llvm::codelayout Namespace Reference

Classes
struct	CDSortConfig
	Algorithm-specific params for Cache-Directed Sort. More...
struct	EdgeCount

Typedefs
using	EdgeT = std::pair< uint64_t, uint64_t >

Functions
std::vector< uint64_t >	computeExtTspLayout (ArrayRef< uint64_t > NodeSizes, ArrayRef< uint64_t > NodeCounts, ArrayRef< EdgeCount > EdgeCounts)
	Find a layout of nodes (basic blocks) of a given CFG optimizing jump locality and thus processor I-cache utilization.
double	calcExtTspScore (ArrayRef< uint64_t > Order, ArrayRef< uint64_t > NodeSizes, ArrayRef< EdgeCount > EdgeCounts)
	Estimate the "quality" of a given node order in CFG.
double	calcExtTspScore (ArrayRef< uint64_t > NodeSizes, ArrayRef< EdgeCount > EdgeCounts)
	Estimate the "quality" of the current node order in CFG.
std::vector< uint64_t >	computeCacheDirectedLayout (ArrayRef< uint64_t > FuncSizes, ArrayRef< uint64_t > FuncCounts, ArrayRef< EdgeCount > CallCounts, ArrayRef< uint64_t > CallOffsets)
	Apply a Cache-Directed Sort for functions represented by a call graph.
std::vector< uint64_t >	computeCacheDirectedLayout (const CDSortConfig &Config, ArrayRef< uint64_t > FuncSizes, ArrayRef< uint64_t > FuncCounts, ArrayRef< EdgeCount > CallCounts, ArrayRef< uint64_t > CallOffsets)
	Apply a Cache-Directed Sort with a custom config.

Typedef Documentation

EdgeT

using llvm::codelayout::EdgeT = typedef std::pair<uint64_t, uint64_t>

Definition at line 24 of file CodeLayout.h.

Function Documentation

calcExtTspScore() [1/2]

double llvm::codelayout::calcExtTspScore	(	ArrayRef< uint64_t >	NodeSizes,
		ArrayRef< EdgeCount >	EdgeCounts
	)

Estimate the "quality" of the current node order in CFG.

Definition at line 1449 of file CodeLayout.cpp.

References calcExtTspScore(), Idx, and llvm::ArrayRef< T >::size().

calcExtTspScore() [2/2]

double llvm::codelayout::calcExtTspScore	(	ArrayRef< uint64_t >	Order,
		ArrayRef< uint64_t >	NodeSizes,
		ArrayRef< EdgeCount >	EdgeCounts
	)

Estimate the "quality" of a given node order in CFG.

The higher the score, the better the order is. The score is designed to reflect the locality of the given order, which is anti-correlated with the number of I-cache misses in a typical execution of the function.

Definition at line 1428 of file CodeLayout.cpp.

References Addr, Idx, and llvm::ArrayRef< T >::size().

Referenced by calcExtTspScore().

computeCacheDirectedLayout() [1/2]

std::vector< uint64_t > llvm::codelayout::computeCacheDirectedLayout	(	ArrayRef< uint64_t >	FuncSizes,
		ArrayRef< uint64_t >	FuncCounts,
		ArrayRef< EdgeCount >	CallCounts,
		ArrayRef< uint64_t >	CallOffsets
	)

Apply a Cache-Directed Sort for functions represented by a call graph.

The placement is done by optimizing the call locality by co-locating frequently executed functions. FuncSizes: The sizes of the nodes (in bytes). FuncCounts: The execution counts of the nodes in the profile. CallCounts: The execution counts of every edge (jump) in the profile. The map also defines the edges in CFG and should include 0-count edges. CallOffsets: The offsets of the calls from their source nodes.

Returns

The best function order found.

Definition at line 1471 of file CodeLayout.cpp.

References CacheEntries, CacheSize, CDMaxChainSize, computeCacheDirectedLayout(), Config, DistancePower, and FrequencyScale.

Referenced by computeCacheDirectedLayout().

computeCacheDirectedLayout() [2/2]

std::vector< uint64_t > llvm::codelayout::computeCacheDirectedLayout	(	const CDSortConfig &	Config,
		ArrayRef< uint64_t >	FuncSizes,
		ArrayRef< uint64_t >	FuncCounts,
		ArrayRef< EdgeCount >	CallCounts,
		ArrayRef< uint64_t >	CallOffsets
	)

Apply a Cache-Directed Sort with a custom config.

Definition at line 1457 of file CodeLayout.cpp.

References assert(), Config, and llvm::ArrayRef< T >::size().

computeExtTspLayout()

std::vector< uint64_t > llvm::codelayout::computeExtTspLayout	(	ArrayRef< uint64_t >	NodeSizes,
		ArrayRef< uint64_t >	NodeCounts,
		ArrayRef< EdgeCount >	EdgeCounts
	)

Find a layout of nodes (basic blocks) of a given CFG optimizing jump locality and thus processor I-cache utilization.

This is achieved via increasing the number of fall-through jumps and co-locating frequently executed nodes together. The nodes are assumed to be indexed by integers from [0, |V|) so that the current order is the identity permutation. NodeSizes: The sizes of the nodes (in bytes). NodeCounts: The execution counts of the nodes in the profile. EdgeCounts: The execution counts of every edge (jump) in the profile. The map also defines the edges in CFG and should include 0-count edges.

Returns

The best block order found.

Definition at line 1411 of file CodeLayout.cpp.

References assert(), and llvm::ArrayRef< T >::size().