Abstract:Byte-addressable non-volatile memory (NVM) features high density, DRAM comparable performance, and persistence. These characteristics position NVM as a promising new tier in the memory hierarchy. Nevertheless, NVM has asymmetric read and write performance, and considerably higher write energy than DRAM. Our work provides an in-depth evaluation of the first commercially available byte-addressable NVM -- the Intel Optane DC persistent memory. The first part of our study quantifies the latency, bandwidth, power efficiency, and energy consumption under eight memory configurations. We also evaluate the real impact on in-memory graph processing workloads. Our results show that augmenting NVM with DRAM is essential, and the combination can effectively bridge the performance gap and provide reasonable performance with higher capacity. We also identify NUMA-related performance characteristics for accesses to memory on a remote socket. In the second part, we employ two fine-grained allocation policies to control traffic distribution between DRAM and NVM. Our results show that bandwidth spilling between DRAM and NVM could provide 2.0x bandwidth and enable $20\%$ larger problems than using DRAM as a cache. Also, write isolation between DRAM and NVM could save up to 3.9x energy and improves bandwidth by 3.1x compared to DRAM-cached NVM. We establish a roofline model to explore power and energy efficiency at various distributions of read-only traffic. Our results show that NVM requires 1.8x lower power than DRAM for data-intensive workloads. Overall, applications can significantly optimize performance and power efficiency by adapting traffic distribution to NVM and DRAM through memory configurations and fine-grained policies to fully exploit the new memory device.

Bridging the Performance Gap for Copy-Based Garbage Collectors Atop Non-Volatile Memory

PVM: A Memory Management under Vertical Architecture

Demystifying the Performance of HPC Scientific Applications on NVM-based Memory Systems

Wear Leveling for Non-Volatile Memory: a Runtime System Approach

Performance Analysis and Optimization of Full Garbage Collection in Memory-hungry Environments

Analysis And Optimizations Of Java Full Garbage Collection

ScissorGC: Scalable and Efficient Compaction for Java Full Garbage Collection

GCPersist: an efficient GC-assisted lazy persistency framework for resilient Java applications on NVM.

Design and Evaluation of an Efficient Semispace-based Memory Reclamation Scheme

Preserving Addressability Upon GC-Triggered Data Movements on Non-Volatile Memory

UHNVM: A Universal Heterogeneous Cache Design with Non-Volatile Memory

Hardware-assisted Middleware: Acceleration of Garbage Collection Operations

Redesign the Memory Allocator for Non-Volatile Main Memory.

NGraph: Parallel Graph Processing in Hybrid Memory Systems

Memory That Never Forgets: Emerging Nonvolatile Memory and the Implication for Architecture Design

Virtual Page Behavior Based Page Management Policy for Hybrid Main Memory in Cloud Computing.

Efficient Memory Management for NVM-Based Hybrid Memory Systems

System Evaluation of the Intel Optane Byte-addressable NVM

Memos: Revisiting Hybrid Memory Management in Modern Operating System

A Performance Study of Software Prefetching for Tracing Garbage Collectors.

VAIL: A Victim-Aware Cache Policy to Improve NVM Lifetime for Hybrid Memory System.