# pacman package¶

## pacman.exceptions module¶

exception pacman.exceptions.MachineHasDisconnectedSubRegion[source]

Some part of the machine has no paths connecting it to the rest of the machine.

exception pacman.exceptions.MinimisationFailedError[source]

Raised when a routing table could not be minimised to reach a specified target.

exception pacman.exceptions.PacmanAlgorithmFailedToGenerateOutputsException[source]

An exception that indicates that an algorithm has not generated the correct outputs for some unknown reason.

exception pacman.exceptions.PacmanAlreadyExistsException(item_type, item_id)[source]

An exception that indicates that something already exists and that adding another would be a conflict.

Parameters: item_type (str) – The type of the item that already exists item_id (str) – The ID of the item which is in conflict
exception pacman.exceptions.PacmanAlreadyPlacedError[source]

Bases: ValueError

An exception that indicates multiple placements are being made for a vertex.

exception pacman.exceptions.PacmanCanNotFindChipException[source]

An exception that indicates the chip was not in the list of chips.

exception pacman.exceptions.PacmanConfigurationException[source]

An exception that indicates that something went wrong with configuring some part of PACMAN.

exception pacman.exceptions.PacmanElementAllocationException[source]

An exception that indicates that something went wrong with element allocation.

exception pacman.exceptions.PacmanException[source]

Bases: Exception

Indicates a general exception from Pacman

exception pacman.exceptions.PacmanExternalAlgorithmFailedToCompleteException[source]

An exception that indicates that an algorithm ran from outside the software stack has failed to complete for some unknown reason.

exception pacman.exceptions.PacmanInvalidParameterException(parameter, value, problem)[source]

An exception that indicates that a parameter has an invalid value.

Parameters: parameter (str) – The name of the parameter value (str) – The value of the parameter problem (str) – The problem with the value of the parameter
exception pacman.exceptions.PacmanNoMergeException[source]

An exception that indicates to indicate that there are no merges worth performing.

exception pacman.exceptions.PacmanNotExistException[source]

An exception that indicates that a routing table entry was attempted to be removed from a routing table which didn’t have such an entry.

exception pacman.exceptions.PacmanNotFoundError[source]

An exception that indicates that some object has not been found when requested.

exception pacman.exceptions.PacmanNotPlacedError[source]

Bases: KeyError

An exception that indicates no placements are made for a vertex.

exception pacman.exceptions.PacmanPartitionException[source]

An exception that indicates that something went wrong with partitioning

exception pacman.exceptions.PacmanPlaceException[source]

An exception that indicates that something went wrong with placement

exception pacman.exceptions.PacmanProcessorAlreadyOccupiedError[source]

Bases: ValueError

An exception that indicates multiple placements are being made to a processor.

exception pacman.exceptions.PacmanProcessorNotOccupiedError[source]

Bases: KeyError

An exception that indicates that no placement has been made to a processor.

exception pacman.exceptions.PacmanPruneException[source]

An exception that indicates that something went wrong with pruning

exception pacman.exceptions.PacmanRouteInfoAllocationException[source]

An exception that indicates that something went wrong with route info allocation.

exception pacman.exceptions.PacmanRoutingException[source]

An exception that indicates that something went wrong with routing.

exception pacman.exceptions.PacmanTypeError[source]

An exception that indicates that an object is of incorrect type.

exception pacman.exceptions.PacmanValueError[source]

An exception that indicates that a value is invalid for some reason.

exception pacman.exceptions.PartitionMissingEdgesException[source]

Raise when after partitioning a partition does h=not have the edges it expects

exception pacman.exceptions.SDRAMEdgeSizeException[source]

raised when a constant SDRAM partition discovers its edges have inconsistent size requests

## Module contents¶

Provides various functions which together can be used to take a graph and split it into pieces that can be loaded on to a machine, along with routes between the pieces.

### Functional Requirements¶

• Creation of an Application Graph of Vertices indicating points of computation within the graph and Edges between the vertices indicating a directional communication between the vertices; and a similar Machine Graph.

• Vertices in the Application Graph will have a number of atoms - an atom cannot be broken down in to anything smaller.

• Vertices in the Application Graph must be able to indicate what machine resources are required by any given subset of the atoms.

• Vertices in the Machine Graph must be able to fit on a single chip of the machine in terms of resource usage.

• A Vertex can have a number of constraints which must be respected by any algorithm which uses the graph. Algorithms must check that they can support the given constraints and must fail if they cannot. Provided constraints include support for:

• The maximum number of atoms which any Machine Graph Vertex can contain for a given Application Graph vertex
• The chip and/or processor on to which a Machine Graph Vertex should be placed.
• A set of Application Graph Vertices whose corresponding Machine Graph vertices should contain the same number of atoms.
• A set of Application Graph Vertices whose corresponding Machine Graph vertices should be placed on the same chip if they contain the same atom.
• It should be possible to create new constraints as the need arises.

