Contexts in ScotchPy

Contexts are implemented in ScotchPy with the Context class.

The Context class

class Context(init=True)

The graph constructor itself calls libscotch’s contextAlloc function, and usually contextInit is called too.

Parameters:

init – If True, the context allocation is followed by the init. If not, the init() method needs to be called before.


init()

This routine Initializes the fields of the Context object instance. Called by a class constructor when init=True (by default). Mandatory before calling context routines or after calling exit() in order to re-use the same Context instance.


exit()

This routine cleans-up all fields of the instance. Is the opposite of the init() routine. Called automatically when the object is destroyed. Calling explicitly this method is only required when an init() call follows.


option_get_num(optinum, optival)

This routine retrieves the value of a given integer context value of index optinum in optival.

Parameters:

  • optinum (Integer) – The index of the context value to retrieve.

  • optival (Integer) – The value of the context value.


option_set_num(optinum, optival)

This routine sets the value of a given integer context value of index optinum to optival.

Parameters:

  • optinum (Integer) – The index of the context value to set.

  • optival (Integer) – The value of the context value.


random_clone()

This routine clones the current state of the random number generator of the context.


random_reset()

This routine resets the state of the random number generator of the context.


random_seed(seedval)

This routine sets the state of the random number generator of the context with the given seed.

Parameters:

seedval (Integer) – The seed value.


thread_import1(thrdnbr)

This routine initiates the capture, into the context, of an existing pool of threads. It must be called only by the master thread providing thrdnbr number of threads (including the master thread itself).

Parameters:

thrdnbr (Integer) – The number of threads to capture.


thread_import2(thrdnum)

This routine captures the thrdnum thread as a slave thread of the context. It must be called by each slave thread and the master thread after them. threading.Thread must be used here. See in “tests/test_context.py” for an example.

Parameters:

thrdnum (Integer) – The number of the thread to capture.


thread_spawn(thrdnbr, coretab)

This routine populates the context with a pool of thrdnbr- 1 slave threads, in addition to the master thread.

Parameters:

  • thrdnbr (Integer) – The number of threads to spawn.

  • coretab (Any iterable of integers) – The core table, when not None, is an array of thrdnbr integers that gives the core number on which each thread should be bound.


bind_graph(orggraf, cntgraf)

This routine initializes a context container object with the graph orggraf. The graph cntgraf is a context container that will contain the graph orggraf. The graph orggraf is not duplicated, but only linked to the context container. The context container is used to store the graph and to provide a context for all graph operations.

Parameters:

  • orggraf (Graph) – The graph to bind.

  • cntgraf (Graph) – The context container.


bind_mesh(orgmesh, cntmesh)

This routine initializes a context container object with the mesh orgmesh. The mesh cntmesh is a context container that will contain the mesh orgmesh. The mesh orgmesh is not duplicated, but only linked to the context container. The context container is used to store the mesh and to provide a context for all mesh operations.

Parameters:

  • orgmesh (Mesh) – The mesh to bind.

  • cntmesh (Mesh) – The context container.


Various functions are defined as wrappers of Context operations:


context_alloc()

This routine creates the contextAlloc function, returning a non-initialized Context instance. It will need the init() method to be called before the instance be built or loaded upon.