AI Navigate

データレート認識型高速CNN推論をFPGA上で実現

arXiv cs.LG / 2026/3/11

Ideas & Deep Analysis

要点

  • 本論文は、プーリングやストライド畳み込みなど、データ出力率を低下させる層によって引き起こされるFPGAベースのCNNアクセラレータの資源未活用問題に対処している。
  • 設計空間探索を活用し、ハードウェア資源の利用効率を最適化しつつ連続したデータフローを維持するデータレート認識型アクセラレータアーキテクチャを紹介している。
  • 実験結果は従来設計と比較して算術資源の使用量を大幅に削減し、幅広いデータレートにわたって単一FPGA上で複雑なCNNを効率的に展開可能であることを示した。
  • 各層固有のデータレートに応じて構成を適応させることでハードウェア効率を向上させ、処理中の全ハードウェアユニットの稼働を確保している。

Computer Science > Hardware Architecture

arXiv:2603.08726 (cs)
[Submitted on 18 Feb 2026]

Title:Data-Rate-Aware High-Speed CNN Inference on FPGAs

View a PDF of the paper titled Data-Rate-Aware High-Speed CNN Inference on FPGAs, by Tobias Habermann and 1 other authors
View PDF HTML (experimental)
Abstract:Dataflow-based CNN accelerators on FPGAs achieve low latency and high throughput by mapping computations of each layer directly to corresponding hardware units. However, layers such as pooling and strided convolutions reduce the data at their output with respect to their input, strongly effecting the data rate of the following layers. This leads to underutilization in fully unrolled designs. While prior work introduced data-rate-aware layer-wise adaptation, determining the most efficient implementation remains challenging.
This paper presents a data-rate-aware CNN accelerator architecture for multi-pixel processing. Building on existing analytical models, the proposed method performs design-space exploration to identify configurations that improve hardware utilization and resource efficiency while preserving continuous flow of data, keeping all hardware units busy. Experimental results show substantial reductions in arithmetic resources compared to previous designs, enabling efficient implementation of complex CNNs on a single FPGA across a wide range of data rates.
Subjects: Hardware Architecture (cs.AR); Machine Learning (cs.LG)
Cite as: arXiv:2603.08726 [cs.AR]
  (or arXiv:2603.08726v1 [cs.AR] for this version)
  https://doi.org/10.48550/arXiv.2603.08726
Focus to learn more
arXiv-issued DOI via DataCite

Submission history

From: Tobias Habermann [view email]
[v1] Wed, 18 Feb 2026 08:46:38 UTC (112 KB)
Full-text links:

Access Paper:

Current browse context:
cs.AR
< prev   |   next >
Change to browse by:

References & Citations

export BibTeX citation Loading...

BibTeX formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Reddit logo
Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle
Bibliographic Explorer (What is the Explorer?)
Connected Papers Toggle
Connected Papers (What is Connected Papers?)
Litmaps Toggle
Litmaps (What is Litmaps?)
scite.ai Toggle
scite Smart Citations (What are Smart Citations?)
Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle
alphaXiv (What is alphaXiv?)
Links to Code Toggle
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub Toggle
DagsHub (What is DagsHub?)
GotitPub Toggle
Gotit.pub (What is GotitPub?)
Huggingface Toggle
Hugging Face (What is Huggingface?)
Links to Code Toggle
Papers with Code (What is Papers with Code?)
ScienceCast Toggle
ScienceCast (What is ScienceCast?)
Demos

Demos

Replicate Toggle
Replicate (What is Replicate?)
Spaces Toggle
Hugging Face Spaces (What is Spaces?)
Spaces Toggle
TXYZ.AI (What is TXYZ.AI?)
Related Papers

Recommenders and Search Tools

Link to Influence Flower
Influence Flower (What are Influence Flowers?)
Core recommender toggle
CORE Recommender (What is CORE?)
About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.