Adaptive task duplication using on-line bottleneck detection for streaming applications

Choi, Yoonseo; Li, Chenghong; Silva, Dilma Da; Bivens, Alan; Schenfeld, Eugen

doi:10.1145/2212908.2212932

Cited by 11 publications

(9 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to alleviate the effort of considering all dynamic cases at the specification phase, online optimization approaches have been proposed. Choi et al [15] proposed to run dataflow applications with an initial configuration, then properly modify it as necessary at runtime. To be more specific, the bottleneck of the dataflow graph in terms of throughput is identified on the fly and the execution of the identified task is accelerated by means of task duplication.…”

Section: Related Workmentioning

confidence: 99%

“…3) The proposed technique is more general in applying various kinds of design constraints and objectives. Note that the existing online optimization techniques [15], [16], [20] only focused on maximizing throughput without guaranteeing any design properties like worstcase power or latency. Due to the fundamental limitation of online optimization, they are not suitable for the design of real-time or power-constrained systems.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Efficiently Switchable Context-Aware Dataflow Adaptation Technique for Low-Power Multi-Core Embedded Systems

Jung

Yang

2019

IEEE Access

View full text Add to dashboard Cite

Today's embedded systems operate under increasingly dynamic conditions. First, computational workloads can be either variable by nature or adjustable. Moreover, as many devices are batterypowered, it is common to have runtime power management technique, which results in dynamic power budget. This paper presents a design methodology for multi-core systems, based on dataflow specification, that can deal with various contexts. We optimize the original dataflow considering various working conditions, then, autonomously adapt it to a pre-defined optimal form in response to context changes. Such context changes at runtime might cause non-negligible delay or power budget violation. In order to overcome them, an efficient mode switching method is proposed. We show the effectiveness of the proposed technique with a real-life case study, stereo-vision, and synthetic benchmarks.INDEX TERMS Context-awareness, dataflow adaptation, low-power electronics, mode switching, multicore processing.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Efficiently Switchable Context-Aware Dataflow Adaptation Technique for Low-Power Multi-Core Embedded Systems

Jung

Yang

2019

IEEE Access

View full text Add to dashboard Cite

show abstract

“…However its adaptation possibilities are defined by particular patterns specified statically. [6] extends this approach by allowing for dynamic splitting of actors. The main limitation of both approaches lie in the fact that programs need to be acyclic.…”

Section: Related Workmentioning

confidence: 99%

A Monitoring System for Runtime Adaptations of Streaming Applications

Selva

Morel

Marquet

et al. 2015

2015 23rd Euromicro International Conference on Parallel, Distributed, and Network-Based Processing

View full text Add to dashboard Cite

Streaming languages are adequate for expressing many applications quite naturally and have been proven to be a good approach for taking advantage of the intrinsic parallelism of modern CPU architectures. While numerous works focus on improving the throughput of streaming programs, we rather focus on satisfying quality-of-service requirements of streaming applications executed along-side non-streaming processes. We monitor synchronous dataflow (SDF) programs at runtime both at the application and system levels in order to identify violations of quality-of-service requirements. Our monitoring requires the programmer to provide the expected throughput of its application (e.g 25 frames per second for a video decoder), then takes full benefit from the compilation of the SDF graph to detect bottlenecks in this graph and identify causes among processor or memory overloading. It can then be used to perform dynamic adaptations of the applications in order to optimize the use of computing and memory resources.

show abstract

“…However, as shown in Section 5.3, the memory usage of the application is virtually independent of the available PEs and the application might have a considerable scheduling overhead on a single PE. Run-time task duplication is used in [3] to maximize the application's throughput. The technique replaces stateless processes by a master thread that distributes the actual work among its sibling threads.…”

Section: Related Workmentioning

confidence: 99%

“…These steps are not trivial if stateful processes are executed asynchronously and the amount of data being produced or consumed by a process is not known beforehand. Previous works, therefore, limited their approaches to programming models with statically specified data production and consumption rates [2] or to the replication of stateless processes [3].…”

Section: Introductionmentioning

confidence: 99%

AdaPNet

Schor

Bacivarov

Yang

et al. 2014

Proceedings of the 2014 International Conference on Compilers, Architecture and Synthesis for Embedded Systems

View full text Add to dashboard Cite

A widely considered strategy to prevent interference issues on multi-processor systems is to isolate the execution of the individual applications by running each of them on a dedicated virtual guest machine. The amount of computing power available to a single application, however, depends on the other applications running on the system and may change over time. A promising approach to maximize the performance under such conditions is to adapt the application's degree of parallelism when the resources allocated to the application are changed. This enables an application to exploit not more parallelism than required, thereby reducing inter-process communication and scheduling overheads. In this paper, we introduce AdaPNet, a run-time system to execute streaming applications, which are modeled as process networks, efficiently on platforms with dynamic resource allocation. AdaPNet responds to changes in the available resources by first calculating a process network that maximizes the performance of the application on the new resources. Then, AdaPNet transparently transforms the application into the alternative network without discarding the program state. Targeting two many-core systems, we demonstrate that AdaPNet outperforms comparable run-time systems, which do not adapt the degree of parallelism, in terms of speed-up and memory usage.

show abstract

Adaptive task duplication using on-line bottleneck detection for streaming applications

Cited by 11 publications

References 14 publications

Efficiently Switchable Context-Aware Dataflow Adaptation Technique for Low-Power Multi-Core Embedded Systems

Efficiently Switchable Context-Aware Dataflow Adaptation Technique for Low-Power Multi-Core Embedded Systems

A Monitoring System for Runtime Adaptations of Streaming Applications

AdaPNet

Contact Info

Product

Resources

About