Ibrahim, Muhammad Sohail and Usman, Muhammad and Lee, Jeong-A (2025) USEFUSE: Uniform stride for enhanced performance in fused layer architecture of deep neural networks. JOURNAL OF SYSTEMS ARCHITECTURE, 166: 103459. ISSN 1383-7621, 1873-6165
Full text not available from this repository. (Request a copy)Abstract
Convolutional Neural Networks (CNNs) are crucial in various applications, but deploying them on resource-constrained edge devices poses challenges. This study presents the Sum-of-Products (SOP) units for convolution, which utilize low-latency left-to-right bit-serial arithmetic to minimize response time and enhance overall performance. The study proposes a methodology for fusing multiple convolution layers to reduce off-chip memory communication and increase the overall performance. An effective mechanism detects and skips inefficient convolutions after ReLU layers, minimizing power consumption without compromising accuracy. Additionally, efficient tile movement guarantees uniform access to the fusion pyramid. An analysis demonstrates the uniform stride strategy improves operational intensity. Two designs cater to varied demands: one focuses on minimal response time for mission-critical applications, and another focuses on resource-constrained devices with comparable latency. This approach notably reduced redundant computations, improving the efficiency of CNN deployment on edge devices.
| Item Type: | Article |
|---|---|
| Uncontrolled Keywords: | EFFICIENT; ACCELERATOR; DESIGN; Convolution neural network; Online arithmetic; Most-significant-digit-first arithmetic; CNN acceleration; Layer fusion |
| Subjects: | 000 Computer science, information & general works > 004 Computer science |
| Divisions: | Informatics and Data Science > Department Computational Life Science > Chair of Image Analysis and Computer Vision (Prof. Dr.-Ing. Dorit Merhof) |
| Depositing User: | Dr. Gernot Deinzer |
| Date Deposited: | 18 Jun 2026 08:26 |
| Last Modified: | 18 Jun 2026 08:26 |
| URI: | https://pred.uni-regensburg.de/id/eprint/66568 |
Actions (login required)
 |
View Item |
