A Reconfigurable 64-Dimension K-Means Clustering Accelerator with Adaptive Overflow Control

Li Du, Yuan Du*, Mau Chung Frank Chang

*Corresponding author for this work

Research output: Contribution to journalArticle

1 Scopus citations

Abstract

This brief presents a novel reconfigurable K -means clustering accelerator that is suitable for integration in both IoT and data center system. The high vector dimension reconfigurability and design cost reduction is achieved through vector-streaming and adaptive overflow control to adapt distance computation using as-needed precision (dynamic 16-bit fixed-point data format). A two-stage shift-bit counted comparator is proposed. It can determine most results through only turning on the shift-bit comparator (3-bit), reducing the power consumption by 7\times compared to the direct full dynamic range comparison. Four vectors with two cluster centroids are processed simultaneously. Up to 8-dimension cluster vectors are stored in local buffer to reduce data exchange between the main memory and the processing engine. A prototype accelerator was implemented in TSMC 65 nm. The accelerator occupied 0.26 mm2 and can support up to 64-D vector clustering. It achieved 31.2M query vectors/s with 41-mW power consumption at 250-MHz clock (cluster number: 2, vector dimension: 64) and an energy efficiency of 0.41 TOPS/W at 30 MHz clock.

Original languageEnglish
Article number8736399
Pages (from-to)760-764
Number of pages5
JournalIEEE Transactions on Circuits and Systems II: Express Briefs
Volume67
Issue number4
DOIs
StatePublished - Apr 2020

Keywords

  • clustering
  • hardware accelerator
  • K-means
  • Machine learning
  • unsupervised learning
  • vector flow

Fingerprint Dive into the research topics of 'A Reconfigurable 64-Dimension K-Means Clustering Accelerator with Adaptive Overflow Control'. Together they form a unique fingerprint.

  • Cite this