The algorithm in scheduleBlock can be improved to make the visualization more accurate. We could use the set of scheduled nodes instead of visited nodes when determining "allPredsScheduled".

before(.pdf) the change 22 g1StoreP is visualized before, but depends on, 23-25 MachTemp in the After Matching phase. This inaccuracy is removed after(.pdf) the proposed patch. Thanks [~rcastanedalo] for the graphs.