34.5 A 818-4094TOPS/W Capacitor-Reconfigured CIM Macro for Unified Acceleration of CNNs and Transformers
Author:
- 18 February 2024
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 67, 574-576
Abstract
In the rapidly evolving landscape of machine learning, workloads using diverse neural-network architectures must be covered: including CNNs for image processing, transformers for natural language processing (NLP), and hybrid architectures that blend CNNs and transformers for audio processing. As illustrated in Fig. 34.5.1, these varied architectures have unique computational precision requirements. While CNNs achieve satisfactory accuracy even with low-computational precision, compute SNR or CSNR[1], transformers require higher CSNR to reach their full potential. This diversity amplifies the need for versatile hardware accelerators that can efficiently handle both CNNs and transformers, while meeting the multifaceted demands of modern machine-learning applications.Keywords
This publication has 6 references indexed in Scilit:
- 7.6 A 70.85-86.27TOPS/W PVT-Insensitive 8b Word-Wise ACIM with Post-Processing Relaxation2023
- A Fully Bit-Flexible Computation in Memory Macro Using Multi-Functional Computing Bit Cell and Embedded Input Sparsity SensingIEEE Journal of Solid-State Circuits, 2023
- Fully Row/Column-Parallel In-memory Computing SRAM Macro employing Capacitor-based Mixed-signal Computation with 5-b Inputs2021
- Fundamental limits on the precision of in-memory architectures2020
- A Programmable Heterogeneous Microprocessor Based on Bit-Scalable In-Memory ComputingIEEE Journal of Solid-State Circuits, 2020
- 15.3 A 351TOPS/W and 372.4GOPS Compute-in-Memory SRAM Macro in 7nm FinFET CMOS for Machine-Learning Applications2020