When you run the PartitionParticipants() command, the PartitioningInfo argument accepts a dictionary specifying what machine resources are assigned to each participant. Each tuple pair in the dictionary consists of:

When you set partitioning details using the PartitioningInfo argument, System Coupling uses the Custom partitioning algorithm automatically and blocks all other partitioning algorithms.

If you provide incomplete information for the PartitioningInfo argument — that is, when either the name of an existing participant is not included or an included participant is associated with an empty machine list — then the affected participant executes using its default parallel settings.

To get values for the PartitioningInfo argument, obtain lists of available participants and machines.

To get participants, query the CouplingParticipant object.

>>> DatamodelRoot().CouplingParticipant.GetChildNames()

['MAPDL-1', 'FLUENT-2', 'FLLUENT-3']

To get machines, run the GetMachines() command.

>>> GetMachines()

[{'machine-name' : 'machine-1', 'core-count' : 4}, {'machine-name' : 'machine-2', 
'core-count' : 8}, {'machine-name' : 'machine-3', 'core-count' : 12}]

You may either obtain these lists as shown above or include the queries in a Python script, as shown in Applying Custom Partitioning by Running a Script.

To apply custom partitioning, run the PartitionParticipants() command, using the PartitioningInfo argument to specify participants and the machines assigned to them.

In the example below, each of the three participants has one machine assigned to it. In this case, the AlgorithmName argument is not required because partitioning details are set by the PartitioningInfo argument.

>>>PartitionParticipants(   
      PartitioningInfo = {
          'MAPDL-1': [{'machine-name' : 'machine-1', 'core-count' : 4}], 
          'CFX-2': [{'machine-name' : 'machine-2', 'core-count' : 4}]
      }
    )

To apply custom partitioning in the GUI, specify partitioning parameters by running the PartitionParticipants() command in the Command Console. For details, see Applying Custom Partitioning by Running a Command.

When the command is executed, the Algorithm Name property is set to Custom automatically and other partitioning algorithms are blocked.

You can apply custom partitioning by running a script in System Coupling's CLI.

allMachines = GetMachines()
 
# all AEDT participants go on the first core
aedtMachines = []
aedtMachines.append({'machine-name' : allMachines[0]['machine-name'], 'core-count' : 1})
 
# all FLUENT participants go on the rest of the cores
fluentMachines = []
fluentMachines.append(allMachines[1])
fluentMachines.append({'machine-name' : allMachines[0]['machine-name'], 'core-count' :
    allMachines[0]['core-count'] - 1})
 
partitioningInfo = {}

 for participant in DatamodelRoot().CouplingParticipant.GetChildren():
    participantName = participant.GetName()
    if participant.ParticipantType == "AEDT":      
        partitioningInfo[participantName] = aedtMachines
    else if participant.ParticipantType == "FLUENT":
        partitioningInfo[participantName] = fluentMachines

PartitionParticipants(PartitioningInfo = partitioningInfo)