Adaptive parallel approximate similarity search for responsive multimedia retrieval

  • Authors:
  • George Teodoro;Eduardo Valle;Nathan Mariano;Ricardo Torres;Wagner Meira, Jr.

  • Affiliations:
  • University of Maryland, College Park, MD, USA;Unicamp, Campinas, Brazil;UFMG, Belo Horizonte, Brazil;Unicamp, Campinas, Brazil;UFMG, Belo Horizonte, Brazil

  • Venue:
  • Proceedings of the 20th ACM international conference on Information and knowledge management
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper introduces Hypercurves, a flexible framework for pro- viding similarity search indexing to high throughput multimedia services. Hypercurves efficiently and effectively answers k-nearest neighbor searches on multigigabyte high-dimensional databases. It supports massively parallel processing and adapts at runtime its parallelization regimens to keep answer times optimal for either low and high demands. In order to achieve its goals, Hypercurves introduces new techniques for selecting parallelism configurations and allocating threads to computation cores, including hyperthreaded cores. Its efficiency gains are throughly validated on a large database of multimedia descriptors, where it presented near linear speedups and superlinear scaleups. The adaptation reduces query response times in 43% and 74% for both platforms tested, when compared to the best static parallelism regimens.