PAIRS: Pruning-AIded Row-Skipping for SDK-Based Convolutional Weight Mapping in Processing-In-Memory Architectures
- Authors
- Rhe, J.[Rhe, Johnny]; Jeon, K.E.[Jeon, Kang Eun]; Kot, J.H.[Kot, Jong Hwan]
- Issue Date
- 2023
- Publisher
- Institute of Electrical and Electronics Engineers Inc.
- Keywords
- convolutional neural network; pattern-based pruning; processing in memory; weight mapping
- Citation
- Proceedings of the International Symposium on Low Power Electronics and Design, v.2023-August
- Indexed
- SCOPUS
- Journal Title
- Proceedings of the International Symposium on Low Power Electronics and Design
- Volume
- 2023-August
- URI
- https://scholarworks.bwise.kr/skku/handle/2021.sw.skku/109023
- DOI
- 10.1109/ISLPED58423.2023.10244367
- ISSN
- 1533-4678
- Abstract
- Processing-in-memory (PIM) architecture is becoming a promising candidate for convolutional neural network (CNN) inference. A recent weight mapping method called shift and duplicate kernel (SDK) improves the utilization by the deployment of shifting the same kernels into idle columns. However, this method inevitably generates idle cells with an irregular distribution, which limits reducing the size of the weight matrix. To effectively compress the weight matrix in the PIM array, prior works have introduced a row-wise pruning scheme, one of the structured weight pruning schemes, that aims to skip the operation on a row by zeroing out all weight in the specific row (we call it row-skipping). However, due to the deployment of shifting kernels, SDK mapping complicates zeroing out all the weight in the same row. To address this issue, we propose pruning-aided row-skipping (PAIRS) that effectively reduces the number of rows of convolutional weights that are mapped with SDK mapping. By pairing the SDK mapping-aware pruning pattern design and row-wise pruning, PAIRS achieves a higher row-skipping ratio. In comparison to pruning methods, PAIRS achieves up to 1.95× rows skipped and 4× higher compression rate with similar or even better inference accuracy. © 2023 IEEE.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - Information and Communication Engineering > School of Electronic and Electrical Engineering > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.