Classes
class	graphlab::async_consensus
	This implements a distributed consensus algorithm which waits for global completion of all computation/RPC events on a given object. More...
class	graphlab::coherent_dht< KeyType, ValueType >
struct	graphlab::dc_init_param
	Distributed control constructor parameters. More...
class	graphlab::distributed_control
	The distributed control object is primary means of communication between the distributed GraphLab processes. More...
class	graphlab::dc_dist_object< T >
	Provides a class with its own distributed communication context, allowing instances of the class to communicate with other remote instances. More...
class	graphlab::dht< KeyType, ValueType >
struct	graphlab::request_future< T >

Macros
#define	RPC_DEFAULT_NUMHANDLERTHREADS (size_t)(-1)
	default number of handler threads to spawn.
#define	RPC_DEFAULT_COMMTYPE TCP_COMM
	default communication method
#define	RPC_MAX_N_PROCS 128
	Maximum number of processes supported.
#define	BUFFER_RELINQUISH_LIMIT 131072

Typedefs
typedef uint16_t	graphlab::procid_t
	The type used for numbering processors.

Functions
bool	graphlab::init_param_from_env (dc_init_param &param)
bool	graphlab::init_param_from_mpi (dc_init_param &param, dc_comm_type commtype)

Detailed Description

Macro Definition Documentation

#define BUFFER_RELINQUISH_LIMIT 131072

Below this limit, a copy of the buffer is made for the sender. Above this limit, the entire buffer is given to the sender.

Definition at line 53 of file dc_compile_parameters.hpp.

bool graphlab::init_param_from_env ( dc_init_param & param )

initializes parameters from environment. Returns true on success

Definition at line 33 of file dc_init_from_env.cpp.

bool graphlab::init_param_from_mpi	(	dc_init_param &	param,
		dc_comm_type	commtype = `TCP_COMM`
	)

initializes parameters from MPI. Returns true on success MPI must be initialized before calling this function

Definition at line 39 of file dc_init_from_mpi.cpp.