CRSNG Subvention à la Découverte/NSERC Discovery Grant (2017-2024)

Researcher : Daniel Lemire

Amount : 294 000 $

We seek to accelerate software indexing techniques, either within search engines or within databases. In this work, we exploit recent and emerging hardware capabilities. In particular, we seek to fully benefit from vector instructions. To keep the memory close to the processor, we seek to improve index compression, whether they are inverted indexes, B-trees or bitmap indexes. We seek to uncompress data at great speed in RAM. We want to accelerate common operations such as intersections and unions.