CIM-Tuner: Balancing the Compute and Storage Capacity of SRAM-CIM Accelerator via Hardware-mapping Co-exploration

TL;DR

CIM-Tuner is an automatic tool that optimizes hardware mapping for SRAM-CIM accelerators, balancing compute and storage to significantly improve energy efficiency and throughput under area constraints.

Contribution

It introduces a hardware-mapping co-exploration framework with a generalized template and two-level strategies, enabling universal and optimized CIM accelerator configurations.

Findings

Achieves 1.58× higher energy efficiency

Achieves 2.11× higher throughput

Provides improvements across various CIM designs

Abstract

As an emerging type of AI computing accelerator, SRAM Computing-In-Memory (CIM) accelerators feature high energy efficiency and throughput. However, various CIM designs and under-explored mapping strategies impede the full exploration of compute and storage balancing in SRAM-CIM accelerator, potentially leading to significant performance degradation. To address this issue, we propose CIM-Tuner, an automatic tool for hardware balancing and optimal mapping strategy under area constraint via hardware-mapping co-exploration. It ensures universality across various CIM designs through a matrix abstraction of CIM macros and a generalized accelerator template. For efficient mapping with different hardware configurations, it employs fine-grained two-level strategies comprising accelerator-level scheduling and macro-level tiling. Compared to prior CIM mapping, CIM-Tuner's extended strategy space achieves 1.58$\times$ higher energy efficiency and 2.11$\times$ higher throughput. Applied to SOTA CIM accelerators with identical area budget, CIM-Tuner also delivers comparable improvements. The simulation accuracy is silicon-verified and CIM-Tuner tool is open-sourced at https://github.com/champloo2878/CIM-Tuner.git.