Throughput-constrained DVFS for scenario-aware dataflow graphs

Damavandpeyma, M.; Stuijk, Sander; Basten, Twan; Geilen, Marc; Corporaal, Henk

doi:10.1109/rtas.2013.6531090

Cited by 19 publications

(17 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These studies do not minimize energy consumption, which is also crucial for battery-operated embedded devices. Modern cores support a wide range of voltages and frequencies, which are often exploited to perform dynamic voltage and frequency scaling (DVFS) to minimize the task computation energy [6], [7]. This has motivated researchers in recent years to jointly optimize lifetime reliability and energy consumption through intelligent task mappings [8]- [10].…”

Section: Introductionmentioning

confidence: 99%

Reliability and Energy-Aware Mapping and Scheduling of Multimedia Applications on Multiprocessor Systems

Das

Kumar

Veeravalli

2016

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

Abstract-Lifetime reliability is an emerging concern in multiprocessor systems as escalating power density and hence temperature variation continues to accelerate wear-out leading to a growing prominence of device defects. In this paper, we propose a system-level approach that involves performance-aware mapping of multimedia applications on a multiprocessor system to jointly minimize energy consumption and temperature related wear-out. Fundamental to this approach is a simplified temperature model that incorporates not only the transient and the steady-state behavior (temporal effect), but also the temperature dependency on the surrounding cores (spatial effect). This model is validated against the temperature obtained using the HotSpot tool with transient and steady-state simulations, and is shown to be accurate within 5.5 • C, leading to an MTTF estimation accuracy of an average 21% with respect to the state-of-the-art approaches. The proposed temperature model is integrated in a gradient-based fast heuristic that controls the voltage and frequency of the cores to limit the average and peak temperature leading to a longer lifetime, simultaneously minimizing the energy consumption. Lifetime computation considers task remapping, which is a common feature available in modern multiprocessor systems. A linear programming approach is then proposed to distribute the cores of a multiprocessor system among concurrent applications to maximize the lifetime. Experiments conducted with a set of synthetic and real-life applications represented as synchronous data flow graphs demonstrate that the proposed approach minimizes energy consumption by an average 24% with 47% increase in lifetime. For concurrent applications, the proposed lifetime-aware core distribution results in an average 10% improvement in lifetime as compared to performance-based core distribution.Index Terms-Lifetime reliability, mean time to failure (MTTF), platform-based design, synchronous data flow graphs.

show abstract

Section: Introductionmentioning

confidence: 99%

Reliability and Energy-Aware Mapping and Scheduling of Multimedia Applications on Multiprocessor Systems

Das

Kumar

Veeravalli

2016

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

show abstract

“…To schedule and analyze the performance of streaming applications on a platform, a computation model is required. The expressivity of Synchronous Data Flow (SDF) [6], [7] fits well with the features of streaming applications and hence SDF graphs (SDFGs) are widely used in literature [8], [9], [10] and also in this paper. In an SDFG, tasks communicate by the FIFO allocated to each edge.…”

Section: Introductionmentioning

confidence: 76%

“…Recently, data flow graphs like SDFG and Scenario Aware Dataflow Graph (SADFG) [10], [26], [27], [9] gain much attention due to their powerful combination of expressivity and analyzability. In this paper, we use SDF to model streaming applications like software defined radio and multimedia applications.…”

Section: B Application Modelmentioning

confidence: 99%

Task-FIFO Co-Scheduling of Streaming Applications on MPSoCs with Predictable Memory Hierarchy

Tang

Basten

Geilen

et al. 2017

ACM Trans. Embed. Comput. Syst.

View full text Add to dashboard Cite

Abstract-Multi-processor systems-on-chips are widely adopted in implementing modern streaming applications to satisfy the ever increasing computing requirements. Predictable memory hierarchies, which make memory access predictable, can better satisfy the strict timing requirements of streaming applications. However, different levels of the memory hierarchy vary in latency and capacity. Hence, the system performance not only depends on the task schedule but also closely relates with the FIFO size distribution and FIFO allocation, which makes the scheduling problem much more complex. We propose an efficient Iterationbased Task-FIFO Co-Scheduling algorithm to optimize the FIFO size distribution and task/FIFO assignment. Randomly generated Synchronous Dataflow Graphs with different sizes and a set of practical applications are used to evaluate the performance of the proposed method. The experimental results demonstrate that the proposed algorithm outperforms the load balancing method and the Highest Access Frequency First algorithm.

show abstract

“…The second challenge is how to achieve energy efficiency of the MPSoC. To address the first challenge, selftimed scheduling has been widely considered as the scheduling policy (e.g., [4], [5]) for streaming applications modeled as data-flow graphs to guarantee the throughput constraints. The analysis techniques for self-timed scheduling, however, necessitate a complex design space exploration (DSE) to determine the minimum number of processors needed to schedule the applications and the assignment of tasks to processors.…”

Section: Introductionmentioning

confidence: 99%

Energy-efficient scheduling of throughput-constrained streaming applications by periodic mode switching

Niknam

Stefanov

2017

2017 International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation (SAMOS)

View full text Add to dashboard Cite

In this paper, we address the problem of energy reduction when scheduling streaming applications with throughput constraints on homogeneous multiprocessor systems with voltage and frequency scaling capability. We propose a novel periodic scheduling framework which allows streaming applications to switch their execution periodically between a few energy-efficient schedules at run-time in order to meet a throughput constraint at long run. Using such periodic switching, we can benefit from adopting Dynamic Voltage and Frequency Scaling (DVFS) techniques to exploit available static slack time in the schedule of an application efficiently. The experimental results, on a set of real-life streaming applications, show that our novel scheduling approach can achieve up to 68% energy reduction depending on the application and the throughput constraint compared to related approaches.

show abstract

Throughput-constrained DVFS for scenario-aware dataflow graphs

Cited by 19 publications

References 20 publications

Reliability and Energy-Aware Mapping and Scheduling of Multimedia Applications on Multiprocessor Systems

Reliability and Energy-Aware Mapping and Scheduling of Multimedia Applications on Multiprocessor Systems

Task-FIFO Co-Scheduling of Streaming Applications on MPSoCs with Predictable Memory Hierarchy

Energy-efficient scheduling of throughput-constrained streaming applications by periodic mode switching

Contact Info

Product

Resources

About