PipeFuser: Building Flexible Pipeline Architecture for DNN Accelerators via Layer Fusion
Authors:
- 22 January 2024
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 921-926
Abstract
In this paper, we propose a fused-pipeline architecture that leverages the layer fusion technique to harness the strengths of both non-pipeline and full-pipeline architectures while mitigating their disadvantages. In particular, we observe that the performance of the fused-pipeline accelerators is significantly influenced by the layer fusion strategies and intra-layer mapping schemes. To optimize and rapidly employ the fused-pipeline architecture, we present an end-to-end automation framework, named PipeFuser. At the core of PipeFuser is a genetic algorithm (GA)-based co-design engine, which is used to acquire near-optimal hardware configurations in the vast design space. Experimental results demonstrate that our fused-pipeline architecture achieves 2.3 × to 3.3 × higher performance over the non-pipeline design and 1.9 × to 2.5 × speedup compared to the full-pipeline architecture, with greater deployment flexibility.Keywords
Funding Information
- Research and Development
This publication has 17 references indexed in Scilit:
- Graph-OPU: A Highly Integrated FPGA-Based Overlay Processor for Graph Neural Networks2023
- DeFiNES: Enabling Fast Exploration of the Depth-first Scheduling Space for DNN Accelerators through Analytical Modeling2023
- DeepBurning-SEG: Generating DNN Accelerators of Segment-Grained Pipeline Architecture2022
- Gemmini: Enabling Systematic Deep-Learning Architecture Evaluation via Full-Stack Integration2021
- HybridDNN: A Framework for High-Performance Hybrid DNN Accelerator Design and Implementation2020
- Uni-OPU: An FPGA-Based Uniform Accelerator for Convolutional and Transposed Convolutional NetworksIEEE Transactions on Very Large Scale Integration (VLSI) Systems, 2020
- OPU: An FPGA-Based Overlay Processor for Convolutional Neural NetworksIEEE Transactions on Very Large Scale Integration (VLSI) Systems, 2019
- Toward an Efficient Deep Pipelined Template-Based Architecture for Accelerating the Entire 2-D and 3-D CNNs on FPGAIEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2019
- Fused-layer CNN accelerators2016
- Deep Residual Learning for Image Recognition2016