Memory Controller Design Under Cloud Workloads

TL;DR

This paper evaluates various memory controller designs under cloud workloads, revealing that simple scheduling policies perform well and increasing memory channels offers minimal benefits, guiding future controller optimization.

Contribution

It provides an empirical analysis of memory controller behaviors with scale-out workloads, highlighting the effectiveness of simple scheduling policies and limited gains from multiple memory channels.

Findings

FR-FCFS outperforms other scheduling policies

Simple FCFS scheduling is nearly as effective as FR-FCFS

Adding more memory channels yields minimal performance improvements

Abstract

This work studies the behavior of state-of-the-art memory controller designs when executing scale-out workloads. It considers memory scheduling techniques, memory page management policies, the number of memory channels, and the address mapping scheme used. Experimental measurements demonstrate: 1)~Several recently proposed memory scheduling policies are not a good match for these scale-out workloads. 2)~The relatively simple First-Ready-First-Come-First-Served (FR-FCFS) policy performs consistently better, and 3)~for most of the studied workloads, the even simpler First-Come-First-Served scheduling policy is within 1\% of FR-FCFS. 4)~Increasing the number of memory channels offers negligible performance benefits, e.g., performance improves by 1.7\% on average for 4-channels vs. 1-channel. 5)~77\%-90\% of DRAM rows activations are accessed only once before closure. These observation can guide future development and optimization of memory controllers for scale-out workloads.