AoCStream: All-on-Chip CNN Accelerator With Stream-Based Line-Buffer Architecture

TL;DR

AoCStream introduces a stream-based line-buffer architecture for CNN acceleration that significantly reduces memory requirements, enabling efficient on-chip implementation of object detection CNNs on low-end FPGAs without external memory.

Contribution

The paper proposes a novel stream-based line-buffer architecture and accelerator-aware pruning, reducing memory and resource usage for CNNs on low-end FPGAs, unlike traditional frame-based methods.

Findings

Achieves on-chip CNN implementation on low-end FPGA without external memory.

Higher throughput and efficiency compared to similar accelerators.

Reduces intermediate data and weight memory significantly.

Abstract

Convolutional neural network (CNN) accelerators are being widely used for their efficiency, but they require a large amount of memory, leading to the use of a slow and power consuming external memory. This paper exploits two schemes to reduce the required memory amount and ultimately to implement a CNN of reasonable performance only with on-chip memory of a practical device like a low-end FPGA. To reduce the memory amount of the intermediate data, a stream-based line-buffer architecture and a dataflow for the architecture are proposed instead of the conventional frame-based architecture, where the amount of the intermediate data memory is proportional to the square of the input image size. The architecture consists of layer-dedicated blocks operating in a pipelined way with the input and output streams. Each convolutional layer block has a line buffer storing just a few rows of input data. The sizes of the line buffers are proportional to the width of the input image, so the architecture requires less intermediate data storage than the conventional frame-based architecture, especially in the trend of getting larger input size in modern object detection CNNs. In addition to the reduced intermediate data storage, the weight memory is reduced by the accelerator-aware pruning. The experimental results show that a whole object detection CNN can be implemented even on a low-end FPGA without an external memory. Compared to previous accelerators with similar object detection accuracy, the proposed accelerator reaches much higher throughput even with less FPGA resources of LUTs, registers, and DSPs, showing much higher efficiency. The trained models and implemented bit files are available at https://github.com/HyeongjuKang/accelerator-aware-pruning and https://github.com/HyeongjuKang/aocstream.