Results show that weight scaling outperforms PBA, while layer-wise bias optimization increases the accuracy from 80.33% to 84.13% without modifying the datapath width.
Abstract
This paper investigates reduced-precision floating-point arithmetic for FFT-based CNN inference on an industrial CPU-FPGA platform. We combine FFT-based convolution with adaptive post-training FP8 quantization and evaluate two FPGA-oriented optimization methods: progressive bias adjustment (PBA) within the FFT and layer-wise exponent-bias selection across the CNN. The methods are implemented in a LeNet-5 accelerator using serial radix-$2^2$ SDF FFT modules and evaluated on an industrial fault detection dataset. Results show that weight scaling outperforms PBA, while layer-wise bias optimization increases the accuracy from 80.33% to 84.13% without modifying the datapath width. Compared with CPU-only inference, the FPGA achieves approximately 2.5$\times$ higher energy efficiency.
Energy efficiency is a critical design constraint in FPGA-based embedded systems that deploy convolutional neural networks (CNNs). Rapid and accurate exploration methods facilitate optimized implementations. This paper proposes a frequency-scalable power modeling framework for CNN intel-lectual property (IP) libraries...
Philippe Magalhães, Virginie Fresse, Olivier Alata et al.· Symposium on Integrated Circ...· 0 citations
This paper proposes an optimization methodology for convolutional neural networks (CNNs) on field programmable gate arrays (FPGAs) under strict latency and hardware constraints. The approach integrates CNN model restructuring, hardware-aware scheduling, and FPGA-based hardware-software co-design to reduce computational...
Serkan Sağlam, Mark Zwolinski· IEEE Access· 0 citations
The growth of Convolutional Neural Networks (CNNs) in applications such as image classification and real-time object detection is limited by high computational and energy costs, especially for resource-constrained embedded systems. To address this challenge, this paper presents a fully open-source and flexible hardware...
Fabricio Lorenzon, P. Pereira, M. Grellert· Symposium on Integrated Circ...· 0 citations
As CPUs alone are often insufficiently efficient for running Vision Transformers (ViTs), contemporary computing platforms, particularly edge systems, increasingly adopt heterogeneous architectures that combine CPUs with GPUs and Neural Processing Units (NPUs). This paper presents a detailed energy assessment of individ...
Li-Xian Ho, C. Kachris· Applied Sciences· 0 citations
Fast convolution is widely used to reduce the arithmetic cost of small-kernel CNN convolutions. However, its hardware efficiency depends not only on the number of transform-domain multiplications, but also on the complexity of the transform stages. This work proposes Inspection Factorization (IF) as an alternative to c...
T. O. Da Silva, F. Moraes· Symposium on Integrated Circ...· 0 citations
Direct-P is introduced for noncausal inference and a causal path that passes the forward quantization directly into backward into forward quantization is addressed, accelerating a complete single-GPU 8-billion-parameter update by up to 1.14$\times$.
Robert Hu· 2 citations· ⚡1
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
