This paper proposes TaskProf2, a parallelism profiler and an adviser for task parallel programs. As a parallelism profiler, TaskProf2 pinpoints regions with serialization bottlenecks, scheduling overheads, and secondary effects of execution. As an adviser, TaskProf2 identifies regions that matter in improving parallelism. To accomplish these objectives, it uses a performance model that captures series-parallel relationships between various dynamic execution fragments of tasks and includes fine-grained measurement of computation in those fragments. Using this performance model, TaskProf2’s what-if analyses identify regions that improve the parallelism of the program while considering tasking overheads. Its differential analyses perform fine-grained differencing of an oracle and the observed performance model to identify static regions experiencing secondary effects. We have used TaskProf2 to identify regions with serialization bottlenecks and secondary effects in many applications.
Tue 25 JunDisplayed time zone: Tijuana, Baja California change
08:30 - 09:30 | Parallelism and Super Computing IPLDI Research Papers at 228AB Chair(s): Veselin Raychev DeepCode AG | ||
08:30 20mTalk | Huron: Hybrid False Sharing Detection and Repair PLDI Research Papers Tanvir Ahmed Khan University of Michigan, USA, Yifan Zhao University of Michigan, USA, Gilles Pokam Intel Corporation, Barzan Mozafari University of Michigan, USA, Baris Kasikci University of Michigan, USA Media Attached | ||
08:50 20mTalk | Model-Driven Transformations for Multi- and Many-Core CPUs PLDI Research Papers Media Attached | ||
09:10 20mTalk | Parallelism-Centric What-If and Differential Analyses PLDI Research Papers Pre-print Media Attached | ||