I want to correctly model my RT/Embedded SW activity as a UML activity diagram.
My activity comprises two phases, where Phase 2 runs some (relatively “long”) time after Phase 1 has completed. (And, btw, so they are not parallel.)

Goal of question — To know if:

• My activity is one activity, or in fact two separate activities?

• If indeed two separate activities, can I nevertheless legally model it on a single UML activity diagram (maybe optionally as two partitions)?

Additional detail:

The activities generate results in two phases. Let’s name them:

GenResultsPhaseONE and GenResultsPhaseTWO_andOutput

These two activities are not necessarily procedurally linked – but there is indeed a repeated synchronous processing Period, where, let’s say:

• At Time T1, GenResults_Phase1 is invoked and generates & saves
  ResultData_Temp.

• At Time T2, GenResults_Phase2_andOutput is invoked and reads
  ResultData_Temp and processes it further, and generates updated
  results, and outputs them.