We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Hardware Architecture

Title: S2Engine: A Novel Systolic Architecture for Sparse Convolutional Neural Networks

Abstract: Convolutional neural networks (CNNs) have achieved great success in performing cognitive tasks. However, execution of CNNs requires a large amount of computing resources and generates heavy memory traffic, which imposes a severe challenge on computing system design. Through optimizing parallel executions and data reuse in convolution, systolic architecture demonstrates great advantages in accelerating CNN computations. However, regular internal data transmission path in traditional systolic architecture prevents the systolic architecture from completely leveraging the benefits introduced by neural network sparsity. Deployment of fine-grained sparsity on the existing systolic architectures is greatly hindered by the incurred computational overheads. In this work, we propose S2Engine $-$ a novel systolic architecture that can fully exploit the sparsity in CNNs with maximized data reuse. S2Engine transmits compressed data internally and allows each processing element to dynamically select an aligned data from the compressed dataflow in convolution. Compared to the naive systolic array, S2Engine achieves about $3.2\times$ and about $3.0\times$ improvements on speed and energy efficiency, respectively.
Comments: 13 pages, 17 figures
Subjects: Hardware Architecture (cs.AR); Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
Journal reference: IEEE Transactions on Computers, 2021
DOI: 10.1109/TC.2021.3087946
Cite as: arXiv:2106.07894 [cs.AR]
  (or arXiv:2106.07894v1 [cs.AR] for this version)

Submission history

From: Jianlei Yang [view email]
[v1] Tue, 15 Jun 2021 06:08:37 GMT (3679kb,D)

Link back to: arXiv, form interface, contact.