Synergy of Dynamic Frequency Scaling and Demotion on DRAM Power Management: Models and Optimizations

Abstract

Main memory (or DRAM) is one of the most significant components to the computer system's performance and energy consumption. Dynamic frequency scaling (DFS) and DRAM low-power states (Demotion) are two main-stream techniques for DRAM power management. DFS reduces the operation frequency of memory channels and DRAM devices when the memory bandwidth is under-utilized, whereas demotion transits individual memory ranks to low-power states during long idle periods. Despite that there have been fruitful research work for DFS and demotion separately, little attention has been paid to the synergy between these two techniques. To bridge this gap, this paper conducts a comprehensive study on the synergy between DFS and demotion. In particular, we leverage queuing theory to develop analytical models for the energy consumption and performance of DRAM systems with DFS and demotion. These models provide valuable insights into the synergy between DFS and demotion. We further attempt to minimize the energy consumption by considering both DFS and demotion while keeping a pre-defined performance penalty budget. To reduce the optimization complexity, we develop simple yet effective heuristics to search near-optimum DFS-demotion configurations. We experimentally compare our design with other state-of-the-art DRAM energy saving policies using detailed simulations of a large set of workloads. Experimental results show the accuracy of our analytical models and the effectiveness of our optimizations.Index Terms-Demotion, dynamic frequency scaling, energy consumption, main memory systems, in-memory processing Ç 1 INTRODUCTION E NERGY consumption has become a major factor in the design and implementation of modern computer systems. In many systems, main memory (or DRAM) is a critical component for the performance and energy consumption. As processors have moved to a multi-/many-core era, more applications run simultaneously with their working sets in the main memory. Emerging applications such as in-memory data analytics [1] and RAMCloud [2] boost the memory capacity of modern computing systems. Such hunger for main memory of larger capacity makes the amount of energy consumed by main memory approaching or even surpassing that consumed by processors in many servers [3], [4]. For example, it has been reported that main memory contributes to as much as 40-46 percent of total energy consumption in server applications [5], [6]. For these reasons, this paper investigates whether and how we can leverage DRAM power management techniques to reduce the energy consumption of main memory.There have been various energy saving techniques on exploiting the power management capability of main memory. Two important and common techniques are DRAM low-power states (Demotion) [7], [8], [9], [10] and dynamic frequency scaling (DFS) [11], [12]. The common theme of demotions is to transit individual memory ranks to low-power states during (long) idle periods, whereas DFS dynamically scales the operation frequency of memory channels and DRAM devices to make the...

Dynamic Frequency Scaling Regarding Memory for Energy Efficiency of Embedded Systems

“…The results of [7] were further extended to consider both CPU and DRAM power consumption in a server [8]. In [9], power management of DRAM using both DFS and low-power states is modeled and studied using simulation. The joint scaling of CPU and DRAM frequencies was also studied in [10] for server systems.…”

“…Switching activities may differ from each other. On the other hand, the power consumption of the DRAM system can be divided into operation power and background power [6], [9]. The operation power is the power required to execute memory reads and writes.…”

A Survey of Non-Volatile Main Memory Technologies: State-of-the-Arts, Practices, and Future Directions

“…Although DRAM scal-ing is continued from 28 nm in 2013 to 10+ nm in 2016 [4,5] , the scaling has slowed down and become more and more difficult. Moreover, recent studies [6][7][8][9][10] have showed that DRAM-based main memory accounts for about 30%-40% of the total energy consumption of a physical server.…”