A Design Framework of Heterogeneous Approximate DCIM-Based Accelerator for Energy-Efficient NN Processing

• Published in: IEEE transactions on circuits and systems. I, Regular papers Vol. 72; no. 8; pp. 3997 - 4006
• Main Authors: Lee, Kyeongho, Lee, Hyeyeong, Park, Jongsun
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.08.2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accuracy; Adders; Approximation; Artificial neural networks; CIM; compute-in-memory; Costs; DCIM generation; Energy consumption; Energy efficiency; Evolutionary algorithms; Genetic algorithms; Hardware; heterogeneous approximation; In-memory computing; Mapping; Neural networks; Power demand; Power management; Sensitivity; SRAM; Static random access memory; Tradeoffs
• ISSN: 1549-8328; 1558-0806
• DOI: 10.1109/TCSI.2025.3530637

Static random-access memory (SRAM) based digital compute-in-memory (DCIM) provides error-resilient computation at the expense of considerable power overhead of adder tree. In recent works, DCIM macro based on approximate computing mitigates the adder tree overheads, however, it faces a trade-off between power and neural network (NN) accuracy. The trade-off becomes more complicated in array-level CIM architecture since output channels of NN model have different sensitivities to approximation errors. In this paper, we propose a heterogeneous approximate DCIM-based accelerator design framework that achieves a good energy-accuracy trade-off for a specific NN model. The framework includes three key features: 1) Evolutionary algorithm-based search finds cost-efficient approximation points by pruning the design space. 2) Genetic algorithm-based channel-wise mapping creates heterogeneous approximation methods that effectively reduce DCIM energy consumption while maintaining high accuracy. 3) A hardware generation strategy decides the number of DCIM macros and their sizes, resulting in an energy-efficient DCIM-based accelerator tailored for the given NN model. Experimental results show that employing the proposed heterogeneous channel-wise mapping significantly enhances the energy efficiency compared to a homogeneous mapping. Moreover, the proposed framework can produce heterogeneous DCIM-based accelerators that consume less energy than state-of-the-art approximate DCIM approaches.