Instructions of dataflow programs correspond with the nodes of the corresponding dataflow graph. Hence, the nodes essentially encode an operation encoded as DataflowProcess. If it is none, it refers to an output node (that we have to add explicitly here). Moreover, we have to encode the potential two successor edges of the dataflow graph which is done by outL,argL and outR,argR which contain the addresses of the successor instructions and the argument index where the generated values will be sent to.
| Record Field | Description |
Full Usage:
argL
Field type: int
|
argL is the index of the argument of the consumer node outL
|
Full Usage:
argR
Field type: int
|
argR is the index of the argument of the consumer node outR
|
|
opc is the opcode of the node's operation; if None it is an output port
|
Full Usage:
outL
Field type: int
|
outL is the address of the consumer node of the produced left token
|
Full Usage:
outR
Field type: int
|
outR is the address of the consumer node of the produced right token
|