Galois::WorkList Namespace Reference

Scheduling policies for Galois iterators. More...

Namespaces
namespace	anonymous_namespace{WorkList.h}
Classes
struct	ChunkHeader
class	AltChunkedQueue
class	AltChunkedStack
class	StealingQueue
struct	AltChunkedMaster
class	AltChunkedLIFO
class	AltChunkedFIFO
class	BulkSynchronous
	Bulk-synchronous scheduling. More...
struct	squeue
struct	squeue< false, PS, TQ >
struct	ChunkedMaster
	Common functionality to all chunked worklists. More...
class	ChunkedFIFO
	Chunked FIFO. More...
class	ChunkedLIFO
	Chunked LIFO. More...
class	dChunkedFIFO
	Distributed chunked FIFO. More...
class	dChunkedLIFO
	Distributed chunked LIFO. More...
class	dChunkedBag
	Distributed chunked bag. More...
struct	ExternRef
struct	FIFO
	Simple FIFO worklist (not scalable). More...
struct	GFIFO
struct	LIFO
	Simple LIFO worklist (not scalable). More...
struct	NoGlobalQueue
struct	LocalQueue
struct	OrderedByIntegerMetric
	Approximate priority scheduling. More...
class	OrderedList
struct	OwnerComputes
struct	StableIterator
class	ConExtListNode
class	ConExtIterator
class	ConExtLinkedStack
class	ConExtLinkedQueue
struct	DummyIndexer

---

Detailed Description

Scheduling policies for Galois iterators.

Unless you have very specific scheduling requirement, dChunkedLIFO or dChunkedFIFO is a reasonable scheduling policy. If you need approximate priority scheduling, use OrderedByIntegerMetric. For debugging, you may be interested in FIFO or LIFO, which try to follow serial order exactly.

The way to use a worklist is to pass it as a template parameter to for_each(). For example,

 Galois::for_each<Galois::WorkList::dChunkedFIFO<32> >(begin, end, fn);