CHIME: Energy-Efficient STT-RAM-based Concurrent Hierarchical In-Memory Processing

TL;DR

CHIME is a novel hierarchical in-memory processing architecture using STT-RAM, strategically placing compute units across memory levels to improve performance and energy efficiency for domain-specific workloads.

Contribution

It introduces a multi-level PiC/PiM architecture with heterogeneous compute units, optimizing data locality, performance, and energy consumption, which surpasses single-level designs.

Findings

Achieves 57.95% speedup over state-of-the-art approaches.

Reduces energy consumption by 78.23%.

Enhances compute unit utilization and concurrency.

Abstract

Processing-in-cache (PiC) and Processing-in-memory (PiM) architectures, especially those utilizing bit-line computing, offer promising solutions to mitigate data movement bottlenecks within the memory hierarchy. While previous studies have explored the integration of compute units within individual memory levels, the complexity and potential overheads associated with these designs have often limited their capabilities. This paper introduces a novel PiC/PiM architecture, Concurrent Hierarchical In-Memory Processing (CHIME), which strategically incorporates heterogeneous compute units across multiple levels of the memory hierarchy. This design targets the efficient execution of diverse, domain-specific workloads by placing computations closest to the data where it optimizes performance, energy consumption, data movement costs, and area. CHIME employs STT-RAM due to its various advantages in PiC/PiM computing, such as high density, low leakage, and better resiliency to data corruption from activating multiple word lines. We demonstrate that CHIME enhances concurrency and improves compute unit utilization at each level of the memory hierarchy. We present strategies for exploring the design space, grouping, and placing the compute units across the memory hierarchy. Experiments reveal that, compared to the state-of-the-art bit-line computing approaches, CHIME achieves significant speedup and energy savings of 57.95% and 78.23% for various domain-specific workloads, while reducing the overheads associated with single-level compute designs.