TurboBŁYSK: Scheduling for Improved Data-Driven Task Performance with Fast Dependency Resolution

Podobas, Artur; Brorsson, Mats; Vlassov, Vladimir

doi:10.1007/978-3-319-11454-5_4

Cited by 11 publications

(6 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [14], the authors propose the dep_pattern clause to cache data dependency patterns reducing dependency management overhead. Our proposal goes further, not only caching dependency structures but preventing task creation altogether.…”

Section: A Task Management Overheadmentioning

confidence: 99%

Optimizing Iterative Data-flow Scientific Applications using Directed Cyclic Graphs

Álvarez

Beltrán

2023

IEEE Access

View full text Add to dashboard Cite

Data-flow programming models have become a popular choice for writing parallel applications as an alternative to traditional work-sharing parallelism. They are better suited to write applications with irregular parallelism that can present load imbalance. However, these programming models suffer from overheads related to task creation, scheduling and dependency management, limiting performance and scalability when tasks become too small. At the same time, many HPC applications implement iterative methods or multi-step simulations that create the same directed acyclic graphs of tasks on each iteration. By giving application programmers a way to express that a specific loop is creating the same task pattern on each iteration, we can create a single task DAG once and transform it into a cyclic graph. This cyclic graph is then reused for successive iterations, minimizing task creation and dependency management overhead. This paper presents the taskiter, a new construct we propose for the OmpSs-2 and OpenMP programming models, allowing the use of directed cyclic task graphs (DCTG) to minimize runtime overheads. Moreover, we present a simple immediate successor locality-aware heuristic that minimizes task scheduling overhead by bypassing the runtime task scheduler. We evaluate the implementation of the taskiter and the immediate successor heuristic in 8 iterative benchmarks. Using small task granularities, we obtain a geometric mean speedup of 2.56x over the reference OmpSs-2 implementation, and a 3.77x and 5.2x speedup over the LLVM and GCC OpenMP runtimes, respectively.INDEX TERMS taskiter, data-flow programming, ompss-2, openmp, iterative applications

show abstract

Section: A Task Management Overheadmentioning

confidence: 99%

Optimizing Iterative Data-flow Scientific Applications using Directed Cyclic Graphs

Álvarez

Beltrán

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The GOMP library and LLVM's OpenMP runtime are also available online as Open Source software [17] [23]. The topic of overhead in dependency resolution has also been tackled from other angles, such as the TurboBLYSK framework, to create dependency patterns [29].…”

Section: Related Workmentioning

confidence: 99%

Advanced synchronization techniques for task-based runtime systems

Álvarez

Sala

Maroñas

et al. 2021

Proceedings of the 26th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming

View full text Add to dashboard Cite

Task-based programming models like OmpSs-2 and OpenMP provide a flexible data-flow execution model to exploit dynamic, irregular and nested parallelism. Providing an efficient implementation that scales well with small granularity tasks remains a challenge, and bottlenecks can manifest in several runtime components. In this paper, we analyze the limiting factors in the scalability of a task-based runtime system and propose individual solutions for each of the challenges, including a wait-free dependency system and a novel scalable scheduler design based on delegation. We evaluate how the optimizations impact the overall performance of the runtime, both individually and in combination. We also compare the resulting runtime against state of the art OpenMP implementations, showing equivalent or better performance, especially for fine-grained tasks.CCS Concepts: • Computing methodologies → Parallel programming languages.

show abstract

“…Previous works discussed the task scheduling and dependences resolution overheads in data-driven task-based models like OpenMP and OmpSs. TurboBLYSK [21] is a framework which implements the OpenMP 4.0 with a custom compiler and a highly efficient runtime schedule of tasks with explicit data-dependence annotations. Its objective is also to reduce the dependence management overheads of the runtime.…”

Section: Related Workmentioning

confidence: 99%