This new pass is basically a clone of parallelization_planner, except it doesn't embed metadata. Its purpose is to print stats on potential time savings from parallelization. Since most of the code is duplicated it should be refactored once TimingModel is brought into the core.
This new pass is basically a clone of parallelization_planner, except it doesn't embed metadata. Its purpose is to print stats on potential time savings from parallelization. Since most of the code is duplicated it should be refactored once TimingModel is brought into the core.