);
ArrayRef<float> tmp0(FVal0);
nval0 = ConstantDataVector::get(context, tmp0);
if (hasTwoResults) {
for (int i=0; i < getVecSize(FInfo); ++i)
FVal1.push_back((float)DVal1[i]);
ArrayRef<float> tmp1(FVal1);
nval1 = ConstantDataVector::get(context, tmp1);
}
} else {
ArrayRef<double> tmp0(DVal0);
nval0 = ConstantDataVector::get(context, tmp0);
if (hasTwoResults) {
ArrayRef<double> tmp1(DVal1);
nval1 = ConstantDataVector::get(context, tmp1);
}
}
}
if (hasTwoResults) {
// sincos
assert(FInfo.getId() == AMDGPULibFunc::EI_SINCOS &&
"math function with ptr arg not supported yet");
new StoreInst(nval1, aCI->getArgOperand(1), aCI);
}
replaceCall(nval0);
return true;
}
// Public interface to the Simplify LibCalls pass.
FunctionPass *llvm::createAMDGPUSimplifyLibCallsPass(const TargetMachine *TM) {
return new AMDGPUSimplifyLibCalls(TM);
}
FunctionPass *llvm::createAMDGPUUseNativeCallsPass() {
return new AMDGPUUseNativeCalls();
}
bool AMDGPUSimplifyLibCalls::runOnFunction(Function &F) {
if (skipFunction(F))
return false;
bool Changed = false;
auto AA = &getAnalysis<AAResultsWrapperPass>().getAAResults();
LLVM_DEBUG(dbgs() << "AMDIC: process function ";
F.printAsOperand(dbgs(), false, F.getParent()); dbgs() << '\n';);
for (auto &BB : F) {
for (BasicBlock::iterator I = BB.begin(), E = BB.end(); I != E; ) {
// Ignore non-calls.
CallInst *CI = dyn_cast<CallInst>(I);
++I;
// Ignore intrinsics that do not become real instructions.
if (!CI || isa<DbgInfoIntrinsic>(CI) || CI->isLifetimeStartOrEnd())
continue;
// Ignore indirect calls.
Function *Callee = CI->getCalledFunction();
if (Callee == 0) continue;
LLVM_DEBUG(dbgs() << "AMDIC: try folding " << *CI << "\n";
dbgs().flush());
if(Simplifier.fold(CI, AA))
Changed = true;
}
}
return Changed;
}
bool AMDGPUUseNativeCalls::runOnFunction(Function &F) {
if (skipFunction(F) || UseNative.empty())
return false;
bool Changed = false;
for (auto &BB : F) {
for (BasicBlock::iterator I = BB.begin(), E = BB.end(); I != E; ) {
// Ignore non-calls.
CallInst *CI = dyn_cast<CallInst>(I);
++I;
if (!CI) continue;
// Ignore indirect calls.
Function *Callee = CI->getCalledFunction();
if (Callee == 0) continue;
if(Simplifier.useNative(CI))
Changed = true;
}
}
return Changed;
}
Reference in New Issue View Git Blame Copy Permalink