• Multiple edges can exist between the same two vertices.

• It must be possible to build the Machine Graph directly without requiring that it is created by one of the other modules.

• It is not required that there is a Machine Graph Edge between every pair of Machine Graph Vertex from the same Application Graph Vertex.

• Where a Machine Graph is created from an Application Graph, it should be possible to find the corresponding Vertices and Edges from one graph to the other.

• Creation of multicast routing info consisting of key/mask combinations assigned to Edges of the Machine Graph.

• It must be possible to build this information directly without requiring that it is created by one of the other modules.
• There should be exactly one key/mask combination for each Edge in the Machine Graph, which will represent all the keys which will be sent in all packets from the Vertex at the start of the Edge down that Edge.
• It is possible for a Vertex to send several different keys down several different Edges, but only one per Edge (but note that it is acceptable for different keys to be assigned to different Edges between the same two Vertices).
• There should be no overlap between the key/mask combinations of Edges which come from different Vertices i.e. no two Edges which start at different Vertices should have the same key/mask combination.
• Partitioning of an Application graph with respect to a machine, such that the resources consumed by each Vertex does not exceed those provided by each chip on the machine.

• It should be possible to select from a range of partitioning algorithms or provide one, although a default should be provided in the absence of such a choice .
• Any partitioning constraints should be met; if there are any that cannot, or that are not understood by the algorithm in use an exception should be thrown. Non-partitioning constraints can be ignored, although these can be used if it makes sense for the given algorithm.
• It must be possible to create at least one grouping of the generated Vertices so that each group fits within the resources provided by a single chip on the machine.
• It should not be assumed that a given grouping of Vertices will be the final grouping on the machine, although it is acceptable to make hints through additional constraints about what is likely to work.
• The machine itself must not be altered by the partitioning, so that it can be used in further processing.
• The graph itself must not be altered by the partitioning, so that it can be used in further processing.
• No two Machine Graph Vertices created from a single Application Graph Vertex can contain the same atom.
• Any Edges in the Application Graph must be split with the Vertices to create a number of Machine Graph edges, such that where there was a vertex v connected to a vertex w by a single edge in the Application Graph, there should be an Edge in the Machine Graph between every Vertex of Application Graph Vertex v and every Vertex of Application Graph Vertex w; for example, if there are 2 Machine Graph Vertices for each of v and w, and one Edge between them in the Application Graph, then there will be 4 new Edges in the Machine Graph for this Edge.
• Placement of a Machine Graph on a given machine, such that the resources required by any combination of Vertices placed on any chip in the machine does not exceed the resources provided by that chip.

• It should be possible to choose from a range of placement algorithms or provide one, although a default should be provided in the absence of such a choice.
• Any placement constraints should be met; if there are any that cannot, or that are not understood by placement algorithm, an exception should be thrown. Non-placement constraints can be ignored, although these can be used if it makes sense for the given algorithm.
• The machine itself should not be altered by placement so that it can be used in further processing.
• The graph itself should not be altered by placement so that it can be used in further processing.
• The returned placements should only contain a single placement for each vertex.
• The placements should be such that the vertices with edges between them must be able to communicate with each other.
• Allocation of multicast routing keys and masks to a Machine Graph such that each vertex sends out packets with a different key/mask combination.

• This can use the placement information if required. If an algorithm requires placement information but none is provided an exception is thrown.
• Routing of edges between vertices with a given allocation of routing keys and masks with respect to a given machine.

• It should be possible to choose from a range of routing algorithms, or provide one, although a default should be provided in the absence of such a choice
• For any vertex, following the routes from the placement of the vertex should result exactly in the set of placements of the destination vertices described by all the edges which start at that vertex. No additional destination should be reached, and no fewer than this set of destinations should be reached.
• It should be possible to call each of the modules independently. There should be no assumption that one of the other modules has produced the data input for any other module.

• There should be no assumption about how the inputs and outputs are stored.

• Any utility functions that provide access to internal structures within a data structure should operate in approximately O(1) time; for example, where an object of type obj holds a number of objects of type subobj with property prop, requesting a list of subobj objects contained within obj with property value prop = value should not iterate through a list of such objects, but should instead maintain a mapping that allows access to such objects in O(1) time. If this is not possible, obj should only provide access to a list of subobj objects, allowing the caller to filter these themselves. This will ensure that no misunderstanding can be made about the speed of operation of a function.