Zimeng Fan

Zimeng Fan

Auto-generated profile, based on publicly available information

Zimeng Fan

School of Computer Science, Wuhan University

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
11 March 2024 in Applied Soft Computing
Verified
Journal Article
Closed access
FLFang Liu
Fang Liu
ZFZimeng Fan
Zimeng Fan
WHWei Hu
Wei Hu
DXDian Xu
Dian Xu
MPMin Peng
Min Peng
YHYanxiang He
Yanxiang He

15 January 2024 in ACM Transactions on Reconfigurable Technology and Systems
Verified
Journal Article
Open access
ZFZimeng Fan
Zimeng Fan
WHWei Hu
Wei Hu
FLFang Liu
Fang Liu
DXDian Xu
Dian Xu
HGHong Guo
Hong Guo
Min PengMin Peng
Min Peng
In computer vision, the joint development of the algorithm and computing dimensions cannot be separated. Models and algorithms are constantly evolving, while hardware designs must adapt to new or updated algorithms. Reconfigurable devices are recognized as important platforms for computer vision applications because of their reconfigurability. There are two typical design approaches: customized and overlay design. However, existing work is unable to achieve both efficient performance and scalability to adapt to a wide range of models. To address both considerations, we propose a design framework based on reconfigurable devices to provide unified support for computer vision models. It provides software-programmable modules while leaving unit design space for problem-specific algorithms. Based on the proposed framework, we design a model mapping method and a hardware architecture with two processor arrays to enable dynamic and static reconfiguration, thereby relieving redesign pressure. In addition, resource consumption and efficiency can be balanced by adjusting the hyperparameter. In experiments on CNN, vision Transformer, and vision MLP models, our work’s throughput is improved by 18.8x–33.6x and 1.4x–2.0x compared to CPU and GPU. Compared to others on the same platform, accelerators based on our framework can better balance resource consumption and efficiency.

19 July 2022
Verified
Book Chapter
Closed access
PDF file
FHFeihu Huang
Feihu Huang
MJMin Jiang
Min Jiang
FLFang Liu
Fang Liu
DXDian Xu
Dian Xu
ZFZimeng Fan
Zimeng Fan
YWYonghao Wang
Yonghao Wang

18 August 2025 in ACM Transactions on Reconfigurable Technology and Systems
Verified
Journal Article
Open access
ZFZimeng Fan
Zimeng Fan
Min PengMin Peng
Min Peng
Graph Neural Networks (GNNs) have seen considerable advancements across various applications. However, the computationally and storage-intensive nature of GNNs presents unique challenges for hardware design. Existing GNN accelerators frequently grapple with low execution efficiency and workload imbalance. Therefore, this article introduces DGMF, a unified dynamic mapping approach that optimizes dataflow and computation mode based on GNN models and datasets. To support this dynamic mapping approach, we also present a dedicated hardware architecture. This architecture dynamically adjusts resource allocation and parallelism based on models and datasets, thereby ensuring optimal utilization. Additionally, we propose a Design Space Exploration (DSE) algorithm that traverses the defined design space to identify the optimal design solution, effectively managing diverse design constraints and objectives. To validate our approach, we developed an accelerator based on three datasets and five types of GNN models. Experimental results demonstrate that DGMF achieves performance by 10.5×–2,548×, and 1.01×–51.7× compared to GPUs and existing accelerators. It also improves resource efficiency, energy efficiency, and DRAM access by 0.87×–1.47×, 0.92×–2.18×, and 1.01×–47.25×, respectively, compared to other works.

14 January 2026 in ACM Transactions on Internet Technology
Verified
Journal Article
Open access
ZFZimeng Fan
Zimeng Fan
Min PengMin Peng
Min Peng
AI has found extensive application in edge analytics, with Deep Neural Networks effectively processing and analyzing data. Some applications need to deal with unstructured data like graphs. This complexity and diversity of unstructured features pose challenges for traditional DNN models. Therefore, Graph Neural Networks have emerged as a significant solution. However, GNNs encounter two primary challenges in edge analytics scenarios: scalability and efficiency. These challenges arise primarily from significant structural differences among various graphs and models. To address these issues, we propose AGF, an Adaptive GNN Framework. AGF consists of three components: subgraph partitioning and model partitioning, scalable hardware architecture, and design space exploration. Subgraph partitioning and model partitioning transform complex and diverse GNNs into a uniform computing flow and subgraphs. These can be directly mapped onto the hardware kernel, leveraging GNN characteristics to achieve independent parallelism. The hardware architecture is based on a multi-level processing element structure, which can efficiently parallelize subgraphs and vertices based on the partitioning. The design space exploration selects kernel allocation strategies and setups based on the hardware, dataset, and model. We validated AGF using six datasets and three GNN models. The results demonstrate that our framework achieves an acceleration ranging from 4.54 to 53.17× compared to the GPU-based GNN framework. When compared to other FPGA-based GNN accelerators, AGF achieves a latency reduction ranging from 1.16 to 33.6× in model execution and from 1.09 to 2.37× in end-to-end scenarios.

1 June 2022
Verified
Conference Paper
Closed access
DXDian Xu
Dian Xu
WHWei Hu
Wei Hu
FLFang Liu
Fang Liu
ZFZimeng Fan
Zimeng Fan
QSQingsong Shi
Qingsong Shi
Recently, the Transformer-based models have achieved leading results in many research areas such as natural language processing and computer vision. However, since Transformer-based models have a huge computational complexities, optimization for Transformer is the focus of current research. the core of Transformer models is the attention module, and in this paper we propose a software and hardware co-designed attention module, which reduces the computational units of the multi-head attention module by balancing the The inference time of the attention module is reduced by balancing the computational units of multiple attention modules. At the same time, we design the corresponding structure on FPGA, which accelerate 14.6x inference time in comparison with the original attention module implementation on GPU. We applied this attention module to multiple visual Transformer models and tested the accuracy difference between the balanced multi-head attention module-based model and the original model on CIFAR-10, and obtained a result of 80.46%, which lower from the original version by 0.18%.