Saved in:
Bibliographic Details
Main Authors: Han, Ruichi, Chen, Yizhi, Lei, Tong, Gonzalez, Jordi Altayo, Hemani, Ahmed
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.14087
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915740741795840
author Han, Ruichi
Chen, Yizhi
Lei, Tong
Gonzalez, Jordi Altayo
Hemani, Ahmed
author_facet Han, Ruichi
Chen, Yizhi
Lei, Tong
Gonzalez, Jordi Altayo
Hemani, Ahmed
contents Interconnect power consumption remains a bottleneck in Deep Neural Network (DNN) accelerators. While ordering data based on '1'-bit counts can mitigate this via reduced switching activity, practical hardware sorting implementations remain underexplored. This work proposes the hardware implementation of a comparison-free sorting unit optimized for Convolutional Neural Networks (CNN). By leveraging approximate computing to group population counts into coarse-grained buckets, our design achieves hardware area reductions while preserving the link power benefits of data reordering. Our approximate sorting unit achieves up to 35.4% area reduction while maintaining 19.50\% BT reduction compared to 20.42% of precise implementation.
format Preprint
id arxiv_https___arxiv_org_abs_2601_14087
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle '1'-bit Count-based Sorting Unit to Reduce Link Power in DNN Accelerators
Han, Ruichi
Chen, Yizhi
Lei, Tong
Gonzalez, Jordi Altayo
Hemani, Ahmed
Hardware Architecture
Artificial Intelligence
Machine Learning
Interconnect power consumption remains a bottleneck in Deep Neural Network (DNN) accelerators. While ordering data based on '1'-bit counts can mitigate this via reduced switching activity, practical hardware sorting implementations remain underexplored. This work proposes the hardware implementation of a comparison-free sorting unit optimized for Convolutional Neural Networks (CNN). By leveraging approximate computing to group population counts into coarse-grained buckets, our design achieves hardware area reductions while preserving the link power benefits of data reordering. Our approximate sorting unit achieves up to 35.4% area reduction while maintaining 19.50\% BT reduction compared to 20.42% of precise implementation.
title '1'-bit Count-based Sorting Unit to Reduce Link Power in DNN Accelerators
topic Hardware Architecture
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2601.14087