Throughput-optimizing Compilation of Dataflow Applications for Multi-Cores using Quasi-Static Scheduling

Throughput-optimizing Compilation of Dataflow Applications for Multi-Cores using Quasi-Static Scheduling
复制标题

使用准静态调度的多核数据流应用程序的吞吐量优化编译

DOI:
10.1145/2764967.2764972
复制
发表时间:
2015
期刊:
Proceedings of the 18th International Workshop on Software and Compilers for Embedded Systems
影响因子:
--
通讯作者:
Christian Haubelt
Christian Haubelt
中科院分区:
--
文献类型:
--
作者:
Tobias Schwarzer;Joachim Falk;Michael Glaß;Jürgen Teich;Christian Zebelein;Christian Haubelt

文献摘要

参考文献

被引文献

相似文献

使用动态数据流图的应用程序建模非常适合多核平台。然而,应用程序的细粒度和平台之间经常存在不匹配的情况。根据平台定制这种粒度可以通过 (a) 减少动态调度开销和 (b) 利用编译器优化来提高性能。在本文中,我们提出了一种吞吐量优化的编译方法,该方法使用准静态调度(QSS)来组合静态数据流子图的参与者。我们提出的方法在设计空间探索 (DSE) 中结合了核心分配、QSS 和参与者绑定,优化了多个可用核心的吞吐量。在 DSE 期间,每个候选实现都被编译到目标硬件(此处为 Intel i7 和 ARM Cortex-A9)并在其上进行评估。包括综合基准测试以及实际控制应用在内的实验结果表明,我们提出的整体编译方法优于与 QSS 无关的经典 DSE 以及采用 QSS 作为后处理步骤的 DSE。其中,我们展示了一个案例,其中编译方法在 4 核实现中获得了 9.91 倍的加速,而经典的 DSE 仅获得了 2.12 倍的加速。
Application modeling using dynamic dataflow graphs is well-suited for multi-core platforms. However, there is often a mismatch between the fine granularity of the application and the platform. Tailoring this granularity to the platform promises performance gains by (a) reducing dynamic scheduling overhead and (b) exploiting compiler optimizations. In this paper, we propose a throughput-optimizing compilation approach that uses Quasi-Static Schedules (QSSs) to combine actors of static dataflow subgraphs. Our proposed approach combines core allocation, QSSs, and actor binding in a Design Space Exploration (DSE), optimizing the throughput for a number of available cores. During the DSE, each implementation candidate is compiled to and evaluated on the target hardware---here an Intel i7 and an ARM Cortex-A9. Experimental results including synthetic benchmarks as well as a real-world control application show that our proposed holistic compilation approach outperforms classic DSEs that are agnostic of QSS as well as a DSE that employs QSS as a post-processing step. Amongst others, we show a case where the compilation approach obtains a speedup of 9.91 x for a 4-core implementation, while a classic DSE only obtains a speedup of 2.12 x.
用于嵌入式芯片级多处理器的联合应用映射/互连综合技术
DOI: --
发表时间: 2005
影响因子: 5.3
作者:
N. Bambha;S. Bhattacharyya
通讯作者: S. Bhattacharyya
与 Compaan/Laura 一起实现过程网络的有效控制
DOI: --
发表时间: 2008
影响因子: 1.2
作者:
Steven Derrien;A. Turjan;Claudiu Zissulescu;B. Kienhuis;E. Deprettere
通讯作者: E. Deprettere
DOI: --
发表时间: 2010
期刊: IEEE Workshop on Embedded Systems for Real-Time Multimedia
影响因子: --
作者:
Sjoerd Meijer;Hristo Nikolov;T. Stefanov
通讯作者: T. Stefanov
吞吐量驱动的同步数据流转换,用于映射到异构 MPSoC
DOI: 10.1109/samos.2012.6404168
发表时间: 2012
期刊: 2012 International Conference on Embedded Computer Systems (SAMOS)
影响因子: --
作者:
Anastasia Stulova;R. Leupers;G. Ascheid
通讯作者: G. Ascheid
DOI: --
发表时间: 2008
期刊: 2008 45th ACM/IEEE Design Automation Conference
影响因子: --
作者:
Chia;J. Pino;S. Bhattacharyya
通讯作者: S. Bhattacharyya