Package solvcon :: Module rpc :: Class Worker

[frames] | no frames]

Class Worker

object --+
         |
        Worker

The whole worker object will run remotely (means in separated process).

Instance Methods

__init__(self, muscle, profiler_data=None)
x.__init__(...) initializes x; see x.__class__.__doc__ for signature

source code

_eventloop(self)
Event loop.

source code

eventloop(self)

source code

register(self, address, authkey, *args, **kw)
Connect to remote listener and run event lop.

source code

run(self, address, authkey, *args, **kw)
Listen to given address and run event loop.

source code

chdir(self, dirname)

source code

remote_setattr(self, name, var)
Remotely set attribute of worker.

source code

remote_loadobj(self, name, objfn)
Remotely unpickle a file and set it to self with the specified name.

source code

barrier(self)
Send barrier signal for synchronization.

source code

accept_peer(self, peern, family, authkey)
Accept connection from specified peer.

source code

connect_peer(self, peern, address, authkey)
Make a connection to specified peer (it has to be accepting connection).

source code

set_peer(self, src, dst)
Create MPI proxy for a pair of p2p connection.

source code

get_port_by_mpi(self, dst, tag)

source code

terminate(self)

source code

create_solver(self, bcmap, dirname, blkfn, iblk, nblk, solvertype, svrkw)
Load a block and create a solver object with the given information, and set it to muscle.

source code

drop_anchor(self, ankcls, ankkw)
Create an anchor object and append it to the solver muscle.

source code

Inherited from object: __delattr__, __format__, __getattribute__, __hash__, __new__, __reduce__, __reduce_ex__, __repr__, __setattr__, __sizeof__, __str__, __subclasshook__

Instance Variables

conn
the connection object to master.

lsnr
the listener object for master.

muscle
the muscle object.

pconns
dictionary of connection objects to peers.

plsnrs
dictionary of listener objects for peers.

serial
serial number of the worker process.

Properties

Inherited from object: __class__

Method Details

init(self, muscle, profiler_data=None)
(Constructor)

x.__init__(...) initializes x; see x.__class__.__doc__ for signature

Overrides: object.__init__: (inherited documentation)

register(self, address, authkey, *args, **kw)

Connect to remote listener and run event lop. In this case, the worker don't have a valid listener since it acts as a client to the dealer.

Parameters:

address (str or tuple) - address to connect.
authkey (str) - authentication key for connection.

run(self, address, authkey, *args, **kw)

Listen to given address and run event loop. In this case, the worker has the listener and acts as a server.

Parameters:

address (str or tuple) - address to listen.
authkey (str) - authentication key for connection.

accept_peer(self, peern, family, authkey)

Accept connection from specified peer.

Parameters:

peern (int) - index of the peer who wants to connect.
family (str) - the family of address needed to be guessed.
authkey (str) - authentication key for connection.

connect_peer(self, peern, address, authkey)

Make a connection to specified peer (it has to be accepting connection).

Parameters:

peern (int) - index of the peer who wants to connect.
address (str or tuple) - the address to connect to.
authkey (str) - authentication key for connection.

set_peer(self, src, dst)

Create MPI proxy for a pair of p2p connection.

Parameters:

src (int) - source worker ID.
dst (int) - destination worker ID.

create_solver(self, bcmap, dirname, blkfn, iblk, nblk, solvertype, svrkw)

Load a block and create a solver object with the given information, and set it to muscle.

Parameters:

bcmap (dict) - BC mapper.
dirname (str) - the directory of saved domain object.
blkfn (str) - the relative path for the block to be loaded.
iblk (int) - index of the block to be loaded.
nblk (int) - number of total blocks (sub-domains).
solvertype (type) - the type of solver to be created.
svrkw (dict) - keywords passed to the constructor of solver.

Returns:

nothing

drop_anchor(self, ankcls, ankkw)

Create an anchor object and append it to the solver muscle.

Parameters:

ankcls (type) - anchor type.
ankkw (dict) - keywords to the constructor of the anchor.

Returns:

nothing