LLVM: lib/Target/AMDGPU/AMDGPUTargetTransformInfo.cpp File Reference (original) (raw)

Go to the source code of this file.

Functions
static bool	dependsOnLocalPhi (const Loop L, const Value Cond, unsigned Depth=0)

static bool	intrinsicHasPackedVectorBenefit (Intrinsic::ID ID)

static unsigned	adjustInliningThresholdUsingCallee (const CallBase CB, const SITargetLowering TLI, const GCNTTIImpl *TTIImpl)

static unsigned	getCallArgsTotalAllocaSize (const CallBase *CB, const DataLayout &DL)

Variables
static cl::opt< unsigned >	UnrollThresholdPrivate ("amdgpu-unroll-threshold-private", cl::desc("Unroll threshold for AMDGPU if private memory used in a loop"), cl::init(2700), cl::Hidden)

static cl::opt< unsigned >	UnrollThresholdLocal ("amdgpu-unroll-threshold-local", cl::desc("Unroll threshold for AMDGPU if local memory used in a loop"), cl::init(1000), cl::Hidden)

static cl::opt< unsigned >	UnrollThresholdIf ("amdgpu-unroll-threshold-if", cl::desc("Unroll threshold increment for AMDGPU for each if statement inside loop"), cl::init(200), cl::Hidden)

static cl::opt< bool >	UnrollRuntimeLocal ("amdgpu-unroll-runtime-local", cl::desc("Allow runtime unroll for AMDGPU if local memory used in a loop"), cl::init(true), cl::Hidden)

static cl::opt< unsigned >	UnrollMaxBlockToAnalyze ("amdgpu-unroll-max-block-to-analyze", cl::desc("Inner loop block size threshold to analyze in unroll for AMDGPU"), cl::init(32), cl::Hidden)

static cl::opt< unsigned >	ArgAllocaCost ("amdgpu-inline-arg-alloca-cost", cl::Hidden, cl::init(4000), cl::desc("Cost of alloca argument"))

static cl::opt< unsigned >	ArgAllocaCutoff ("amdgpu-inline-arg-alloca-cutoff", cl::Hidden, cl::init(256), cl::desc("Maximum alloca size to use for inline cost"))

static cl::opt< size_t >	InlineMaxBB ("amdgpu-inline-max-bb", cl::Hidden, cl::init(1100), cl::desc("Maximum number of BBs allowed in a function after inlining" " (compile time constraint)"))

static cl::opt< unsigned >	MemcpyLoopUnroll ("amdgpu-memcpy-loop-unroll", cl::desc("Unroll factor (affecting 4x32-bit operations) to use for memory " "operations when lowering memcpy as a loop"), cl::init(16), cl::Hidden)

◆ DEBUG_TYPE

#define DEBUG_TYPE "AMDGPUtti"

◆ adjustInliningThresholdUsingCallee()

Definition at line 1264 of file AMDGPUTargetTransformInfo.cpp.

References A, llvm::CallBase::args(), llvm::ComputeValueVTs(), DL, llvm::CallBase::getArgOperandNo(), llvm::CallBase::getCallingConv(), llvm::Value::getContext(), llvm::TargetTransformInfoImplBase::getDataLayout(), llvm::InlineConstants::getInstrCost(), llvm::Type::getInt32Ty(), llvm::SITargetLowering::getNumRegistersForCallingConv(), llvm::InstructionCost::getValue(), llvm::AMDGPU::isArgPassedInSGPR(), llvm::AMDGPUAS::PRIVATE_ADDRESS, and llvm::TargetTransformInfo::TCK_SizeAndLatency.

Referenced by llvm::GCNTTIImpl::adjustInliningThreshold().

◆ dependsOnLocalPhi()

◆ getCallArgsTotalAllocaSize()

◆ intrinsicHasPackedVectorBenefit()

◆ ArgAllocaCost

cl::opt< unsigned > ArgAllocaCost("amdgpu-inline-arg-alloca-cost", cl::Hidden, cl::init(4000), cl::desc("Cost of alloca argument")) ( "amdgpu-inline-arg-alloca-cost" , cl::Hidden , cl::init(4000) , cl::desc("Cost of alloca argument") )	static

◆ ArgAllocaCutoff

◆ InlineMaxBB

cl::opt< size_t > InlineMaxBB("amdgpu-inline-max-bb", cl::Hidden, cl::init(1100), cl::desc("Maximum number of BBs allowed in a function after inlining" " (compile time constraint)")) ( "amdgpu-inline-max-bb" , cl::Hidden , cl::init(1100) , cl::desc("Maximum number of BBs allowed in a function after inlining" " (compile time constraint)") )	static

◆ MemcpyLoopUnroll

◆ UnrollMaxBlockToAnalyze

cl::opt< unsigned > UnrollMaxBlockToAnalyze("amdgpu-unroll-max-block-to-analyze", cl::desc("Inner loop block size threshold to analyze in unroll for AMDGPU"), cl::init(32), cl::Hidden) ( "amdgpu-unroll-max-block-to-analyze" , cl::desc("Inner loop block size threshold to analyze in unroll for AMDGPU") , cl::init(32) , cl::Hidden )	static

◆ UnrollRuntimeLocal

cl::opt< bool > UnrollRuntimeLocal("amdgpu-unroll-runtime-local", cl::desc("Allow runtime unroll for AMDGPU if local memory used in a loop"), cl::init(true), cl::Hidden) ( "amdgpu-unroll-runtime-local" , cl::desc("Allow runtime unroll for AMDGPU if local memory used in a loop") , cl::init(true) , cl::Hidden )	static

◆ UnrollThresholdIf

cl::opt< unsigned > UnrollThresholdIf("amdgpu-unroll-threshold-if", cl::desc("Unroll threshold increment for AMDGPU for each if statement inside loop"), cl::init(200), cl::Hidden) ( "amdgpu-unroll-threshold-if" , cl::desc("Unroll threshold increment for AMDGPU for each if statement inside loop") , cl::init(200) , cl::Hidden )	static

◆ UnrollThresholdLocal

cl::opt< unsigned > UnrollThresholdLocal("amdgpu-unroll-threshold-local", cl::desc("Unroll threshold for AMDGPU if local memory used in a loop"), cl::init(1000), cl::Hidden) ( "amdgpu-unroll-threshold-local" , cl::desc("Unroll threshold for AMDGPU if local memory used in a loop") , cl::init(1000) , cl::Hidden )	static

◆ UnrollThresholdPrivate

cl::opt< unsigned > UnrollThresholdPrivate("amdgpu-unroll-threshold-private", cl::desc("Unroll threshold for AMDGPU if private memory used in a loop"), cl::init(2700), cl::Hidden) ( "amdgpu-unroll-threshold-private" , cl::desc("Unroll threshold for AMDGPU if private memory used in a loop") , cl::init(2700) , cl::Hidden )	static