SEGA-DCIM: Design Space Exploration-Guided Automatic Digital CIM Compiler with Multiple Precision Support

2025-05-14

AI Summary

Digital Computing-in-Memory (DCIM) design remains highly manual, inefficient, and time-consuming. Method: This paper proposes the first end-to-end compiler framework supporting automatic integer and floating-point multi-precision mapping for DCIM. It integrates template-based hardware generation, precision-aware mapping, RTL-and-layout co-synthesis, and introduces a novel multi-objective genetic algorithm (MOGA)-driven design space exploration (DSE) mechanism to enable fully automated compilation—from algorithmic specification to physical layout. Contribution/Results: It is the first DCIM compiler to unify multi-precision automatic mapping with joint optimization across hardware layers. Under concurrent area, power, and latency constraints, the framework efficiently explores a vast design space, generating implementations matching or exceeding state-of-the-art hand-designed solutions in performance while drastically reducing design cycle time.

Technology Category

• Machine Learning: Hardware-aware ML
• Search and Optimization: Mixed Discrete/Continuous Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Digital computing-in-memory (DCIM) has been a popular solution for addressing the memory wall problem in recent years. However, the DCIM design still heavily relies on manual efforts, and the optimization of DCIM is often based on human experience. These disadvantages limit the time to market while increasing the design difficulty of DCIMs. This work proposes a design space exploration-guided automatic DCIM compiler (SEGA-DCIM) with multiple precision support, including integer and floating-point data precision operations. SEGA-DCIM can automatically generate netlists and layouts of DCIM designs by leveraging a template-based method. With a multi-objective genetic algorithm (MOGA)-based design space explorer, SEGA-DCIM can easily select appropriate DCIM designs for a specific application considering the trade-offs among area, power, and delay. As demonstrated by the experimental results, SEGA-DCIM offers solutions with wide design space, including integer and floating-point precision designs, while maintaining competitive performance compared to state-of-the-art (SOTA) DCIMs.

Problem

• Automating DCIM design to reduce manual effort
• Optimizing DCIM performance across area, power, delay
• Supporting multiple precision operations in DCIM

Innovation

• Template-based automatic DCIM netlist generation
• MOGA-guided multi-objective design space exploration
• Support for integer and floating-point precision