WhenMathParallel vs Sequential Work Comparator
Group tasks into stages. Within a stage, tasks run in parallel and the stage takes as long as its longest task, while stages themselves always run one after another. See the total time either way.
Worked example
Stage 1 has a 2-day task and a 5-day task running in parallel, taking 5 days total since the stage is bound by its longest task, and Stage 2 has a single 1-day task after that, bringing the parallel total to 6 days. Running every task fully sequentially instead adds them all up directly, 2 plus 5 plus 1, coming to 8 days total, a 2-day difference driven entirely by the 2-day task in Stage 1 running alongside the 5-day one instead of after it.
Common mistakes
- Assuming parallel execution always cuts the timeline roughly in half. The actual saving equals the sum of each stage's non-longest tasks, not a fixed fraction of the total.
- Assuming every task within a stage can genuinely run at the same time with no resource conflict. That assumption needs to hold true for the parallel figure to be realistic.
Limitations
This assumes every task in a stage can genuinely run at the same time with no resource conflict. It does not check real dependencies between individual tasks, so the Phase Dependency Planner should be used instead when tasks depend on each other directly.
Frequently asked questions
- Can a stage have just one task?
- Yes, a single-task stage behaves the same either way, since there is nothing to parallelize within a stage that has only one task.
- Does adding more tasks to a stage always increase its total time?
- No, in the parallel view the stage still takes as long as its single longest task, so adding a shorter task alongside existing ones has no effect on that stage's duration at all.