CoChat

Eyeriss: a spatial architecture for energy-efficient dataflow for convolutional neural networks

Chen, Yu-Hsin, Emer, Joel, Sze, Vivienne · 2016 · ACM SIGARCH Computer Architecture News

The DOI matches its Crossref record; 45 of the 48 references it cites were found in a registry or an index.

Abstract

Deep convolutional neural networks (CNNs) are widely used in modern AI systems for their superior accuracy but at the cost of high computational complexity. The complexity comes from the need to simultaneously process hundreds of filters and channels in the high-dimensional convolutions, which involve a significant amount of data movement. Although highly-parallel compute paradigms, such as SIMD/SIMT, effectively address the computation requirement to achieve high throughput, energy consumption still remains high as data movement can be more expensive than computation. Accordingly, finding a dataflow that supports parallel processing with minimal data movement cost is crucial to achieving energy-efficient CNN processing without compromising accuracy. In this paper, we present a novel dataflow, called row-stationary (RS), that minimizes data movement energy consumption on a spatial architecture. This is realized by exploiting local data reuse of filter weights and feature map pixels, i.e., activations, in the high-dimensional convolutions, and minimizing data movement of partial sum accumulations. Unlike dataflows used in existing designs, which only reduce certain types of data movement, the proposed RS dataflow can adapt to different CNN shape configurations and reduces all types of data movement through maximally utilizing the processing engine (PE) local storage, direct inter-PE communication and spatial parallelism. To evaluate the energy efficiency of the different dataflows, we propose an analysis framework that compares energy cost under the same hardware area and processing parallelism constraints. Experiments using the CNN configurations of AlexNet show that the proposed RS dataflow is more energy efficient than existing dataflows in both convolutional (1.4× to 2.5×) and fully-connected layers (at least 1.3× for batch size larger than 16). The RS dataflow has also been demonstrated on a fabricated chip, which verifies our energy analysis.

Overview

Integrity signals

Registry

Crossref · 10.1145/3007787.3001177

Cites

45 of 45 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Verification

Claims the paper makes

This paper’s full text isn’t openly available, so its claims and self-check weren’t run. The registry and citation checks still apply.

Citations

Cites

Show all 45 references

Research works printed but not linked: 1

  1. S. Han H. Mao and W. J. Dally "Deep Compression: Compressing Deep Neural Network with Pruning Trained Quantization and Huffman Coding " in ICLR 2016. S. Han H. Mao and W. J. Dally "Deep Compression: Compressing Deep Neural Network with Pruning Trained Quantization and Huffman Coding " in ICLR 2016.

Cited by

See all 1,182

Metadata

Authors
Chen, Y.-H., Emer, J., & Sze, V.
Affiliations
  • Massachusetts Institute of Technology (US)
  • NVIDIA Research, NVIDIA
Venue
ACM SIGARCH Computer Architecture News
Published
2016-06-18
Volume
44
Issue
3
Pages
367-379
ISSN
0163-5964 (ISSN-L)
OpenAlex
W2442974303
Cited by
1,182
References
48
Citations by year
2026: 35 · 2025: 65 · 2024: 88 · 2023: 102 · 2022: 126 · 2021: 195 · 2020: 195 · 2019: 166 · 2018: 131 · 2017: 72 · 2016: 7
FWCI
189.92
Type
journal article
Language
en
Open access
closed
Locations
ACM SIGARCH Computer Architecture News (publishedVersion)
Topics
  • Physical Sciences › Computer Science › Hardware and Architecture › Parallel Computing and Optimization Techniques
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
  • Physical Sciences › Computer Science › Hardware and Architecture › Embedded Systems Design Techniques
Keywords
spatial parallelism, AlexNet

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Eyeriss: a spatial architecture for energy-efficient dataflow for convolutional neural networks. https://app.cochat.ai/papers/10.1145/3007787.3001177
The paper
Chen, Y.-H., Emer, J., & Sze, V. (2016). Eyeriss: a spatial architecture for energy-efficient dataflow for convolutional neural networks. ACM SIGARCH Computer Architecture News. https://doi.org/10.1145/3007787.3001177

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.