Energy-Efficient Runtime Management of Heterogeneous Multicores using Online Projection

Abstract

Heterogeneous multicores offer flexibility in the form of different core types and Dynamic Voltage and Frequency Scaling (DVFS), defining a vast configuration space. The optimal configuration choice is not always straightforward, even for single applications, and becomes a very difficult problem for dynamically changing scenarios of concurrent applications with unpredictable spawn and termination times and individual performance requirements. This article proposes an integrated approach for runtime decision making for energy efficiency on such systems. The approach consists of a model that predicts performance and power for any possible decision and low-complexity heuristics that use this model to evaluate a subset of possible decisions to choose the best. The model predicts performance by projecting standalone application profiling data to the current status of the system and power by using a set of platform-specific parameters that are determined only once for a given system and are independent of the application mix. Our approach is evaluated with a plethora of dynamic, multi-application scenarios. When considering best effort performance to be adequate, our runtime achieves on average 3% higher energy efficiency compared to the powersave governor and 2× better compared to the other linux governors. Moreover, when also considering individual applications' performance requirements, our runtime is able to satisfy them, giving away 18% of the system's energy efficiency compared to the powersave, which, however, misses the performance targets by 23%; at the same time, our runtime maintains an efficiency advantage of about 55% compared to the other governors, which also satisfy the performance constraints. 63:2 S. Tzilis et al. INTRODUCTIONMulticore architectures often offer heterogeneity and Dynamic Voltage and Frequency Scaling (DVFS). This flexibility provides knobs for trading performance for energy and vice versa thereby improving system efficiency. Commercial examples of such, single-ISA, architectures include the Samsung Exynos [21] and Qualcomm Snapdragon [19] processors. They employ multiple types of cores (power-efficient and high performance), and multiple voltage-frequency pairs per cluster of cores. In this manner, they can potentially adapt to the fluctuating system workload and its needs at runtime. Such runtime variations stem from the number, type, and performance requirements of the workloads running at any given time on the system. Adapting the system to these dynamic changes can balance smooth application operation and performance with energy efficiency.In this context, choosing the optimal system configuration is often not straightforward. Executing an application on a power-efficient core might prolong execution enough to reduce energy efficiency, while choosing a high-performance core poses the opposite risk-an increase of power consumption outweighing performance gains. DVFS adds more complexity to this problem, as it is unclear how a high-performance core at a low frequency and a power-effi...

Concurrent Application Bias Scheduling for Energy Efficiency of Heterogeneous Multi-Core Platforms

“…[3] Imitation [16] Online Conc. [22] Inter-cluster [21] Power-aware [2] Projection [26] DyPO [10] AdaMD [4] Proposed approach (c) show energy consumption of the suggested solutions, which are not optimal. On the other hand, it should be noted that the best configuration (red point) in fact is 4b for the first application and 4L for the second application, which requires migrations of the first application from LIT-TLE cores to big cores.…”

“…Table 1 comparatively summarizes the most recent and relevant state-of-the-art approaches as per controlling knobs, ability to handle dynamic multiprogrammed workloads, methodology, and decision strategies that they used. The frameworks presented in [3], [10], [21], [22], [26] cannot handle multi-thread applications and/or combinational optimization of different knobs, restricting their scope. The approach in [16] does consider multi-thread applications and combinational knobs setting, however, this approach neglects multi-application and dynamic workload scenarios.…”

“…The approach in [16] does consider multi-thread applications and combinational knobs setting, however, this approach neglects multi-application and dynamic workload scenarios. Power-aware [2], Projection [26], and DyPO [10] consider the application bias for guiding resource allocation, which we also consider in this work. Thus, we select DyPO [10] for comparison against our work to show the effect of using weighted bias for multiapplication workload scenarios.…”

“…Both off-line and on-line approaches are confined to extract an optimal resource configuration of a single application running in isolation [3], [10], [12], [18], and are thus not readily adaptable to multiple concurrent application scenarios. While some approaches [21], [22] consider multiple applications, those i) overlook the combination of all the existing knobs (DoP, DVFS, and core selection) in heterogeneous platforms [4], [21], [22], [26], ii) ignore dynamic workload scenarios where applications arrive and leave the system in an unknown manner, limiting their efficiency and adaptability in optimizing resource allocation, and iii) do not consider the weight of total energy consumption per application, restricting those from prioritizing among applications [1], [22], [26]. This subsequently limits the significance of energy gains that can be achieved, since an optimal resource management configuration for one application (obtained from off-line/on-line profiling) which results in lower energy could be non-optimal when another application arrives for • Off-line characterization and modeling of weighted bias and performance for single and multi-threaded applications.…”

A Group-Based Energy-Efficient Dual Priority Scheduling for Real-Time Embedded Systems

“…However, the battery capacity is limited. To ensure the reliability and long time running of real-time embedded system, the research is working on ways to reduce energy consumption, which can reduce system energy demands and improve the reliability and stability of energy supply [4][5][6].…”