Parallel R-tree–based Spatial Query Processing on a Commercial Processing-in-Memory System

Tasmia Jannat 1 , Michael Gowanlock 2 , Satish Puri 1 · 2026

Read Parallel R-tree–based Spatial Query Processing on a Commercial Processing-in-Memory System by Tasmia Jannat 1 , Michael Gowanlock 2 , Satish Puri 1 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-14445.txt.

Description

Jannat, Gowanlock, and Puri run R-tree range queries on commercial memory chips, cutting data-movement cost. Their broadcast method builds the tree bottom-up on the CPU, sends top levels to UPMEM units to filter globally, and farms lower levels out in parallel. On Sports and Lakes data and synthetic tests on up to 2,540 memory units, broadcast execution beats subtree partitioning by keeping communication from swamping work. On Lakes, kernel speedups reach up to 3.66 times over CPU search, while the memory kernel uses roughly a third of the energy.

Catalog record

CREATORTasmia Jannat 1 , Michael Gowanlock 2 , Satish Puri 1
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53167
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.14445

Browse the full library · About the Tower of Babel