DAM SRAM CORE: An Efficient High-Speed and Low-Power CIM SRAM CORE Design for Feature Extraction Convolutional Layers in Binary Neural Networks.
In: Micromachines, Jg. 15 (2024-04-30), Heft 5
Online
academicJournal
Zugriff:
This article proposes a novel design for an in-memory computing SRAM, the DAM SRAM CORE, which integrates storage and computational functionality within a unified 11T SRAM cell and enables the performance of large-scale parallel Multiply-Accumulate (MAC) operations within the SRAM array. This design not only improves the area efficiency of the individual cells but also realizes a compact layout. A key highlight of this design is its employment of a dynamic aXNOR-based computation mode, which significantly reduces the consumption of both dynamic and static power during the computational process within the array. Additionally, the design innovatively incorporates a self-stabilizing voltage gradient quantization circuit, which enhances the computational accuracy of the overall system. The 64 × 64 bit DAM SRAM CORE in-memory computing core was fabricated using the 55 nm CMOS logic process and validated via simulations. The experimental results show that this core can deliver 5-bit output results with 1-bit input feature data and 1-bit weight data, while maintaining a static power consumption of 0.48 mW/mm 2 and a computational power consumption of 11.367 mW/mm 2 . This showcases its excellent low-power characteristics. Furthermore, the core achieves a data throughput of 109.75 GOPS and exhibits an impressive energy efficiency of 21.95 TOPS/W, which robustly validate the effectiveness and advanced nature of the proposed in-memory computing core design.
Titel: |
DAM SRAM CORE: An Efficient High-Speed and Low-Power CIM SRAM CORE Design for Feature Extraction Convolutional Layers in Binary Neural Networks.
|
---|---|
Autor/in / Beteiligte Person: | Zhao, R ; Gong, Z ; Liu, Y ; Chen, J |
Link: | |
Zeitschrift: | Micromachines, Jg. 15 (2024-04-30), Heft 5 |
Veröffentlichung: | Basel, Switzerland : MDPI, [2010]-, 2024 |
Medientyp: | academicJournal |
ISSN: | 2072-666X (print) |
DOI: | 10.3390/mi15050617 |
Sonstiges: |
|