paper-with-me

Papers

FastMamba: A High-Speed and Efficient Mamba Accelerator on FPGA with Accurate Quantization

2025-05-25 · Aotao Wang, Haikuo Shao, Shaobo Ma, Zhongfeng Wang

State Space Models (SSMs), like recent Mamba2, have achieved remarkable performance and received extensive attention. However, deploying Mamba2 on resource-constrained edge devices encounters many problems: severe outliers within the linear layer challenging the quantization, diverse and irregular element-wise tensor operations, and hardware-unfriendly nonlinear functions in the SSM block. To address these issues, this paper presents FastMamba, a dedicated accelerator on FPGA with hardware-algorithm co-design to promote the deployment efficiency of Mamba2. Specifically, we successfully achieve 8-bit quantization for linear layers through Hadamard transformation to eliminate outliers. Moreover, a hardware-friendly and fine-grained power-of-two quantization framework is presented for the SSM block and convolution layer, and a first-order linear approximation is developed to optimize the nonlinear functions. Based on the accurate algorithm quantization, we propose an accelerator that integrates parallel vector processing units, pipelined execution dataflow, and an efficient SSM Nonlinear Approximation Unit, which enhances computational efficiency and reduces hardware complexity. Finally, we evaluate FastMamba on Xilinx VC709 FPGA. For the input prefill task on Mamba2-130M, FastMamba achieves 68.80\times and 8.90\times speedup over Intel Xeon 4210R CPU and NVIDIA RTX 3090 GPU, respectively. In the output decode experiment with Mamba2-2.7B, FastMamba attains 6\times higher energy efficiency than RTX 3090 GPU.

📄 PDF Abstract BibTeX arXiv:2505.18975

Code (0)

등록된 구현이 없습니다.

Tasks

Computational EfficiencyCPUGPUMambaQuantizationState Space Models

Methods 이 논문이 사용한 방법론

Convolution A convolution is a type of matrix operation, consisting of a kernel, a small matrix of weights, that slides over input data performing element-wise multiplication with the…
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.

Similar Papers 제목 키워드 기반

A CNN Accelerator on FPGA Using Depthwise Separable Convolution

2018-09-03 · Lin Bai, Yiming Zhao, Xinming Huang

Convolutional neural networks (CNNs) have been widely deployed in the fields of computer vision and pattern recognition because of their high accuracy. However, large convolution operations are computing-intensive that o…

CPUGPU

HybridDNN: A Framework for High-Performance Hybrid DNN Accelerator Design and Implementation

2020-04-08 · Hanchen Ye, Xiaofan Zhang, Zhize Huang, Gengsheng Chen 외

To speedup Deep Neural Networks (DNN) accelerator design and enable effective implementation, we propose HybridDNN, a framework for building high-performance hybrid DNN accelerators and delivering FPGA-based hardware imp…

ViM-Q: Scalable Algorithm-Hardware Co-Design for Vision Mamba Model Inference on FPGA

2026-05-03 · Shengzhe Lyu, Yuhan She, Patrick S. Y. Hung, Ray C. C. Cheung 외 arxiv

Vision Mamba (ViM) models offer a compelling efficiency advantage over Transformers by leveraging the linear complexity of State Space Models (SSMs), yet efficiently deploying them on FPGAs remains challenging. Linear la…

Synetgy: Algorithm-hardware Co-design for ConvNet Accelerators on Embedded FPGAs

2018-11-21 · Yifan Yang, Qijing Huang, Bichen Wu, Tianjun Zhang 외

Using FPGAs to accelerate ConvNets has attracted significant attention in recent years. However, FPGA accelerator design has not leveraged the latest progress of ConvNets. As a result, the key application characteristics…

High-Level Synthesis

OpenCL-based FPGA accelerator for disparity map generation with stereoscopic event cameras

2019-03-08 · David Castells-Rufas, Jordi Carrabina

Although event-based cameras are already commercially available. Vision algorithms based on them are still not common. As a consequence, there are few Hardware Accelerators for them. In this work we present some experime…

Stereo MatchingStereo Matching Hand