'1'-bit Count-based Sorting Unit to Reduce Link Power in DNN Accelerators
Research paper by Ruichi Han and collaborators. arXiv record dated 2026-01-20. Research area: Artificial intelligence.
Resource information
- Authors
- Ruichi Han, Yizhi Chen, Tong Lei, Jordi Altayo Gonzalez, Ahmed Hemani
- Research area
- Artificial intelligence
- Application domain
- General AI research
- AI methods
- Neural networks
- Research topics
- cs.AR, cs.AI, cs.LG
- arXiv record date
- 2026-01-20
- Publication
- arXiv
- Upstream last modified
- Jan 21, 2026
- Last record update
- Nov 4, 2026
Abstract
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.